examples: make the libmgmt example more fun

You can try it out yourself by running `go build` and then calling it.
Use a bare integer argument to create that number of noop resources.
There are clearly some performance optimizations that we could do for
extremely large graphs.

.dockerignore

@@ -0,0 +1 @@
+docker

.editorconfig

@@ -0,0 +1,19 @@
+; This file is for unifying the coding style for different editors and IDEs.
+; Plugins are available for notepad++, emacs, vim, gedit,
+; textmate, visual studio, and more.
+;
+; See http://editorconfig.org for details.
+
+# Top-most EditorConfig file.
+root = true
+
+[*]
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true
+
+[*.go]
+indent_style = tab
+
+[Makefile]
+indent_style = tab

.gitignore

@@ -6,4 +6,5 @@ old/
 tmp/
 *_stringer.go
 mgmt
+mgmt.static
 rpmbuild/

.gitmodules

@@ -0,0 +1,12 @@
+[submodule "vendor/github.com/coreos/etcd"]
+	path = vendor/github.com/coreos/etcd
+	url = https://github.com/coreos/etcd/
+[submodule "vendor/google.golang.org/grpc"]
+	path = vendor/google.golang.org/grpc
+	url = https://github.com/grpc/grpc-go
+[submodule "vendor/github.com/grpc-ecosystem/grpc-gateway"]
+	path = vendor/github.com/grpc-ecosystem/grpc-gateway
+	url = https://github.com/grpc-ecosystem/grpc-gateway
+[submodule "vendor/gopkg.in/fsnotify.v1"]
+	path = vendor/gopkg.in/fsnotify.v1
+	url = https://gopkg.in/fsnotify.v1

.travis.yml

@@ -1,17 +1,18 @@
 language: go
 go:
-  - 1.4.3
-  - 1.5.3
   - 1.6
+  - 1.7
   - tip
-sudo: false
+sudo: true
+dist: trusty
 before_install: 'git fetch --unshallow'
 install: 'make deps'
 script: 'make test'
 matrix:
+  fast_finish: true
   allow_failures:
     - go: tip
-    - go: 1.4.3
+    - go: 1.7
 notifications:
   irc:
     channels:

AUTHORS

@@ -5,3 +5,4 @@ For a more exhaustive list please run: git log --format='%aN' | sort -u
 This list is sorted alphabetically by first name.
 
 James Shubin
+Paul Morgan

DOCUMENTATION.md

@@ -33,18 +33,23 @@ along with this program.  If not, see <http://www.gnu.org/licenses/>.
 4. [Features - All things mgmt can do](#features)
 	* [Autoedges - Automatic resource relationships](#autoedges)
 	* [Autogrouping - Automatic resource grouping](#autogrouping)
-5. [Usage/FAQ - Notes on usage and frequently asked questions](#usage-and-frequently-asked-questions)
+	* [Automatic clustering - Automatic cluster management](#automatic-clustering)
-6. [Reference - Detailed reference](#reference)
+	* [Remote mode - Remote "agent-less" execution](#remote-agent-less-mode)
+	* [Puppet support - write manifest code for mgmt](#puppet-support)
+5. [Resources - All built-in primitives](#resources)
+6. [Usage/FAQ - Notes on usage and frequently asked questions](#usage-and-frequently-asked-questions)
+7. [Reference - Detailed reference](#reference)
+	* [Meta parameters](#meta-parameters)
 	* [Graph definition file](#graph-definition-file)
 	* [Command line](#command-line)
-7. [Examples - Example configurations](#examples)
+8. [Examples - Example configurations](#examples)
-8. [Development - Background on module development and reporting bugs](#development)
+9. [Development - Background on module development and reporting bugs](#development)
-9. [Authors - Authors and contact information](#authors)
+10. [Authors - Authors and contact information](#authors)
 
 ##Overview
 
-The `mgmt` tool is a research prototype to demonstrate next generation config
+The `mgmt` tool is a next generation config management prototype. It's not yet
-management techniques. Hopefully it will evolve into a useful, robust tool.
+ready for production, but we hope to get there soon. Get involved today!
 
 ##Project Description
 
@@ -56,13 +61,16 @@ For more information, you may like to read some blog posts from the author:
 
 * [Next generation config mgmt](https://ttboj.wordpress.com/2016/01/18/next-generation-configuration-mgmt/)
 * [Automatic edges in mgmt](https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/)
+* [Automatic grouping in mgmt](https://ttboj.wordpress.com/2016/03/30/automatic-grouping-in-mgmt/)
+* [Automatic clustering in mgmt](https://ttboj.wordpress.com/2016/06/20/automatic-clustering-in-mgmt/)
 
-There is also an [introductory video](https://www.youtube.com/watch?v=GVhpPF0j-iE&html5=1) available.
+There is also an [introductory video](http://meetings-archive.debian.net/pub/debian-meetings/2016/debconf16/Next_Generation_Config_Mgmt.webm) available.
+Older videos and other material [is available](https://github.com/purpleidea/mgmt/#on-the-web).
 
 ##Setup
 
 During this prototype phase, the tool can be run out of the source directory.
-You'll probably want to use ```./run.sh run --file examples/graph1.yaml``` to
+You'll probably want to use ```./run.sh run --yaml examples/graph1.yaml``` to
 get started. Beware that this _can_ cause data loss. Understand what you're
 doing first, or perform these actions in a virtual environment such as the one
 provided by [Oh-My-Vagrant](https://github.com/purpleidea/oh-my-vagrant).
@@ -81,7 +89,7 @@ automatically ensure that any file resource you declare that matches a
 file installed by your package resource will only be processed after the
 package is installed.
 
-####Controlling autodeges
+####Controlling autoedges
 
 Though autoedges is likely to be very helpful and avoid you having to declare
 all dependencies explicitly, there are cases where this behaviour is
@@ -98,6 +106,11 @@ installation of the `mysql-server` package.
 You can disable autoedges for a resource by setting the `autoedge` key on
 the meta attributes of that resource to `false`.
 
+####Blog post
+
+You can read the introductory blog post about this topic here:
+[https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/](https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/)
+
 ###Autogrouping
 
 Automatic grouping or AutoGroup is the mechanism in mgmt by which it will
@@ -112,6 +125,173 @@ used for other use cases too.
 You can disable autogrouping for a resource by setting the `autogroup` key on
 the meta attributes of that resource to `false`.
 
+####Blog post
+
+You can read the introductory blog post about this topic here:
+[https://ttboj.wordpress.com/2016/03/30/automatic-grouping-in-mgmt/](https://ttboj.wordpress.com/2016/03/30/automatic-grouping-in-mgmt/)
+
+###Automatic clustering
+
+Automatic clustering is a feature by which mgmt automatically builds, scales,
+and manages the embedded etcd cluster which is compiled into mgmt itself. It is
+quite helpful for rapidly bootstrapping clusters and avoiding the extra work to
+setup etcd.
+
+If you prefer to avoid this feature. you can always opt to use an existing etcd
+cluster that is managed separately from mgmt by pointing your mgmt agents at it
+with the `--seeds` variable.
+
+####Blog post
+
+You can read the introductory blog post about this topic here:
+[https://ttboj.wordpress.com/2016/06/20/automatic-clustering-in-mgmt/](https://ttboj.wordpress.com/2016/06/20/automatic-clustering-in-mgmt/)
+
+###Remote ("agent-less") mode
+
+Remote mode is a special mode that lets you kick off mgmt runs on one or more
+remote machines which are only accessible via SSH. In this mode the initiating
+host connects over SSH, copies over the `mgmt` binary, opens an SSH tunnel, and
+runs the remote program while simultaneously passing the etcd traffic back
+through the tunnel so that the initiators etcd cluster can be used to exchange
+resource data.
+
+The interesting benefit of this architecture is that multiple hosts which can't
+connect directly use the initiator to pass the important traffic through to each
+other. Once the cluster has converged all the remote programs can shutdown
+leaving no residual agent.
+
+This mode can also be useful for bootstrapping a new host where you'd like to
+have the service run continuously and as part of an mgmt cluster normally.
+
+In particular, when combined with the `--converged-timeout` parameter, the
+entire set of running mgmt agents will need to all simultaneously converge for
+the group to exit. This is particularly useful for bootstrapping new clusters
+which need to exchange information that is only available at run time.
+
+####Blog post
+
+You can read the introductory blog post about this topic here:
+[https://ttboj.wordpress.com/2016/10/07/remote-execution-in-mgmt/](https://ttboj.wordpress.com/2016/10/07/remote-execution-in-mgmt/)
+
+###Puppet support
+
+You can supply a Puppet manifest instead of creating the (YAML) graph manually.
+Puppet must be installed and in `mgmt`'s search path. You also need the
+[ffrank-mgmtgraph Puppet module](https://forge.puppet.com/ffrank/mgmtgraph).
+
+Invoke `mgmt` with the `--puppet` switch, which supports 3 variants:
+
+1. Request the configuration from the Puppet Master (like `puppet agent` does)
+
+        mgmt run --puppet agent
+
+2. Compile a local manifest file (like `puppet apply`)
+
+        mgmt run --puppet /path/to/my/manifest.pp
+
+3. Compile an ad hoc manifest from the commandline (like `puppet apply -e`)
+
+        mgmt run --puppet 'file { "/etc/ntp.conf": ensure => file }'
+
+For more details and caveats see [Puppet.md](Puppet.md).
+
+####Blog post
+
+An introductory post on the Puppet support is on
+[Felix's blog](http://ffrank.github.io/features/2016/06/19/puppet-powered-mgmt/).
+
+##Resources
+
+This section lists all the built-in resources and their properties. The
+resource primitives in `mgmt` are typically more powerful than resources in
+other configuration management systems because they can be event based which
+lets them respond in real-time to converge to the desired state. This property
+allows you to build more complex resources that you probably hadn't considered
+in the past.
+
+In addition to the resource specific properties, there are resource properties
+(otherwise known as parameters) which can apply to every resource. These are
+called [meta parameters](#meta-parameters) and are listed separately. Certain
+meta parameters aren't very useful when combined with certain resources, but
+in general, it should be fairly obvious, such as when combining the `noop` meta
+parameter with the [Noop](#Noop) resource.
+
+* [Exec](#Exec): Execute shell commands on the system.
+* [File](#File): Manage files and directories.
+* [Msg](#Msg): Send log messages.
+* [Noop](#Noop): A simple resource that does nothing.
+* [Pkg](#Pkg):  Manage system packages with PackageKit.
+* [Svc](#Svc): Manage system systemd services.
+* [Timer](#Timer): Manage system systemd services.
+* [Virt](#Virt): Manage virtual machines with libvirt.
+
+###Exec
+
+The exec resource can execute commands on your system.
+
+###File
+
+The file resource manages files and directories. In `mgmt`, directories are
+identified by a trailing slash in their path name. File have no such slash.
+
+####Path
+
+The path property specifies the file or directory that we are managing.
+
+####Content
+
+The content property is a string that specifies the desired file contents.
+
+####Source
+
+The source property points to a source file or directory path that we wish to
+copy over and use as the desired contents for our resource.
+
+####State
+
+The state property describes the action we'd like to apply for the resource. The
+possible values are: `exists` and `absent`.
+
+####Recurse
+
+The recurse property limits whether file resource operations should recurse into
+and monitor directory contents with a depth greater than one.
+
+####Force
+
+The force property is required if we want the file resource to be able to change
+a file into a directory or vice-versa. If such a change is needed, but the force
+property is not set to `true`, then this file resource will error.
+
+###Msg
+
+The msg resource sends messages to the main log, or an external service such
+as systemd's journal.
+
+###Noop
+
+The noop resource does absolutely nothing. It does have some utility in testing
+`mgmt` and also as a placeholder in the resource graph.
+
+###Pkg
+
+The pkg resource is used to manage system packages. This resource works on many
+different distributions because it uses the underlying packagekit facility which
+supports different backends for different environments. This ensures that we
+have great Debian (deb/dpkg) and Fedora (rpm/dnf) support simultaneously.
+
+###Svc
+
+The service resource is still very WIP. Please help us my improving it!
+
+###Timer
+
+This resource needs better documentation. Please help us my improving it!
+
+###Virt
+
+The virt resource can manage virtual machines via libvirt.
+
 ##Usage and frequently asked questions
 (Send your questions as a patch to this FAQ! I'll review it, merge it, and
 respond by commit with the answer.)
@@ -130,6 +310,58 @@ chosen, but it was also somewhat arbitrary. If there is available interest,
 good reasoning, *and* patches, then we would consider either switching or
 supporting both, but this is not a high priority at this time.
 
+###Can I use an existing etcd cluster instead of the automatic embedded servers?
+
+Yes, it's possible to use an existing etcd cluster instead of the automatic,
+elastic embedded etcd servers. To do so, simply point to the cluster with the
+`--seeds` variable, the same way you would if you were seeding a new member to
+an existing mgmt cluster.
+
+The downside to this approach is that you won't benefit from the automatic
+elastic nature of the embedded etcd servers, and that you're responsible if you
+accidentally break your etcd cluster, or if you use an unsupported version.
+
+###What does the error message about an inconsistent dataDir mean?
+
+If you get an error message similar to:
+
+```
+Etcd: Connect: CtxError...
+Etcd: CtxError: Reason: CtxDelayErr(5s): No endpoints available yet!
+Etcd: Connect: Endpoints: []
+Etcd: The dataDir (/var/lib/mgmt/etcd) might be inconsistent or corrupt.
+```
+
+This happens when there are a series of fatal connect errors in a row. This can
+happen when you start `mgmt` using a dataDir that doesn't correspond to the
+current cluster view. As a result, the embedded etcd server never finishes
+starting up, and as a result, a default endpoint never gets added. The solution
+is to either reconcile the mistake, and if there is no important data saved, you
+can remove the etcd dataDir. This is typically `/var/lib/mgmt/etcd/member/`.
+
+###Why do resources have both a `Compare` method and an `IFF` (on the UID) method?
+
+The `Compare()` methods are for determining if two resources are effectively the
+same, which is used to make graph change delta's efficient. This is when we want
+to change from the current running graph to a new graph, but preserve the common
+vertices. Since we want to make this process efficient, we only update the parts
+that are different, and leave everything else alone. This `Compare()` method can
+tell us if two resources are the same.
+
+The `IFF()` method is part of the whole UID system, which is for discerning if a
+resource meets the requirements another expects for an automatic edge. This is
+because the automatic edge system assumes a unified UID pattern to test for
+equality. In the future it might be helpful or sane to merge the two similar
+comparison functions although for now they are separate because they are
+actually answer different questions.
+
+###Did you know that there is a band named `MGMT`?
+
+I didn't realize this when naming the project, and it is accidental. After much
+anguishing, I chose the name because it was short and I thought it was
+appropriately descriptive. If you need a less ambiguous search term or phrase,
+you can try using `mgmtconfig` or `mgmt config`.
+
 ###You didn't answer my question, or I have a question!
 
 It's best to ask on [IRC](https://webchat.freenode.net/?channels=#mgmtconfig)
@@ -147,9 +379,40 @@ information on these options, please view the source at:
 If you feel that a well used option needs documenting here, please patch it!
 
 ###Overview of reference
+* [Meta parameters](#meta-parameters): List of available resource meta parameters.
 * [Graph definition file](#graph-definition-file): Main graph definition file.
 * [Command line](#command-line): Command line parameters.
 
+###Meta parameters
+These meta parameters are special parameters (or properties) which can apply to
+any resource. The usefulness of doing so will depend on the particular meta
+parameter and resource combination.
+
+####AutoEdge
+Boolean. Should we generate auto edges for this resource?
+
+####AutoGroup
+Boolean. Should we attempt to automatically group this resource with others?
+
+####Noop
+Boolean. Should the Apply portion of the CheckApply method of the resource
+make any changes? Noop is a concatenation of no-operation.
+
+####Retry
+Integer. The number of times to retry running the resource on error. Use -1 for
+infinite. This currently applies for both the Watch operation (which can fail)
+and for the CheckApply operation. While they could have separate values, I've
+decided to use the same ones for both until there's a proper reason to want to
+do something differently for the Watch errors.
+
+####Delay
+Integer. Number of milliseconds to wait between retries. The same value is
+shared between the Watch and CheckApply retries. This currently applies for both
+the Watch operation (which can fail) and for the CheckApply operation. While
+they could have separate values, I've decided to use the same ones for both
+until there's a proper reason to want to do something differently for the Watch
+errors.
+
 ###Graph definition file
 graph.yaml is the compiled graph definition file. The format is currently
 undocumented, but by looking through the [examples/](https://github.com/purpleidea/mgmt/tree/master/examples)
@@ -159,7 +422,7 @@ you can probably figure out most of it, as it's fairly intuitive.
 The main interface to the `mgmt` tool is the command line. For the most recent
 documentation, please run `mgmt --help`.
 
-####`--file <graph.yaml>`
+####`--yaml <graph.yaml>`
 Point to a graph file to run.
 
 ####`--converged-timeout <seconds>`
@@ -169,12 +432,79 @@ Exit if the machine has converged for approximately this many seconds.
 Exit when the agent has run for approximately this many seconds. This is not
 generally recommended, but may be useful for users who know what they're doing.
 
+####`--noop`
+Globally force all resources into no-op mode. This also disables the export to
+etcd functionality, but does not disable resource collection, however all
+resources that are collected will have their individual noop settings set.
+
+####`--remote <graph.yaml>`
+Point to a graph file to run on the remote host specified within. This parameter
+can be used multiple times if you'd like to remotely run on multiple hosts in
+parallel.
+
+####`--allow-interactive`
+Allow interactive prompting for SSH passwords if there is no authentication
+method that works.
+
+####`--ssh-priv-id-rsa`
+Specify the path for finding SSH keys. This defaults to `~/.ssh/id_rsa`. To
+never use this method of authentication, set this to the empty string.
+
+####`--cconns`
+The maximum number of concurrent remote ssh connections to run. This defaults
+to `0`, which means unlimited.
+
+####`--no-caching`
+Don't allow remote caching of the remote execution binary. This will require
+the binary to be copied over for every remote execution, but it limits the
+likelihood that there is leftover information from the configuration process.
+
+####`--prefix <path>`
+Specify a path to a custom working directory prefix. This directory will get
+created if it does not exist. This usually defaults to `/var/lib/mgmt/`. This
+can't be combined with the `--tmp-prefix` option. It can be combined with the
+`--allow-tmp-prefix` option.
+
+####`--tmp-prefix`
+If this option is specified, a temporary prefix will be used instead of the
+default prefix. This can't be combined with the `--prefix` option.
+
+####`--allow-tmp-prefix`
+If this option is specified, we will attempt to fall back to a temporary prefix
+if the primary prefix couldn't be created. This is useful for avoiding failures
+in environments where the primary prefix may or may not be available, but you'd
+like to try. The canonical example is when running `mgmt` with `--remote` there
+might be a cached copy of the binary in the primary prefix, but in case there's
+no binary available continue working in a temporary directory to avoid failure.
+
 ##Examples
 For example configurations, please consult the [examples/](https://github.com/purpleidea/mgmt/tree/master/examples) directory in the git
 source repository. It is available from:
 
 [https://github.com/purpleidea/mgmt/tree/master/examples](https://github.com/purpleidea/mgmt/tree/master/examples)
 
+### Systemd:
+See [`misc/mgmt.service`](misc/mgmt.service) for a sample systemd unit file.
+This unit file is part of the RPM.
+
+To specify your custom options for `mgmt` on a systemd distro:
+
+```bash
+sudo mkdir -p /etc/systemd/system/mgmt.service.d/
+
+cat > /etc/systemd/system/mgmt.service.d/env.conf <<EOF
+# Environment variables:
+MGMT_SEEDS=http://127.0.0.1:2379
+MGMT_CONVERGED_TIMEOUT=-1
+MGMT_MAX_RUNTIME=0
+
+# Other CLI options if necessary.
+#OPTS="--max-runtime=0"
+EOF
+
+sudo systemctl daemon-reload
+```
+
 ##Development
 
 This is a project that I started in my free time in 2013. Development is driven

Makefile

@@ -15,12 +15,12 @@
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-SHELL = /bin/bash
+SHELL = /usr/bin/env bash
-.PHONY: all version program path deps run race build clean test format docs rpmbuild mkdirs rpm srpm spec tar upload upload-sources upload-srpms upload-rpms copr
+.PHONY: all art cleanart version program path deps run race generate build clean test gofmt yamlfmt format docs rpmbuild mkdirs rpm srpm spec tar upload upload-sources upload-srpms upload-rpms copr
 .SILENT: clean
 
-SVERSION := $(shell git describe --match '[0-9]*\.[0-9]*\.[0-9]*' --tags --dirty --always)
+SVERSION := $(or $(SVERSION),$(shell git describe --match '[0-9]*\.[0-9]*\.[0-9]*' --tags --dirty --always))
-VERSION := $(shell git describe --match '[0-9]*\.[0-9]*\.[0-9]*' --tags --abbrev=0)
+VERSION := $(or $(VERSION),$(shell git describe --match '[0-9]*\.[0-9]*\.[0-9]*' --tags --abbrev=0))
 PROGRAM := $(shell echo $(notdir $(CURDIR)) | cut -f1 -d"-")
 OLDGOLANG := $(shell go version | grep -E 'go1.3|go1.4')
 ifeq ($(VERSION),$(SVERSION))
@@ -28,7 +28,7 @@ ifeq ($(VERSION),$(SVERSION))
 else
 	RELEASE = untagged
 endif
-ARCH = $(shell arch)
+ARCH = $(uname -m)
 SPEC = rpmbuild/SPECS/$(PROGRAM).spec
 SOURCE = rpmbuild/SOURCES/$(PROGRAM)-$(VERSION).tar.bz2
 SRPM = rpmbuild/SRPMS/$(PROGRAM)-$(VERSION)-$(RELEASE).src.rpm
@@ -38,7 +38,43 @@ USERNAME := $(shell cat ~/.config/copr 2>/dev/null | grep username | awk -F '='
 SERVER = 'dl.fedoraproject.org'
 REMOTE_PATH = 'pub/alt/$(USERNAME)/$(PROGRAM)'
 
-all: docs
+#
+#	art
+#
+art: art/mgmt_logo_default_symbol.png art/mgmt_logo_default_tall.png art/mgmt_logo_default_wide.png art/mgmt_logo_reversed_symbol.png art/mgmt_logo_reversed_tall.png art/mgmt_logo_reversed_wide.png art/mgmt_logo_white_symbol.png art/mgmt_logo_white_tall.png art/mgmt_logo_white_wide.png
+
+cleanart:
+	rm -f art/mgmt_logo_default_symbol.png art/mgmt_logo_default_tall.png art/mgmt_logo_default_wide.png art/mgmt_logo_reversed_symbol.png art/mgmt_logo_reversed_tall.png art/mgmt_logo_reversed_wide.png art/mgmt_logo_white_symbol.png art/mgmt_logo_white_tall.png art/mgmt_logo_white_wide.png
+
+# NOTE: the widths are arbitrary
+art/mgmt_logo_default_symbol.png: art/mgmt_logo_default_symbol.svg
+	inkscape --export-background='#ffffff' --without-gui --export-png "$@" --export-width 300 $(@:png=svg)
+
+art/mgmt_logo_default_tall.png: art/mgmt_logo_default_tall.svg
+	inkscape --export-background='#ffffff' --without-gui --export-png "$@" --export-width 400 $(@:png=svg)
+
+art/mgmt_logo_default_wide.png: art/mgmt_logo_default_wide.svg
+	inkscape --export-background='#ffffff' --without-gui --export-png "$@" --export-width 800 $(@:png=svg)
+
+art/mgmt_logo_reversed_symbol.png: art/mgmt_logo_reversed_symbol.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 300 $(@:png=svg)
+
+art/mgmt_logo_reversed_tall.png: art/mgmt_logo_reversed_tall.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 400 $(@:png=svg)
+
+art/mgmt_logo_reversed_wide.png: art/mgmt_logo_reversed_wide.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 800 $(@:png=svg)
+
+art/mgmt_logo_white_symbol.png: art/mgmt_logo_white_symbol.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 300 $(@:png=svg)
+
+art/mgmt_logo_white_tall.png: art/mgmt_logo_white_tall.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 400 $(@:png=svg)
+
+art/mgmt_logo_white_wide.png: art/mgmt_logo_white_wide.svg
+	inkscape --export-background='#231f20' --without-gui --export-png "$@" --export-width 800 $(@:png=svg)
+
+all: docs $(PROGRAM).static
 
 # show the current version
 version:
@@ -54,34 +90,51 @@ deps:
 	./misc/make-deps.sh
 
 run:
-	find -maxdepth 1 -type f -name '*.go' -not -name '*_test.go' | xargs go run -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)"
+	find . -maxdepth 1 -type f -name '*.go' -not -name '*_test.go' | xargs go run -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)"
 
 # include race flag
 race:
-	find -maxdepth 1 -type f -name '*.go' -not -name '*_test.go' | xargs go run -race -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)"
+	find . -maxdepth 1 -type f -name '*.go' -not -name '*_test.go' | xargs go run -race -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)"
+
+generate:
+	go generate
 
 build: $(PROGRAM)
 
 $(PROGRAM): main.go
 	@echo "Building: $(PROGRAM), version: $(SVERSION)..."
+ifneq ($(OLDGOLANG),)
+	@# avoid equals sign in old golang versions eg in: -X foo=bar
+	time go build -ldflags "-X main.program $(PROGRAM) -X main.version $(SVERSION)" -o $(PROGRAM);
+else
+	time go build -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)" -o $(PROGRAM);
+endif
+
+$(PROGRAM).static: main.go
+	@echo "Building: $(PROGRAM).static, version: $(SVERSION)..."
 	go generate
 ifneq ($(OLDGOLANG),)
 	@# avoid equals sign in old golang versions eg in: -X foo=bar
-	go build -ldflags "-X main.program $(PROGRAM) -X main.version $(SVERSION)" -o $(PROGRAM);
+	go build -a -installsuffix cgo -tags netgo -ldflags '-extldflags "-static" -X main.program $(PROGRAM) -X main.version $(SVERSION)' -o $(PROGRAM).static;
 else
-	go build -ldflags "-X main.program=$(PROGRAM) -X main.version=$(SVERSION)" -o $(PROGRAM);
+	go build -a -installsuffix cgo -tags netgo -ldflags '-extldflags "-static" -X main.program=$(PROGRAM) -X main.version=$(SVERSION)' -o $(PROGRAM).static;
 endif
 
 clean:
 	[ ! -e $(PROGRAM) ] || rm $(PROGRAM)
 	rm -f *_stringer.go	# generated by `go generate`
+	rm -f *_mock.go		# generated by `go generate`
 
 test:
 	./test.sh
 
-format:
+gofmt:
-	find -type f -name '*.go' -not -path './old/*' -not -path './tmp/*' -exec gofmt -w {} \;
+	find . -maxdepth 3 -type f -name '*.go' -not -path './old/*' -not -path './tmp/*' -exec gofmt -w {} \;
-	find -type f -name '*.yaml' -not -path './old/*' -not -path './tmp/*' -not -path './omv.yaml' -exec ruby -e "require 'yaml'; x=YAML.load_file('{}').to_yaml.each_line.map(&:rstrip).join(10.chr)+10.chr; File.open('{}', 'w').write x" \;
+
+yamlfmt:
+	find . -maxdepth 3 -type f -name '*.yaml' -not -path './old/*' -not -path './tmp/*' -not -path './omv.yaml' -exec ruby -e "require 'yaml'; x=YAML.load_file('{}').to_yaml.each_line.map(&:rstrip).join(10.chr)+10.chr; File.open('{}', 'w').write x" \;
+
+format: gofmt yamlfmt
 
 docs: $(PROGRAM)-documentation.pdf
 
@@ -131,7 +184,7 @@ $(SRPM): $(SPEC) $(SOURCE)
 $(SPEC): rpmbuild/ spec.in
 	@echo Running templater...
 	#cat spec.in > $(SPEC)
-	sed -e s/__PROGRAM__/$(PROGRAM)/ -e s/__VERSION__/$(VERSION)/ -e s/__RELEASE__/$(RELEASE)/ < spec.in > $(SPEC)
+	sed -e s/__PROGRAM__/$(PROGRAM)/g -e s/__VERSION__/$(VERSION)/g -e s/__RELEASE__/$(RELEASE)/g < spec.in > $(SPEC)
 	# append a changelog to the .spec file
 	git log --format="* %cd %aN <%aE>%n- (%h) %s%d%n" --date=local | sed -r 's/[0-9]+:[0-9]+:[0-9]+ //' >> $(SPEC)
 

Puppet.md

@@ -0,0 +1,163 @@
+#mgmt Puppet support
+
+1. [Prerequisites](#prerequisites)
+    * [Testing the Puppet side](#testing-the-puppet-side)
+2. [Writing a suitable manifest](#writing-a-suitable-manifest)
+    * [Unsupported attributes](#unsupported-attributes)
+    * [Unsupported resources](#unsupported-resources)
+    * [Avoiding common warnings](#avoiding-common-warnings)
+3. [Configuring Puppet](#configuring-puppet)
+4. [Caveats](#caveats)
+
+`mgmt` can use Puppet as its source for the configuration graph.
+This document goes into detail on how this works, and lists
+some pitfalls and limitations.
+
+For basic instructions on how to use the Puppet support, see
+the [main documentation](DOCUMENTATION.md#puppet-support).
+
+##Prerequisites
+
+You need Puppet installed in your system. It is not important how you
+get it. On the most common Linux distributions, you can use packages
+from the OS maintainer, or upstream Puppet repositories. An alternative
+that will also work on OSX is the `puppet` Ruby gem. It also has the
+advantage that you can install any desired version in your home directory
+or any other location.
+
+Any release of Puppet's 3.x and 4.x series should be suitable for use with
+`mgmt`. Most importantly, make sure to install the `ffrank-mgmtgraph` Puppet
+module (referred to below as "the translator module").
+
+    puppet module install ffrank-mgmtgraph
+
+Please note that the module is not required on your Puppet master (if you
+use a master/agent setup). It's needed on the machine that runs `mgmt`.
+You can install the module on the master anyway, so that it gets distributed
+to your agents through Puppet's `pluginsync` mechanism.
+
+###Testing the Puppet side
+
+The following command should run successfully and print a YAML hash on your
+terminal:
+
+```puppet
+puppet mgmtgraph print --code 'file { "/tmp/mgmt-test": ensure => present }'
+```
+
+You can use this CLI to test any manifests before handing them straight
+to `mgmt`.
+
+##Writing a suitable manifest
+
+###Unsupported attributes
+
+`mgmt` inherited its resource module from Puppet, so by and large, it's quite
+possible to express `mgmt` graphs in terms of Puppet manifests. However,
+there isn't (and likely never will be) full feature parity between the
+respective resource types. In consequence, a manifest can have semantics that
+cannot be transferred to `mgmt`.
+
+For example, at the time of writing this, the `file` type in `mgmt` had no
+notion of permissions (the file `mode`) yet. This lead to the following
+warning (among others that will be discussed below):
+
+    $ puppet mgmtgraph print --code 'file { "/tmp/foo": mode => "0600" }'
+    Warning: cannot translate: File[/tmp/foo] { mode => "600" } (attribute is ignored)
+
+This is a heads-up for the user, because the resulting `mgmt` graph will
+in fact not pass this information to the `/tmp/foo` file resource, and
+`mgmt` will ignore this file's permissions. Including such attributes in
+manifests that are written expressly for `mgmt` is not sensible and should
+be avoided.
+
+###Unsupported resources
+
+Puppet has a fairly large number of
+[built-in types](https://docs.puppet.com/puppet/latest/reference/type.html),
+and countless more are available through
+[modules](https://forge.puppet.com/). It's unlikely that all of them will
+eventually receive native counterparts in `mgmt`.
+
+When encountering an unknown resource, the translator module will replace
+it with an `exec` resource in its output. This resource will run the equivalent
+of a `puppet resource` command to make Puppet apply the original resource
+itself. This has quite abysmal performance, because processing such a
+resource requires the forking of at least one Puppet process (two if it
+is found to be out of sync). This comes with considerable overhead.
+On most systems, starting up any Puppet command takes several seconds.
+Compared to the split second that the actual work usually takes,
+this overhead can amount to several orders of magnitude.
+
+Avoid Puppet types that `mgmt` does not implement (yet).
+
+###Avoiding common warnings
+
+Many resource parameters in Puppet take default values. For the most part,
+the translator module just ignores them. However, there are cases in which
+Puppet will default to convenient behavior that `mgmt` cannot quite replicate.
+For example, translating a plain `file` resource will lead to a warning message:
+
+    $ puppet mgmtgraph print --code 'file { "/tmp/mgmt-test": }'
+    Warning: File[/tmp/mgmt-test] uses the 'puppet' file bucket, which mgmt cannot do. There will be no backup copies!
+
+The reason is that per default, Puppet assumes the following parameter value
+(among others)
+
+```puppet
+file { "/tmp/mgmt-test":
+  backup => 'puppet',
+}
+```
+
+To avoid this, specify the parameter explicitly:
+
+    $ puppet mgmtgraph print --code 'file { "/tmp/mgmt-test": backup => false }'
+
+This is tedious in a more complex manifest. A good simplification is the
+following [resource default](https://docs.puppet.com/puppet/latest/reference/lang_defaults.html)
+anywhere on the top scope of your manifest:
+
+```puppet
+File { backup => false }
+```
+
+If you encounter similar warnings from other types and/or parameters,
+use the same approach to silence them if possible.
+
+##Configuring Puppet
+
+Since `mgmt` uses an actual Puppet CLI behind the scenes, you might
+need to tweak some of Puppet's runtime options in order to make it
+do what you want. Reasons for this could be among the following:
+
+* You use the `--puppet agent` variant and need to configure
+`servername`, `certname` and other master/agent-related options.
+* You don't want runtime information to end up in the `vardir`
+that is used by your regular `puppet agent`.
+* You install specific Puppet modules for `mgmt` in a non-standard
+location.
+
+`mgmt` exposes only one Puppet option in order to allow you to
+control all of them, through its `--puppet-conf` option. It allows
+you to specify which `puppet.conf` file should be used during
+translation.
+
+    mgmt run --puppet /opt/my-manifest.pp --puppet-conf /etc/mgmt/puppet.conf
+
+Within this file, you can just specify any needed options in the
+`[main]` section:
+
+    [main]
+    server=mgmt-master.example.net
+    vardir=/var/lib/mgmt/puppet
+
+##Caveats
+
+Please see the [README](https://github.com/ffrank/puppet-mgmtgraph/blob/master/README.md)
+of the translator module for the current state of supported and unsupported
+language features.
+
+You should probably make sure to always use the latest release of
+both `ffrank-mgmtgraph` and `ffrank-yamlresource` (the latter is
+getting pulled in as a dependency of the former).

README.md

@@ -1,24 +1,47 @@
-# *mgmt*: This is: mgmt!
+# *mgmt*: next generation config management!
 
+[![mgmt!](art/mgmt.png)](art/)
+
+[![Go Report Card](https://goreportcard.com/badge/github.com/purpleidea/mgmt)](https://goreportcard.com/report/github.com/purpleidea/mgmt)
 [![Build Status](https://secure.travis-ci.org/purpleidea/mgmt.png?branch=master)](http://travis-ci.org/purpleidea/mgmt)
 [![Documentation](https://img.shields.io/docs/markdown.png)](DOCUMENTATION.md)
+[![GoDoc](https://godoc.org/github.com/purpleidea/mgmt?status.svg)](https://godoc.org/github.com/purpleidea/mgmt)
 [![IRC](https://img.shields.io/irc/%23mgmtconfig.png)](https://webchat.freenode.net/?channels=#mgmtconfig)
 [![Jenkins](https://img.shields.io/jenkins/status.png)](https://ci.centos.org/job/purpleidea-mgmt/)
 [![COPR](https://img.shields.io/copr/builds.png)](https://copr.fedoraproject.org/coprs/purpleidea/mgmt/)
+[![arch](https://img.shields.io/arch/aur.png)](https://aur.archlinux.org/packages/mgmt/)
 
 ## Community:
 Come join us on IRC in [#mgmtconfig](https://webchat.freenode.net/?channels=#mgmtconfig) on Freenode!
 You may like the [#mgmtconfig](https://twitter.com/hashtag/mgmtconfig) hashtag if you're on [Twitter](https://twitter.com/#!/purpleidea).
 
+## Status:
+Mgmt is a fairly new project.
+We're working towards being minimally useful for production environments.
+We aren't feature complete for what we'd consider a 1.x release yet.
+With your help you'll be able to influence our design and get us there sooner!
+
 ## Questions:
 Please join the [#mgmtconfig](https://webchat.freenode.net/?channels=#mgmtconfig) IRC community!
 If you have a well phrased question that might benefit others, consider asking it by sending a patch to the documentation [FAQ](https://github.com/purpleidea/mgmt/blob/master/DOCUMENTATION.md#usage-and-frequently-asked-questions) section. I'll merge your question, and a patch with the answer!
 
 ## Quick start:
-* Either get the golang dependencies on your own, or run `make deps` if you're comfortable with how we install them.
+* Make sure you have golang version 1.6 or greater installed.
+* If you do not have a GOPATH yet, create one and export it:
+```
+mkdir $HOME/gopath
+export GOPATH=$HOME/gopath
+```
+* You might also want to add the GOPATH to your `~/.bashrc` or `~/.profile`.
+* For more information you can read the [GOPATH documentation](https://golang.org/cmd/go/#hdr-GOPATH_environment_variable).
+* Next download the mgmt code base, and switch to that directory:
+```
+go get -u github.com/purpleidea/mgmt
+cd $GOPATH/src/github.com/purpleidea/mgmt
+```
+* Get the remaining golang deps with `go get ./...`, or run `make deps` if you're comfortable with how we install them.
 * Run `make build` to get a freshly built `mgmt` binary.
-* Run `cd $(mktemp --tmpdir -d tmp.XXX) && etcd` to get etcd running. The `mgmt` software will do this automatically for you in the future.
+* Run `time ./mgmt run --yaml examples/graph0.yaml --converged-timeout=5 --tmp-prefix` to try out a very simple example!
-* Run `time ./mgmt run --file examples/graph0.yaml --converged-timeout=1` to try out a very simple example!
 * To run continuously in the default mode of operation, omit the `--converged-timeout` option.
 * Have fun hacking on our future technology!
 
@@ -26,7 +49,7 @@ If you have a well phrased question that might benefit others, consider asking i
 Please look in the [examples/](examples/) folder for more examples!
 
 ## Documentation:
-Please see: [DOCUMENTATION.md](DOCUMENTATION.md) or [PDF](https://pdfdoc-purpleidea.rhcloud.com/pdf/https://github.com/purpleidea/mgmt/blob/master/DOCUMENTATION.md).
+Please see: the manually created [DOCUMENTATION.md](DOCUMENTATION.md) (also available as [PDF](https://pdfdoc-purpleidea.rhcloud.com/pdf/https://github.com/purpleidea/mgmt/blob/master/DOCUMENTATION.md)) and the automatically generated [GoDoc documentation](https://godoc.org/github.com/purpleidea/mgmt).
 
 ## Roadmap:
 Please see: [TODO.md](TODO.md) for a list of upcoming work and TODO items.
@@ -39,20 +62,22 @@ Bonus points if you provide a [shell](https://github.com/purpleidea/mgmt/tree/ma
 Feel free to read my article on [debugging golang programs](https://ttboj.wordpress.com/2016/02/15/debugging-golang-programs/).
 
 ## Dependencies:
-* golang 1.4 or higher (required, available in most distros)
+* golang 1.6 or higher (required, available in most distros)
 * golang libraries (required, available with `go get`)
-
+```
-        go get github.com/coreos/etcd/client
+go get github.com/coreos/etcd/client
-        go get gopkg.in/yaml.v2
+go get gopkg.in/yaml.v2
-        go get gopkg.in/fsnotify.v1
+go get gopkg.in/fsnotify.v1
-        go get github.com/codegangsta/cli
+go get github.com/urfave/cli
-        go get github.com/coreos/go-systemd/dbus
+go get github.com/coreos/go-systemd/dbus
-        go get github.com/coreos/go-systemd/util
+go get github.com/coreos/go-systemd/util
-
+go get github.com/coreos/pkg/capnslog
-* stringer (required for building), available as a package on some platforms, otherwise via `go get`
+go get github.com/rgbkrk/libvirt-go
-
+```
-        go get golang.org/x/tools/cmd/stringer
+* stringer (optional for building), available as a package on some platforms, otherwise via `go get`
-
+```
+go get golang.org/x/tools/cmd/stringer
+```
 * pandoc (optional, for building a pdf of the documentation)
 * graphviz (optional, for building a visual representation of the graph)
 
@@ -60,14 +85,25 @@ Feel free to read my article on [debugging golang programs](https://ttboj.wordpr
 We'd love to have your patches! Please send them by email, or as a pull request.
 
 ## On the web:
-* Introductory blog post: [https://ttboj.wordpress.com/2016/01/18/next-generation-configuration-mgmt/](https://ttboj.wordpress.com/2016/01/18/next-generation-configuration-mgmt/)
+* James Shubin; blog: [Next generation configuration mgmt](https://ttboj.wordpress.com/2016/01/18/next-generation-configuration-mgmt/)
-* Introductory recording from DevConf.cz 2016: [https://www.youtube.com/watch?v=GVhpPF0j-iE&html5=1](https://www.youtube.com/watch?v=GVhpPF0j-iE&html5=1)
+* James Shubin; video: [Introductory recording from DevConf.cz 2016](https://www.youtube.com/watch?v=GVhpPF0j-iE&html5=1)
-* Introductory recording from CfgMgmtCamp.eu 2016: [https://www.youtube.com/watch?v=fNeooSiIRnA&html5=1](https://www.youtube.com/watch?v=fNeooSiIRnA&html5=1)
+* James Shubin; video: [Introductory recording from CfgMgmtCamp.eu 2016](https://www.youtube.com/watch?v=fNeooSiIRnA&html5=1)
-* Julian Dunn at CfgMgmtCamp.eu 2016: [https://www.youtube.com/watch?v=kfF9IATUask&t=1949&html5=1](https://www.youtube.com/watch?v=kfF9IATUask&t=1949&html5=1)
+* Julian Dunn; video: [On mgmt at CfgMgmtCamp.eu 2016](https://www.youtube.com/watch?v=kfF9IATUask&t=1949&html5=1)
-* Walter Heck at CfgMgmtCamp.eu 2016: [http://www.slideshare.net/olindata/configuration-management-time-for-a-4th-generation/3](http://www.slideshare.net/olindata/configuration-management-time-for-a-4th-generation/3)
+* Walter Heck; slides: [On mgmt at CfgMgmtCamp.eu 2016](http://www.slideshare.net/olindata/configuration-management-time-for-a-4th-generation/3)
-* Marco Marongiu on mgmt: [http://syslog.me/2016/02/15/leap-or-die/](http://syslog.me/2016/02/15/leap-or-die/)
+* Marco Marongiu; blog: [On mgmt](http://syslog.me/2016/02/15/leap-or-die/)
-* Felix Frank on puppet to mgmt "transpiling" [https://ffrank.github.io/features/2016/02/18/from-catalog-to-mgmt/](https://ffrank.github.io/features/2016/02/18/from-catalog-to-mgmt/)
+* Felix Frank; blog: [From Catalog To Mgmt (on puppet to mgmt "transpiling")](https://ffrank.github.io/features/2016/02/18/from-catalog-to-mgmt/)
-* Blog post on automatic edges and the pkg resource: [https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/](https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/)
+* James Shubin; blog: [Automatic edges in mgmt (...and the pkg resource)](https://ttboj.wordpress.com/2016/03/14/automatic-edges-in-mgmt/)
+* James Shubin; blog: [Automatic grouping in mgmt](https://ttboj.wordpress.com/2016/03/30/automatic-grouping-in-mgmt/)
+* John Arundel; tweet: [“Puppet’s days are numbered.”](https://twitter.com/bitfield/status/732157519142002688)
+* Felix Frank; blog: [Puppet, Meet Mgmt (on puppet to mgmt internals)](https://ffrank.github.io/features/2016/06/12/puppet,-meet-mgmt/)
+* Felix Frank; blog: [Puppet Powered Mgmt (puppet to mgmt tl;dr)](https://ffrank.github.io/features/2016/06/19/puppet-powered-mgmt/)
+* James Shubin; blog: [Automatic clustering in mgmt](https://ttboj.wordpress.com/2016/06/20/automatic-clustering-in-mgmt/)
+* James Shubin; video: [Recording from CoreOSFest 2016](https://www.youtube.com/watch?v=KVmDCUA42wc&html5=1)
+* James Shubin; video: [Recording from DebConf16](http://meetings-archive.debian.net/pub/debian-meetings/2016/debconf16/Next_Generation_Config_Mgmt.webm) ([Slides](https://annex.debconf.org//debconf-share/debconf16/slides/15-next-generation-config-mgmt.pdf))
+* Felix Frank; blog: [Edging It All In (puppet and mgmt edges)](https://ffrank.github.io/features/2016/07/12/edging-it-all-in/)
+* Felix Frank; blog: [Translating All The Things (puppet to mgmt translation warnings)](https://ffrank.github.io/features/2016/08/19/translating-all-the-things/)
+* James Shubin; video: [Recording from systemd.conf 2016](https://www.youtube.com/watch?v=jB992Zb3nH0&html5=1)
+* James Shubin; blog: [Remote execution in mgmt](https://ttboj.wordpress.com/2016/10/07/remote-execution-in-mgmt/)
 
 ##
 

THANKS

@@ -9,10 +9,16 @@ Chris Wright - For encouraging me to continue work on my prototype.
 
 Daniel Riek - For supporting and sheltering this project from bureaucracy.
 
+Diego Ongaro - For good chats, particularly around distributed systems.
+
+Felix Frank - For taking a difficult problem and building an inspiring solution.
+
 Ira Cooper - For having an algorithmic design discussion with me.
 
 Jeff Darcy - For some algorithm recommendations, and NACKing my TopoSort idea!
 
 Red Hat, inc. - For paying my salary, thus financially supporting my hacking.
 
+Samuel Gélineau - For help with programming language theory and design.
+
 And many others...

TODO.md

@@ -3,27 +3,54 @@ If you're looking for something to do, look here!
 Let us know if you're working on one of the items.
 
 ## Package resource
-- [ ] base resource [bug](https://github.com/purpleidea/mgmt/issues/11)
+- [ ] getfiles support on debian [bug](https://github.com/hughsie/PackageKit/issues/118)
+- [ ] directory info on fedora [bug](https://github.com/hughsie/PackageKit/issues/117)
 - [ ] dnf blocker [bug](https://github.com/hughsie/PackageKit/issues/110)
 - [ ] install signal blocker [bug](https://github.com/hughsie/PackageKit/issues/109)
 
 ## File resource [bug](https://github.com/purpleidea/mgmt/issues/13) [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
-- [ ] ability to make/delete folders
+- [ ] chown/chmod support [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
-- [ ] recursive argument (can recursively watch/modify contents)
+- [ ] user/group support [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
-- [ ] force argument (can cause switch from file <-> folder)
+- [ ] recurse limit support [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
 - [ ] fanotify support [bug](https://github.com/go-fsnotify/fsnotify/issues/114)
 
+## Svc resource
+- [ ] base resource improvements
+
 ## Exec resource
 - [ ] base resource improvements
 
 ## Timer resource
-- [ ] base resource [bug](https://github.com/purpleidea/mgmt/issues/15) [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
 - [ ] reset on recompile
-- [ ] increment algorithm (linear, exponential, etc...)
+- [ ] increment algorithm (linear, exponential, etc...) [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+
+## User/Group resource
+- [ ] base resource [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+- [ ] automatic edges to file resource [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+
+## Virt (libvirt) resource
+- [ ] base resource [bug](https://github.com/purpleidea/mgmt/issues/25)
+
+## Net (systemd-networkd) resource
+- [ ] base resource
+
+## Nspawn (systemd-nspawn) resource
+- [ ] base resource [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+
+## Mount (systemd-mount) resource
+- [ ] base resource [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+
+## Cron (systemd-timer) resource
+- [ ] base resource [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
+
+## Http resource
+- [ ] base resource
 
 ## Etcd improvements
-- [ ] embedded etcd master
+- [ ] fix embedded etcd master race
-- [ ] capnslog fixes [bug](https://github.com/coreos/etcd/issues/4115)
+
+## Torrent/dht file transfer
+- [ ] base plumbing
 
 ## Language improvements
 - [ ] language design
@@ -35,9 +62,6 @@ Let us know if you're working on one of the items.
 
 ## Other
 - [ ] better error/retry handling
-- [ ] resource grouping
-- [ ] automatic dependency adding (eg: packagekit file dependencies)
-- [ ] mgmt systemd service file [bug](https://github.com/purpleidea/mgmt/issues/12) [:heart:](https://github.com/purpleidea/mgmt/labels/mgmtlove)
 - [ ] deb package target in Makefile
 - [ ] reproducible builds
 - [ ] add your suggestions!

art/.gitignore

@@ -0,0 +1,2 @@
+*.png
+misc/

art/mgmt_logo_default_symbol.svg

@@ -0,0 +1,94 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 120 107.1" style="enable-background:new 0 0 120 107.1;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<line class="st10" x1="29.2" y1="24.1" x2="52.1" y2="42.8"/>
+				<g>
+					<polygon class="st9" points="27.7,27.8 24.9,20.5 32.6,21.8 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st5" cx="16.1" cy="12.2" r="12.1"/>
+		<g>
+			<g>
+				<line class="st10" x1="52.1" y1="42.1" x2="74.1" y2="80"/>
+				<g>
+					<polygon class="st9" points="70.1,80.9 76.9,84.8 76.8,77.1 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st10" x1="69.4" y1="46.7" x2="95.8" y2="52.4"/>
+				<g>
+					<polygon class="st9" points="69.7,50.7 63.9,45.6 71.3,43.2 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st10" x1="52.1" y1="42.8" x2="71.9" y2="30.1"/>
+				<g>
+					<polygon class="st9" points="73.1,34 76.6,27.1 68.9,27.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st10" x1="16.8" y1="49.6" x2="34.8" y2="46.5"/>
+				<g>
+					<polygon class="st9" points="34.3,50.5 40.3,45.6 33,42.9 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st10" x1="92.3" y1="79.5" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st9" points="96.1,80.6 89.3,84.3 89.5,76.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st10" x1="97.3" y1="36.5" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st9" points="94.6,39.4 94.5,31.7 101.2,35.5 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st3" cx="52.1" cy="42.8" r="12.1"/>
+		<circle class="st4" cx="12.2" cy="50.8" r="12.1"/>
+		<circle class="st7" cx="87.5" cy="21.7" r="12.1"/>
+		<circle class="st8" cx="83.5" cy="95" r="12.1"/>
+		<circle class="st6" cx="107.8" cy="54.2" r="12.1"/>
+	</g>
+</g>
+</svg>

art/mgmt_logo_default_tall.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 168.3 131.6" style="enable-background:new 0 0 168.3 131.6;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st3" d="M4.7,105l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9V124h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6V124H9v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4V124H0v-19H4.7z"/>
+				<path class="st3" d="M26.4,113.9c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2L37,105h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8s0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V113.9z
+					 M31.4,115.2c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1V110c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					s-0.7,2.1-0.7,3.9V115.2z"/>
+				<path class="st3" d="M50.1,105l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9V124h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4s-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6V124h-5v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4s-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4V124h-5v-19H50.1z"/>
+				<path class="st3" d="M78.2,100.3v4.7h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1
+					l0,3.9c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5v-10.1h-2.2V105h2.2v-4.7H78.2z"/>
+				<path class="st4" d="M90.6,122.6c1.4,0,2.4-0.4,3.1-1.1c0.7-0.8,1.1-1.9,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6
+					c-1.1,1.1-2.6,1.7-4.3,1.7c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4v-2.3c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2
+					c1.9,0,3.4,0.6,4.5,1.8s1.7,2.9,1.7,5H95c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7s-1.2,2.9-1.2,5.2
+					v2.2c0,2.4,0.4,4.2,1.2,5.3S89,122.6,90.6,122.6z"/>
+				<path class="st4" d="M100.5,113.6c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4s1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3c-2.1,0-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V113.6z M102.5,115.5c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8s1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8s-1.3,2.9-1.4,5.1V115.5z"/>
+				<path class="st4" d="M121.1,105l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6V124h-1.9v-12.5
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2V124h-2v-19H121.1z"/>
+				<path class="st4" d="M138.2,124v-17.3h-2.6V105h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3l-0.1,1.8
+					c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3v2.3h3.7v1.8h-3.7V124H138.2z"/>
+				<path class="st4" d="M148,99.5c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1s-0.5,0.4-1,0.4
+					s-0.7-0.1-0.9-0.4S148,99.9,148,99.5z M150.2,124h-2v-19h2V124z"/>
+				<path class="st4" d="M155.3,113.6c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8v19.6c0,2.3-0.6,4-1.6,5.2
+					c-1.1,1.2-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V113.6z M157.2,115.4
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V115.4z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st11" x1="58.9" y1="20.1" x2="77.9" y2="35.6"/>
+					<g>
+						<polygon class="st9" points="57.6,23.2 55.3,17.1 61.7,18.2 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st5" cx="48" cy="10.1" r="10.1"/>
+			<g>
+				<g>
+					<line class="st11" x1="77.9" y1="35.1" x2="96.2" y2="66.7"/>
+					<g>
+						<polygon class="st9" points="93,67.5 98.6,70.7 98.6,64.2 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st11" x1="92.3" y1="38.9" x2="114.4" y2="43.6"/>
+					<g>
+						<polygon class="st9" points="92.6,42.3 87.8,38 93.9,36 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st11" x1="77.9" y1="35.6" x2="94.5" y2="25.1"/>
+					<g>
+						<polygon class="st9" points="95.4,28.3 98.4,22.6 91.9,22.9 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st11" x1="48.5" y1="41.3" x2="63.5" y2="38.7"/>
+					<g>
+						<polygon class="st9" points="63.1,42.1 68.1,38 62,35.7 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st11" x1="111.4" y1="66.3" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st9" points="114.6,67.2 109,70.2 109.1,63.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st11" x1="115.6" y1="30.4" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st9" points="113.3,32.8 113.3,26.4 118.9,29.5 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st3" cx="77.9" cy="35.6" r="10.1"/>
+			<circle class="st4" cx="44.6" cy="42.4" r="10.1"/>
+			<circle class="st7" cx="107.4" cy="18.1" r="10.1"/>
+			<circle class="st8" cx="104.1" cy="79.1" r="10.1"/>
+			<circle class="st6" cx="124.4" cy="45.2" r="10.1"/>
+		</g>
+	</g>
+</g>
+</svg>

art/mgmt_logo_default_wide.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 260.4 71.4" style="enable-background:new 0 0 260.4 71.4;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st3" d="M96.7,25.7l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5V32.6c0-1.1-0.2-1.9-0.5-2.4s-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5V32.6
+					c0-1.1-0.1-1.9-0.4-2.4s-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H96.7z"/>
+				<path class="st3" d="M118.5,34.6c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2l0.2-1.7h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8s0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V34.6z
+					 M123.5,35.9c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1v-9.1c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					s-0.7,2.1-0.7,3.9V35.9z"/>
+				<path class="st3" d="M142.2,25.7l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5V32.6c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5V32.6
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H142.2z"/>
+				<path class="st3" d="M170.3,21v4.7h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1l0,3.9
+					c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5V29.4h-2.2v-3.7h2.2V21H170.3z"/>
+				<path class="st4" d="M182.7,43.2c1.4,0,2.4-0.4,3.1-1.1s1.1-1.8,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6s-2.6,1.7-4.3,1.7
+					c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4v-2.3c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2c1.9,0,3.4,0.6,4.5,1.8s1.7,2.9,1.7,5H187
+					c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7c-0.8,1.1-1.2,2.9-1.2,5.2v2.2c0,2.4,0.4,4.2,1.2,5.3
+					S181,43.2,182.7,43.2z"/>
+				<path class="st4" d="M192.6,34.2c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4c1.2,1.6,1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3s-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V34.2z M194.6,36.2c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8c0.8-1.2,1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8c-0.9,1.2-1.3,2.9-1.4,5.1V36.2z"/>
+				<path class="st4" d="M213.2,25.7l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6v12.7h-1.9V32.2
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2v13.2h-2v-19H213.2z"/>
+				<path class="st4" d="M230.3,44.7V27.4h-2.6v-1.8h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3l-0.1,1.8
+					c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3v2.3h3.7v1.8h-3.7v17.3H230.3z"/>
+				<path class="st4" d="M240.1,20.2c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1
+					s-0.5,0.4-1,0.4s-0.7-0.1-0.9-0.4S240.1,20.6,240.1,20.2z M242.3,44.7h-2v-19h2V44.7z"/>
+				<path class="st4" d="M247.4,34.3c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8v19.6c0,2.3-0.6,4-1.6,5.2
+					s-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V34.3z M249.3,36.1
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V36.1z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st12" x1="19.5" y1="16" x2="34.7" y2="28.5"/>
+					<g>
+						<polygon class="st9" points="18.4,18.5 16.6,13.7 21.7,14.5 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st5" cx="10.8" cy="8.1" r="8.1"/>
+			<g>
+				<g>
+					<line class="st12" x1="34.7" y1="28" x2="49.4" y2="53.3"/>
+					<g>
+						<polygon class="st9" points="46.8,54 51.2,56.5 51.2,51.4 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st12" x1="46.2" y1="31.1" x2="63.9" y2="34.9"/>
+					<g>
+						<polygon class="st9" points="46.4,33.8 42.6,30.4 47.5,28.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st12" x1="34.7" y1="28.5" x2="47.9" y2="20.1"/>
+					<g>
+						<polygon class="st9" points="48.7,22.7 51.1,18.1 45.9,18.3 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st12" x1="11.2" y1="33.1" x2="23.2" y2="31"/>
+					<g>
+						<polygon class="st9" points="22.9,33.7 26.8,30.4 22,28.6 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st12" x1="61.5" y1="53" x2="71.9" y2="36.1"/>
+					<g>
+						<polygon class="st9" points="64.1,53.7 59.5,56.2 59.7,51 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st12" x1="64.2" y1="24.2" x2="70.7" y2="33.9"/>
+					<g>
+						<polygon class="st9" points="62.5,26.3 62.2,21.1 66.8,23.4 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st3" cx="34.7" cy="28.5" r="8.1"/>
+			<circle class="st4" cx="8.1" cy="33.9" r="8.1"/>
+			<circle class="st7" cx="58.3" cy="14.5" r="8.1"/>
+			<circle class="st8" cx="55.7" cy="63.3" r="8.1"/>
+			<circle class="st6" cx="71.9" cy="36.1" r="8.1"/>
+		</g>
+	</g>
+</g>
+</svg>

art/mgmt_logo_reversed_symbol.svg

@@ -0,0 +1,94 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 120 107.1" style="enable-background:new 0 0 120 107.1;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<line class="st13" x1="29.2" y1="24.1" x2="52.1" y2="42.8"/>
+				<g>
+					<polygon class="st14" points="27.7,27.8 24.9,20.5 32.6,21.8 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st3" cx="16.1" cy="12.2" r="12.1"/>
+		<g>
+			<g>
+				<line class="st13" x1="52.1" y1="42.1" x2="74.1" y2="80"/>
+				<g>
+					<polygon class="st14" points="70.1,80.9 76.9,84.8 76.8,77.1 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st13" x1="69.4" y1="46.7" x2="95.8" y2="52.4"/>
+				<g>
+					<polygon class="st14" points="69.7,50.7 63.9,45.6 71.3,43.2 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st13" x1="52.1" y1="42.8" x2="71.9" y2="30.1"/>
+				<g>
+					<polygon class="st14" points="73.1,34 76.6,27.1 68.9,27.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st13" x1="16.8" y1="49.6" x2="34.8" y2="46.5"/>
+				<g>
+					<polygon class="st14" points="34.3,50.5 40.3,45.6 33,42.9 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st13" x1="92.3" y1="79.5" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st14" points="96.1,80.6 89.3,84.3 89.5,76.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st13" x1="97.2" y1="36.3" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st14" points="94.4,39.3 94.3,31.5 101.1,35.3 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st5" cx="52.1" cy="42.8" r="12.1"/>
+		<circle class="st4" cx="12.2" cy="50.8" r="12.1"/>
+		<circle class="st7" cx="87.5" cy="21.7" r="12.1"/>
+		<circle class="st8" cx="83.5" cy="95" r="12.1"/>
+		<circle class="st6" cx="107.8" cy="54.2" r="12.1"/>
+	</g>
+</g>
+</svg>

art/mgmt_logo_reversed_tall.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 168.3 133" style="enable-background:new 0 0 168.3 133;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st0" d="M4.7,106.4l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9H9v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8H0v-19H4.7z"/>
+				<path class="st0" d="M26.4,115.3c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2l0.2-1.7h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8c0.5-0.5,0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V115.3z
+					 M31.4,116.6c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1v-9.1c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					s-0.7,2.1-0.7,3.9V116.6z"/>
+				<path class="st0" d="M50.1,106.4l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4s-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4s-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H50.1z"/>
+				<path class="st0" d="M78.2,101.7v4.7h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1
+					l0,3.9c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5v-10.1h-2.2v-3.7h2.2v-4.7H78.2z"/>
+				<path class="st4" d="M90.6,124c1.4,0,2.4-0.4,3.1-1.1c0.7-0.8,1.1-1.9,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6
+					c-1.1,1.1-2.6,1.7-4.3,1.7c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4v-2.3c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2
+					c1.9,0,3.4,0.6,4.5,1.8s1.7,2.9,1.7,5H95c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7s-1.2,2.9-1.2,5.2
+					v2.2c0,2.4,0.4,4.2,1.2,5.3S89,124,90.6,124z"/>
+				<path class="st4" d="M100.5,115c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4s1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3c-2.1,0-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V115z M102.5,116.9c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8s1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8s-1.3,2.9-1.4,5.1V116.9z"/>
+				<path class="st4" d="M121.1,106.4l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6v12.7h-1.9v-12.5
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2v13.2h-2v-19H121.1z"/>
+				<path class="st4" d="M138.2,125.4v-17.3h-2.6v-1.8h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3
+					l-0.1,1.8c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3v2.3h3.7v1.8h-3.7v17.3H138.2z"/>
+				<path class="st4" d="M148,100.9c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1
+					s-0.5,0.4-1,0.4s-0.7-0.1-0.9-0.4S148,101.3,148,100.9z M150.2,125.4h-2v-19h2V125.4z"/>
+				<path class="st4" d="M155.3,115c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8V126c0,2.3-0.6,4-1.6,5.2
+					c-1.1,1.2-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V115z M157.2,116.8
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V116.8z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st16" x1="58.9" y1="20.1" x2="77.9" y2="35.6"/>
+					<g>
+						<polygon class="st14" points="57.6,23.2 55.3,17.1 61.7,18.2 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st3" cx="48" cy="10.1" r="10.1"/>
+			<g>
+				<g>
+					<line class="st16" x1="77.9" y1="35.1" x2="96.2" y2="66.7"/>
+					<g>
+						<polygon class="st14" points="93,67.5 98.6,70.7 98.6,64.2 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st16" x1="92.3" y1="38.9" x2="114.4" y2="43.6"/>
+					<g>
+						<polygon class="st14" points="92.6,42.3 87.8,38 93.9,36 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st16" x1="77.9" y1="35.6" x2="94.5" y2="25.1"/>
+					<g>
+						<polygon class="st14" points="95.4,28.3 98.4,22.6 91.9,22.9 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st16" x1="48.5" y1="41.3" x2="63.5" y2="38.7"/>
+					<g>
+						<polygon class="st14" points="63.1,42.1 68.1,38 62,35.7 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st16" x1="111.4" y1="66.3" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st14" points="114.6,67.2 109,70.2 109.1,63.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st16" x1="115.5" y1="30.5" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st14" points="113.2,32.9 113.1,26.5 118.8,29.6 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st5" cx="77.9" cy="35.6" r="10.1"/>
+			<circle class="st4" cx="44.6" cy="42.4" r="10.1"/>
+			<circle class="st7" cx="107.4" cy="18.1" r="10.1"/>
+			<circle class="st8" cx="104.1" cy="79.1" r="10.1"/>
+			<circle class="st6" cx="124.4" cy="45.2" r="10.1"/>
+		</g>
+	</g>
+</g>
+</svg>

art/mgmt_logo_reversed_wide.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 260.4 71.4" style="enable-background:new 0 0 260.4 71.4;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st0" d="M96.7,27.6l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5V34.4c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5V34.5
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H96.7z"/>
+				<path class="st0" d="M118.5,36.5c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2l0.2-1.7h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8s0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V36.5z
+					 M123.5,37.7c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1v-9.1c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					c-0.5,0.8-0.7,2.1-0.7,3.9V37.7z"/>
+				<path class="st0" d="M142.2,27.6l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5V34.4c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5V34.5
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H142.2z"/>
+				<path class="st0" d="M170.3,22.9v4.7h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1
+					l0,3.9c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5V31.3h-2.2v-3.7h2.2v-4.7H170.3z"/>
+				<path class="st4" d="M182.7,45.1c1.4,0,2.4-0.4,3.1-1.1c0.7-0.8,1.1-1.9,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6s-2.6,1.7-4.3,1.7
+					c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4V36c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2c1.9,0,3.4,0.6,4.5,1.8s1.7,2.9,1.7,5H187
+					c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7s-1.2,2.9-1.2,5.2v2.2c0,2.4,0.4,4.2,1.2,5.3
+					S181,45.1,182.7,45.1z"/>
+				<path class="st4" d="M192.6,36.1c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4c1.2,1.6,1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3s-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V36.1z M194.6,38.1c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8c0.8-1.2,1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8c-0.9,1.2-1.3,2.9-1.4,5.1V38.1z"/>
+				<path class="st4" d="M213.2,27.6l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6v12.7h-1.9V34.1
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2v13.2h-2v-19H213.2z"/>
+				<path class="st4" d="M230.3,46.6V29.3h-2.6v-1.8h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3l-0.1,1.8
+					c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3v2.3h3.7v1.8h-3.7v17.3H230.3z"/>
+				<path class="st4" d="M240.1,22.1c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1
+					s-0.5,0.4-1,0.4s-0.7-0.1-0.9-0.4S240.1,22.5,240.1,22.1z M242.3,46.6h-2v-19h2V46.6z"/>
+				<path class="st4" d="M247.4,36.2c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8v19.6c0,2.3-0.6,4-1.6,5.2
+					s-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V36.2z M249.3,38
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V38z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st15" x1="19.5" y1="16" x2="34.7" y2="28.5"/>
+					<g>
+						<polygon class="st14" points="18.4,18.5 16.6,13.7 21.7,14.5 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st3" cx="10.8" cy="8.1" r="8.1"/>
+			<g>
+				<g>
+					<line class="st15" x1="34.7" y1="28" x2="49.4" y2="53.3"/>
+					<g>
+						<polygon class="st14" points="46.8,54 51.2,56.5 51.2,51.4 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st15" x1="46.2" y1="31.1" x2="63.9" y2="34.9"/>
+					<g>
+						<polygon class="st14" points="46.4,33.8 42.6,30.4 47.5,28.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st15" x1="34.7" y1="28.5" x2="47.9" y2="20.1"/>
+					<g>
+						<polygon class="st14" points="48.7,22.7 51.1,18.1 45.9,18.3 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st15" x1="11.2" y1="33.1" x2="23.2" y2="31"/>
+					<g>
+						<polygon class="st14" points="22.9,33.7 26.8,30.4 22,28.6 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st15" x1="61.5" y1="53" x2="71.9" y2="36.1"/>
+					<g>
+						<polygon class="st14" points="64.1,53.7 59.5,56.2 59.7,51 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st15" x1="64.8" y1="24.4" x2="71.9" y2="36.1"/>
+					<g>
+						<polygon class="st14" points="63,26.4 62.9,21.2 67.4,23.7 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st5" cx="34.7" cy="28.5" r="8.1"/>
+			<circle class="st4" cx="8.1" cy="33.9" r="8.1"/>
+			<circle class="st7" cx="58.3" cy="14.5" r="8.1"/>
+			<circle class="st8" cx="55.7" cy="63.3" r="8.1"/>
+			<circle class="st6" cx="71.9" cy="36.1" r="8.1"/>
+		</g>
+	</g>
+</g>
+</svg>

art/mgmt_logo_white_symbol.svg

@@ -0,0 +1,94 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 120 107.1" style="enable-background:new 0 0 120 107.1;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<line class="st17" x1="29.2" y1="24.1" x2="52.1" y2="42.8"/>
+				<g>
+					<polygon class="st0" points="27.7,27.8 24.9,20.5 32.6,21.8 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st0" cx="16.1" cy="12.2" r="12.1"/>
+		<g>
+			<g>
+				<line class="st17" x1="52.1" y1="42.1" x2="74.1" y2="80"/>
+				<g>
+					<polygon class="st0" points="70.1,80.9 76.9,84.8 76.8,77.1 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st17" x1="69.4" y1="46.7" x2="95.8" y2="52.4"/>
+				<g>
+					<polygon class="st0" points="69.7,50.7 63.9,45.6 71.3,43.2 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st17" x1="52.1" y1="42.8" x2="71.9" y2="30.1"/>
+				<g>
+					<polygon class="st0" points="73.1,34 76.6,27.1 68.9,27.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st17" x1="16.8" y1="49.6" x2="34.8" y2="46.5"/>
+				<g>
+					<polygon class="st0" points="34.3,50.5 40.3,45.6 33,42.9 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st17" x1="92.3" y1="79.5" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st0" points="96.1,80.6 89.3,84.3 89.5,76.5 					"/>
+				</g>
+			</g>
+		</g>
+		<g>
+			<g>
+				<line class="st17" x1="97.2" y1="36.6" x2="107.8" y2="54.2"/>
+				<g>
+					<polygon class="st0" points="94.4,39.5 94.3,31.8 101.1,35.5 					"/>
+				</g>
+			</g>
+		</g>
+		<circle class="st0" cx="52.1" cy="42.8" r="12.1"/>
+		<circle class="st0" cx="12.2" cy="50.8" r="12.1"/>
+		<circle class="st0" cx="87.5" cy="21.7" r="12.1"/>
+		<circle class="st0" cx="83.5" cy="95" r="12.1"/>
+		<circle class="st0" cx="107.8" cy="54.2" r="12.1"/>
+	</g>
+</g>
+</svg>

art/mgmt_logo_white_tall.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 168.3 131.4" style="enable-background:new 0 0 168.3 131.4;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st0" d="M4.7,104.8l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9H9v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8H0v-19H4.7z"/>
+				<path class="st0" d="M26.4,113.8c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2l0.2-1.7h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8s0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V113.8z M31.4,115
+					c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1v-9.1c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					s-0.7,2.1-0.7,3.9V115z"/>
+				<path class="st0" d="M50.1,104.8l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9v12.5h-5v-12.1c0-1.1-0.2-1.9-0.5-2.4s-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6v12.9h-5v-12.1
+					c0-1.1-0.1-1.9-0.4-2.4s-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4v13.8h-5v-19H50.1z"/>
+				<path class="st0" d="M78.2,100.2v4.7h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1
+					l0,3.9c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5v-10.1h-2.2v-3.7h2.2v-4.7H78.2z"/>
+				<path class="st0" d="M90.6,122.4c1.4,0,2.4-0.4,3.1-1.1c0.7-0.8,1.1-1.8,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6
+					c-1.1,1.1-2.6,1.7-4.3,1.7c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4v-2.3c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2
+					c1.9,0,3.4,0.6,4.5,1.8s1.7,2.9,1.7,5H95c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7
+					c-0.8,1.1-1.2,2.9-1.2,5.2v2.2c0,2.4,0.4,4.2,1.2,5.3S89,122.4,90.6,122.4z"/>
+				<path class="st0" d="M100.5,113.4c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4s1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3c-2.1,0-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V113.4z M102.5,115.3c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8c0.8-1.2,1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8c-0.9,1.2-1.3,2.9-1.4,5.1V115.3z"/>
+				<path class="st0" d="M121.1,104.8l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6v12.7h-1.9v-12.5
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2v13.2h-2v-19H121.1z"/>
+				<path class="st0" d="M138.2,123.8v-17.3h-2.6v-1.8h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3
+					l-0.1,1.8c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3v2.3h3.7v1.8h-3.7v17.3H138.2z"/>
+				<path class="st0" d="M148,99.3c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1s-0.5,0.4-1,0.4
+					s-0.7-0.1-0.9-0.4S148,99.7,148,99.3z M150.2,123.8h-2v-19h2V123.8z"/>
+				<path class="st0" d="M155.3,113.5c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8v19.6c0,2.3-0.6,4-1.6,5.2
+					c-1.1,1.2-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V113.5z M157.2,115.2
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V115.2z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st19" x1="58.9" y1="20.1" x2="77.9" y2="35.6"/>
+					<g>
+						<polygon class="st0" points="57.6,23.2 55.3,17.1 61.7,18.2 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st0" cx="48" cy="10.1" r="10.1"/>
+			<g>
+				<g>
+					<line class="st19" x1="77.9" y1="35.1" x2="96.2" y2="66.7"/>
+					<g>
+						<polygon class="st0" points="93,67.5 98.6,70.7 98.6,64.2 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st19" x1="92.3" y1="38.9" x2="114.4" y2="43.6"/>
+					<g>
+						<polygon class="st0" points="92.6,42.3 87.8,38 93.9,36 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st19" x1="77.9" y1="35.6" x2="94.5" y2="25.1"/>
+					<g>
+						<polygon class="st0" points="95.4,28.3 98.4,22.6 91.9,22.9 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st19" x1="48.5" y1="41.3" x2="63.5" y2="38.7"/>
+					<g>
+						<polygon class="st0" points="63.1,42.1 68.1,38 62,35.7 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st19" x1="111.4" y1="66.3" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st0" points="114.6,67.2 109,70.2 109.1,63.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st19" x1="115.4" y1="30.4" x2="124.4" y2="45.2"/>
+					<g>
+						<polygon class="st0" points="113.2,32.8 113,26.4 118.7,29.5 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st0" cx="77.9" cy="35.6" r="10.1"/>
+			<circle class="st0" cx="44.6" cy="42.4" r="10.1"/>
+			<circle class="st0" cx="107.4" cy="18.1" r="10.1"/>
+			<circle class="st0" cx="104.1" cy="79.1" r="10.1"/>
+			<circle class="st0" cx="124.4" cy="45.2" r="10.1"/>
+		</g>
+	</g>
+</g>
+</svg>

art/mgmt_logo_white_wide.svg

@@ -0,0 +1,132 @@
+<?xml version="1.0" encoding="utf-8"?>
+<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 260.4 71.4" style="enable-background:new 0 0 260.4 71.4;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#FFFFFF;}
+	.st1{fill:#E22434;}
+	.st2{display:none;}
+	.st3{fill:#1B3663;}
+	.st4{fill:#00B1D1;}
+	.st5{fill:#BFE6EF;}
+	.st6{fill:#69CBE0;}
+	.st7{fill:#0080BD;}
+	.st8{fill:#005DAB;}
+	.st9{fill:#183660;}
+	.st10{fill:none;stroke:#183660;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st11{fill:none;stroke:#183660;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st12{fill:none;stroke:#183660;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st13{fill:none;stroke:#C0E6EF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st14{fill:#C0E6EF;}
+	.st15{fill:none;stroke:#C0E6EF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st16{fill:none;stroke:#C0E6EF;stroke-width:1.6201;stroke-miterlimit:10;}
+	.st17{fill:none;stroke:#FFFFFF;stroke-width:1.9441;stroke-miterlimit:10;}
+	.st18{fill:none;stroke:#FFFFFF;stroke-width:1.2961;stroke-miterlimit:10;}
+	.st19{fill:none;stroke:#FFFFFF;stroke-width:1.6201;stroke-miterlimit:10;}
+</style>
+<g id="Layer_2">
+</g>
+<g id="Layer_1">
+	<g>
+		<g>
+			<g>
+				<path class="st0" d="M96.7,26l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9V45h-5V32.9c0-1.1-0.2-1.9-0.5-2.4s-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6V45h-5V32.9
+					c0-1.1-0.1-1.9-0.4-2.4s-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4V45h-5V26H96.7z"/>
+				<path class="st0" d="M118.5,34.9c0-3.1,0.6-5.4,1.7-7s2.7-2.3,4.7-2.3c1.7,0,3.1,0.7,4,2l0.2-1.7h4.5v19c0,2.4-0.7,4.3-2,5.6
+					c-1.4,1.3-3.3,1.9-5.9,1.9c-1,0-2.1-0.2-3.3-0.6s-2-0.9-2.6-1.6l1.7-3.4c0.5,0.5,1.1,0.9,1.8,1.2c0.7,0.3,1.5,0.5,2.1,0.5
+					c1.1,0,1.9-0.3,2.4-0.8s0.7-1.4,0.7-2.6v-1.6c-0.9,1.2-2.2,1.9-3.7,1.9c-2,0-3.6-0.8-4.7-2.4s-1.7-3.8-1.7-6.7V34.9z
+					 M123.5,36.2c0,1.8,0.2,3,0.7,3.8s1.2,1.2,2.2,1.2c1,0,1.8-0.4,2.3-1.1V31c-0.5-0.8-1.3-1.2-2.2-1.2c-1,0-1.7,0.4-2.2,1.2
+					s-0.7,2.1-0.7,3.9V36.2z"/>
+				<path class="st0" d="M142.2,26l0.1,1.8c1.1-1.4,2.6-2.1,4.4-2.1c1.9,0,3.2,0.9,4,2.6c1.1-1.7,2.6-2.6,4.7-2.6
+					c3.3,0,5,2.3,5.1,6.9V45h-5V32.9c0-1.1-0.2-1.9-0.5-2.4c-0.3-0.5-0.8-0.7-1.5-0.7c-0.9,0-1.6,0.6-2.1,1.7l0,0.6V45h-5V32.9
+					c0-1.1-0.1-1.9-0.4-2.4c-0.3-0.5-0.8-0.7-1.6-0.7c-0.9,0-1.5,0.5-2,1.4V45h-5V26H142.2z"/>
+				<path class="st0" d="M170.3,21.3V26h2.5v3.7h-2.5v9.5c0,0.8,0.1,1.3,0.3,1.5c0.2,0.3,0.6,0.4,1.2,0.4c0.5,0,0.9,0,1.2-0.1l0,3.9
+					c-0.8,0.3-1.8,0.5-2.7,0.5c-3.2,0-4.8-1.8-4.9-5.5V29.7h-2.2V26h2.2v-4.7H170.3z"/>
+				<path class="st0" d="M182.7,43.5c1.4,0,2.4-0.4,3.1-1.1c0.7-0.8,1.1-1.9,1.2-3.3h1.9c-0.1,1.9-0.7,3.5-1.9,4.6s-2.6,1.7-4.3,1.7
+					c-2.3,0-4-0.7-5.1-2.2s-1.7-3.6-1.8-6.4v-2.3c0-2.9,0.6-5.1,1.7-6.6s2.9-2.2,5.1-2.2c1.9,0,3.4,0.6,4.5,1.8s1.7,2.8,1.7,5H187
+					c-0.1-1.6-0.5-2.8-1.2-3.6s-1.8-1.3-3.1-1.3c-1.7,0-2.9,0.6-3.7,1.7c-0.8,1.1-1.2,2.9-1.2,5.2v2.2c0,2.4,0.4,4.2,1.2,5.3
+					C179.8,43,181,43.5,182.7,43.5z"/>
+				<path class="st0" d="M192.6,34.5c0-2.7,0.6-4.9,1.9-6.5s2.9-2.4,5.1-2.4c2.2,0,3.9,0.8,5.1,2.4c1.2,1.6,1.9,3.7,1.9,6.5v2
+					c0,2.8-0.6,5-1.9,6.5s-2.9,2.3-5.1,2.3s-3.8-0.8-5-2.3s-1.9-3.6-1.9-6.3V34.5z M194.6,36.5c0,2.2,0.4,3.9,1.3,5.2
+					c0.9,1.3,2.1,1.9,3.7,1.9c1.6,0,2.8-0.6,3.7-1.8c0.8-1.2,1.3-2.9,1.3-5.2v-2c0-2.2-0.4-3.9-1.3-5.2c-0.9-1.3-2.1-1.9-3.7-1.9
+					c-1.5,0-2.7,0.6-3.6,1.8c-0.9,1.2-1.3,2.9-1.4,5.1V36.5z"/>
+				<path class="st0" d="M213.2,26l0.1,3c0.6-1,1.3-1.9,2.2-2.5s1.9-0.9,3-0.9c3.3,0,5,2.2,5.1,6.6V45h-1.9V32.5
+					c0-1.7-0.3-3-0.9-3.8c-0.6-0.8-1.5-1.2-2.8-1.2c-1,0-1.9,0.4-2.8,1.2s-1.4,1.8-1.9,3.2V45h-2V26H213.2z"/>
+				<path class="st0" d="M230.3,45V27.7h-2.6V26h2.6v-2.5c0-1.9,0.5-3.3,1.3-4.3s2-1.5,3.5-1.5c0.7,0,1.3,0.1,1.9,0.3l-0.1,1.8
+					c-0.5-0.1-1-0.2-1.6-0.2c-0.9,0-1.7,0.4-2.2,1.1s-0.8,1.7-0.8,3V26h3.7v1.8h-3.7V45H230.3z"/>
+				<path class="st0" d="M240.1,20.5c0-0.4,0.1-0.7,0.3-1s0.5-0.4,0.9-0.4s0.7,0.1,1,0.4s0.3,0.6,0.3,1s-0.1,0.7-0.3,1
+					s-0.5,0.4-1,0.4s-0.7-0.1-0.9-0.4S240.1,20.9,240.1,20.5z M242.3,45h-2V26h2V45z"/>
+				<path class="st0" d="M247.4,34.6c0-3,0.5-5.2,1.5-6.7s2.6-2.2,4.6-2.2c2.2,0,3.8,1,4.9,3l0.1-2.7h1.8v19.6c0,2.3-0.6,4-1.6,5.2
+					s-2.7,1.8-4.7,1.8c-1,0-2.1-0.3-3.2-0.8s-1.9-1.1-2.4-1.9l0.9-1.4c1.3,1.5,2.8,2.2,4.5,2.2c1.6,0,2.7-0.4,3.5-1.3
+					c0.7-0.8,1.1-2.1,1.1-3.8v-3c-1.1,1.8-2.7,2.7-4.9,2.7c-2,0-3.5-0.7-4.5-2.2s-1.6-3.6-1.6-6.5V34.6z M249.3,36.4
+					c0,2.4,0.4,4.2,1.1,5.4s1.9,1.7,3.5,1.7c2.1,0,3.6-1,4.5-3v-9.7c-0.9-2.2-2.4-3.3-4.4-3.3c-1.6,0-2.8,0.6-3.5,1.7
+					s-1.1,2.9-1.1,5.3V36.4z"/>
+			</g>
+		</g>
+		<g>
+			<g>
+				<g>
+					<line class="st18" x1="19.5" y1="16" x2="34.7" y2="28.5"/>
+					<g>
+						<polygon class="st0" points="18.4,18.5 16.6,13.7 21.7,14.5 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st0" cx="10.8" cy="8.1" r="8.1"/>
+			<g>
+				<g>
+					<line class="st18" x1="34.7" y1="28" x2="49.4" y2="53.3"/>
+					<g>
+						<polygon class="st0" points="46.8,54 51.2,56.5 51.2,51.4 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st18" x1="46.2" y1="31.1" x2="63.9" y2="34.9"/>
+					<g>
+						<polygon class="st0" points="46.4,33.8 42.6,30.4 47.5,28.8 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st18" x1="34.7" y1="28.5" x2="47.9" y2="20.1"/>
+					<g>
+						<polygon class="st0" points="48.7,22.7 51.1,18.1 45.9,18.3 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st18" x1="11.2" y1="33.1" x2="23.2" y2="31"/>
+					<g>
+						<polygon class="st0" points="22.9,33.7 26.8,30.4 22,28.6 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st18" x1="61.5" y1="53" x2="71.9" y2="36.1"/>
+					<g>
+						<polygon class="st0" points="64.1,53.7 59.5,56.2 59.7,51 						"/>
+					</g>
+				</g>
+			</g>
+			<g>
+				<g>
+					<line class="st18" x1="64.7" y1="24.4" x2="71.9" y2="36.6"/>
+					<g>
+						<polygon class="st0" points="62.9,26.4 62.9,21.2 67.3,23.7 						"/>
+					</g>
+				</g>
+			</g>
+			<circle class="st0" cx="34.7" cy="28.5" r="8.1"/>
+			<circle class="st0" cx="8.1" cy="33.9" r="8.1"/>
+			<circle class="st0" cx="58.3" cy="14.5" r="8.1"/>
+			<circle class="st0" cx="55.7" cy="63.3" r="8.1"/>
+			<circle class="st0" cx="71.9" cy="36.1" r="8.1"/>
+		</g>
+	</g>
+</g>
+</svg>

config.go

@@ -1,535 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-import (
-	"errors"
-	"fmt"
-	"gopkg.in/yaml.v2"
-	"io/ioutil"
-	"log"
-	"reflect"
-	"strings"
-)
-
-type collectorResConfig struct {
-	Kind    string `yaml:"kind"`
-	Pattern string `yaml:"pattern"` // XXX: Not Implemented
-}
-
-type vertexConfig struct {
-	Kind string `yaml:"kind"`
-	Name string `yaml:"name"`
-}
-
-type edgeConfig struct {
-	Name string       `yaml:"name"`
-	From vertexConfig `yaml:"from"`
-	To   vertexConfig `yaml:"to"`
-}
-
-type GraphConfig struct {
-	Graph     string `yaml:"graph"`
-	Resources struct {
-		Noop []*NoopRes `yaml:"noop"`
-		Pkg  []*PkgRes  `yaml:"pkg"`
-		File []*FileRes `yaml:"file"`
-		Svc  []*SvcRes  `yaml:"svc"`
-		Exec []*ExecRes `yaml:"exec"`
-	} `yaml:"resources"`
-	Collector []collectorResConfig `yaml:"collect"`
-	Edges     []edgeConfig         `yaml:"edges"`
-	Comment   string               `yaml:"comment"`
-}
-
-func (c *GraphConfig) Parse(data []byte) error {
-	if err := yaml.Unmarshal(data, c); err != nil {
-		return err
-	}
-	if c.Graph == "" {
-		return errors.New("Graph config: invalid `graph`")
-	}
-	return nil
-}
-
-func ParseConfigFromFile(filename string) *GraphConfig {
-	data, err := ioutil.ReadFile(filename)
-	if err != nil {
-		log.Printf("Error: Config: ParseConfigFromFile: File: %v", err)
-		return nil
-	}
-
-	var config GraphConfig
-	if err := config.Parse(data); err != nil {
-		log.Printf("Error: Config: ParseConfigFromFile: Parse: %v", err)
-		return nil
-	}
-
-	return &config
-}
-
-// NewGraphFromConfig returns a new graph from existing input, such as from the
-// existing graph, and a GraphConfig struct.
-func (g *Graph) NewGraphFromConfig(config *GraphConfig, etcdO *EtcdWObject, hostname string) (*Graph, error) {
-
-	var graph *Graph // new graph to return
-	if g == nil {    // FIXME: how can we check for an empty graph?
-		graph = NewGraph("Graph") // give graph a default name
-	} else {
-		graph = g.Copy() // same vertices, since they're pointers!
-	}
-
-	var lookup = make(map[string]map[string]*Vertex)
-
-	//log.Printf("%+v", config) // debug
-
-	// TODO: if defined (somehow)...
-	graph.SetName(config.Graph) // set graph name
-
-	var keep []*Vertex // list of vertex which are the same in new graph
-
-	// use reflection to avoid duplicating code... better options welcome!
-	value := reflect.Indirect(reflect.ValueOf(config.Resources))
-	vtype := value.Type()
-	for i := 0; i < vtype.NumField(); i++ { // number of fields in struct
-		name := vtype.Field(i).Name // string of field name
-		field := value.FieldByName(name)
-		iface := field.Interface() // interface type of value
-		slice := reflect.ValueOf(iface)
-		// XXX: should we just drop these everywhere and have the kind strings be all lowercase?
-		kind := FirstToUpper(name)
-		if DEBUG {
-			log.Printf("Config: Processing: %v...", kind)
-		}
-		for j := 0; j < slice.Len(); j++ { // loop through resources of same kind
-			x := slice.Index(j).Interface()
-			obj, ok := x.(Res) // convert to Res type
-			if !ok {
-				return nil, fmt.Errorf("Error: Config: Can't convert: %v of type: %T to Res.", x, x)
-			}
-
-			if _, exists := lookup[kind]; !exists {
-				lookup[kind] = make(map[string]*Vertex)
-			}
-			// XXX: should we export based on a @@ prefix, or a metaparam
-			// like exported => true || exported => (host pattern)||(other pattern?)
-			if !strings.HasPrefix(obj.GetName(), "@@") { // exported resource
-				// XXX: we don't have a way of knowing if any of the
-				// metaparams are undefined, and as a result to set the
-				// defaults that we want! I hate the go yaml parser!!!
-				v := graph.GetVertexMatch(obj)
-				if v == nil { // no match found
-					obj.Init()
-					v = NewVertex(obj)
-					graph.AddVertex(v) // call standalone in case not part of an edge
-				}
-				lookup[kind][obj.GetName()] = v // used for constructing edges
-				keep = append(keep, v)          // append
-
-			} else {
-				// XXX: do this in a different function...
-				// add to etcd storage...
-				obj.SetName(obj.GetName()[2:]) //slice off @@
-
-				data, err := ResToB64(obj)
-				if err != nil {
-					return nil, fmt.Errorf("Config: Could not encode %v resource: %v, error: %v", kind, obj.GetName(), err)
-				}
-
-				if !etcdO.EtcdPut(hostname, obj.GetName(), kind, data) {
-					return nil, fmt.Errorf("Config: Could not export %v resource: %v", kind, obj.GetName())
-				}
-			}
-		}
-	}
-
-	// lookup from etcd graph
-	// do all the graph look ups in one single step, so that if the etcd
-	// database changes, we don't have a partial state of affairs...
-	nodes, ok := etcdO.EtcdGet()
-	if ok {
-		for _, t := range config.Collector {
-			// XXX: should we just drop these everywhere and have the kind strings be all lowercase?
-			kind := FirstToUpper(t.Kind)
-
-			// use t.Kind and optionally t.Pattern to collect from etcd storage
-			log.Printf("Collect: %v; Pattern: %v", kind, t.Pattern)
-			for _, str := range etcdO.EtcdGetProcess(nodes, kind) {
-				obj, err := B64ToRes(str)
-				if err != nil {
-					log.Printf("B64ToRes failed to decode: %v", err)
-					log.Printf("Collect: %v: not collected!", kind)
-					continue
-				}
-
-				if t.Pattern != "" { // XXX: simplistic for now
-					obj.CollectPattern(t.Pattern) // obj.Dirname = t.Pattern
-				}
-
-				log.Printf("Collect: %v[%v]: collected!", kind, obj.GetName())
-
-				// XXX: similar to other resource add code:
-				if _, exists := lookup[kind]; !exists {
-					lookup[kind] = make(map[string]*Vertex)
-				}
-				v := graph.GetVertexMatch(obj)
-				if v == nil { // no match found
-					obj.Init() // initialize go channels or things won't work!!!
-					v = NewVertex(obj)
-					graph.AddVertex(v) // call standalone in case not part of an edge
-				}
-				lookup[kind][obj.GetName()] = v // used for constructing edges
-				keep = append(keep, v)          // append
-			}
-		}
-	}
-
-	// get rid of any vertices we shouldn't "keep" (that aren't in new graph)
-	for _, v := range graph.GetVertices() {
-		if !VertexContains(v, keep) {
-			// wait for exit before starting new graph!
-			v.SendEvent(eventExit, true, false)
-			graph.DeleteVertex(v)
-		}
-	}
-
-	for _, e := range config.Edges {
-		if _, ok := lookup[FirstToUpper(e.From.Kind)]; !ok {
-			return nil, fmt.Errorf("Can't find 'from' resource!")
-		}
-		if _, ok := lookup[FirstToUpper(e.To.Kind)]; !ok {
-			return nil, fmt.Errorf("Can't find 'to' resource!")
-		}
-		if _, ok := lookup[FirstToUpper(e.From.Kind)][e.From.Name]; !ok {
-			return nil, fmt.Errorf("Can't find 'from' name!")
-		}
-		if _, ok := lookup[FirstToUpper(e.To.Kind)][e.To.Name]; !ok {
-			return nil, fmt.Errorf("Can't find 'to' name!")
-		}
-		graph.AddEdge(lookup[FirstToUpper(e.From.Kind)][e.From.Name], lookup[FirstToUpper(e.To.Kind)][e.To.Name], NewEdge(e.Name))
-	}
-
-	return graph, nil
-}
-
-// add edges to the vertex in a graph based on if it matches a uuid list
-func (g *Graph) addEdgesByMatchingUUIDS(v *Vertex, uuids []ResUUID) []bool {
-	// search for edges and see what matches!
-	var result []bool
-
-	// loop through each uuid, and see if it matches any vertex
-	for _, uuid := range uuids {
-		var found = false
-		// uuid is a ResUUID object
-		for _, vv := range g.GetVertices() { // search
-			if v == vv { // skip self
-				continue
-			}
-			if DEBUG {
-				log.Printf("Compile: AutoEdge: Match: %v[%v] with UUID: %v[%v]", vv.Kind(), vv.GetName(), uuid.Kind(), uuid.GetName())
-			}
-			// we must match to an effective UUID for the resource,
-			// that is to say, the name value of a res is a helpful
-			// handle, but it is not necessarily a unique identity!
-			// remember, resources can return multiple UUID's each!
-			if UUIDExistsInUUIDs(uuid, vv.GetUUIDs()) {
-				// add edge from: vv -> v
-				if uuid.Reversed() {
-					txt := fmt.Sprintf("AutoEdge: %v[%v] -> %v[%v]", vv.Kind(), vv.GetName(), v.Kind(), v.GetName())
-					log.Printf("Compile: Adding %v", txt)
-					g.AddEdge(vv, v, NewEdge(txt))
-				} else { // edges go the "normal" way, eg: pkg resource
-					txt := fmt.Sprintf("AutoEdge: %v[%v] -> %v[%v]", v.Kind(), v.GetName(), vv.Kind(), vv.GetName())
-					log.Printf("Compile: Adding %v", txt)
-					g.AddEdge(v, vv, NewEdge(txt))
-				}
-				found = true
-				break
-			}
-		}
-		result = append(result, found)
-	}
-	return result
-}
-
-// add auto edges to graph
-func (g *Graph) AutoEdges() {
-	log.Println("Compile: Adding AutoEdges...")
-	for _, v := range g.GetVertices() { // for each vertexes autoedges
-		if !v.GetMeta().AutoEdge { // is the metaparam true?
-			continue
-		}
-		autoEdgeObj := v.AutoEdges()
-		if autoEdgeObj == nil {
-			log.Printf("%v[%v]: Config: No auto edges were found!", v.Kind(), v.GetName())
-			continue // next vertex
-		}
-
-		for { // while the autoEdgeObj has more uuids to add...
-			uuids := autoEdgeObj.Next() // get some!
-			if uuids == nil {
-				log.Printf("%v[%v]: Config: The auto edge list is empty!", v.Kind(), v.GetName())
-				break // inner loop
-			}
-			if DEBUG {
-				log.Println("Compile: AutoEdge: UUIDS:")
-				for i, u := range uuids {
-					log.Printf("Compile: AutoEdge: UUID%d: %v", i, u)
-				}
-			}
-
-			// match and add edges
-			result := g.addEdgesByMatchingUUIDS(v, uuids)
-
-			// report back, and find out if we should continue
-			if !autoEdgeObj.Test(result) {
-				break
-			}
-		}
-	}
-}
-
-// AutoGrouper is the required interface to implement for an autogroup algorithm
-type AutoGrouper interface {
-	// listed in the order these are typically called in...
-	name() string                                  // friendly identifier
-	init(*Graph) error                             // only call once
-	vertexNext() (*Vertex, *Vertex, error)         // mostly algorithmic
-	vertexCmp(*Vertex, *Vertex) error              // can we merge these ?
-	vertexMerge(*Vertex, *Vertex) (*Vertex, error) // vertex merge fn to use
-	edgeMerge(*Edge, *Edge) *Edge                  // edge merge fn to use
-	vertexTest(bool) (bool, error)                 // call until false
-}
-
-// baseGrouper is the base type for implementing the AutoGrouper interface
-type baseGrouper struct {
-	graph    *Graph    // store a pointer to the graph
-	vertices []*Vertex // cached list of vertices
-	i        int
-	j        int
-	done     bool
-}
-
-// name provides a friendly name for the logs to see
-func (ag *baseGrouper) name() string {
-	return "baseGrouper"
-}
-
-// init is called only once and before using other AutoGrouper interface methods
-// the name method is the only exception: call it any time without side effects!
-func (ag *baseGrouper) init(g *Graph) error {
-	if ag.graph != nil {
-		return fmt.Errorf("The init method has already been called!")
-	}
-	ag.graph = g                               // pointer
-	ag.vertices = ag.graph.GetVerticesSorted() // cache in deterministic order!
-	ag.i = 0
-	ag.j = 0
-	if len(ag.vertices) == 0 { // empty graph
-		ag.done = true
-		return nil
-	}
-	return nil
-}
-
-// vertexNext is a simple iterator that loops through vertex (pair) combinations
-// an intelligent algorithm would selectively offer only valid pairs of vertices
-// these should satisfy logical grouping requirements for the autogroup designs!
-// the desired algorithms can override, but keep this method as a base iterator!
-func (ag *baseGrouper) vertexNext() (v1, v2 *Vertex, err error) {
-	// this does a for v... { for w... { return v, w }} but stepwise!
-	l := len(ag.vertices)
-	if ag.i < l {
-		v1 = ag.vertices[ag.i]
-	}
-	if ag.j < l {
-		v2 = ag.vertices[ag.j]
-	}
-
-	// in case the vertex was deleted
-	if !ag.graph.HasVertex(v1) {
-		v1 = nil
-	}
-	if !ag.graph.HasVertex(v2) {
-		v2 = nil
-	}
-
-	// two nested loops...
-	if ag.j < l {
-		ag.j++
-	}
-	if ag.j == l {
-		ag.j = 0
-		if ag.i < l {
-			ag.i++
-		}
-		if ag.i == l {
-			ag.done = true
-		}
-	}
-
-	return
-}
-
-func (ag *baseGrouper) vertexCmp(v1, v2 *Vertex) error {
-	if v1 == nil || v2 == nil {
-		return fmt.Errorf("Vertex is nil!")
-	}
-	if v1 == v2 { // skip yourself
-		return fmt.Errorf("Vertices are the same!")
-	}
-	if v1.Kind() != v2.Kind() { // we must group similar kinds
-		// TODO: maybe future resources won't need this limitation?
-		return fmt.Errorf("The two resources aren't the same kind!")
-	}
-	// someone doesn't want to group!
-	if !v1.GetMeta().AutoGroup || !v2.GetMeta().AutoGroup {
-		return fmt.Errorf("One of the autogroup flags is false!")
-	}
-	if v1.Res.IsGrouped() { // already grouped!
-		return fmt.Errorf("Already grouped!")
-	}
-	if len(v2.Res.GetGroup()) > 0 { // already has children grouped!
-		return fmt.Errorf("Already has groups!")
-	}
-	if !v1.Res.GroupCmp(v2.Res) { // resource groupcmp failed!
-		return fmt.Errorf("The GroupCmp failed!")
-	}
-	return nil // success
-}
-
-func (ag *baseGrouper) vertexMerge(v1, v2 *Vertex) (v *Vertex, err error) {
-	// NOTE: it's important to use w.Res instead of w, b/c
-	// the w by itself is the *Vertex obj, not the *Res obj
-	// which is contained within it! They both satisfy the
-	// Res interface, which is why both will compile! :(
-	err = v1.Res.GroupRes(v2.Res) // GroupRes skips stupid groupings
-	return                        // success or fail, and no need to merge the actual vertices!
-}
-
-func (ag *baseGrouper) edgeMerge(e1, e2 *Edge) *Edge {
-	return e1 // noop
-}
-
-// vertexTest processes the results of the grouping for the algorithm to know
-// return an error if something went horribly wrong, and bool false to stop
-func (ag *baseGrouper) vertexTest(b bool) (bool, error) {
-	// NOTE: this particular baseGrouper version doesn't track what happens
-	// because since we iterate over every pair, we don't care which merge!
-	if ag.done {
-		return false, nil
-	}
-	return true, nil
-}
-
-// TODO: this algorithm may not be correct in all cases. replace if needed!
-type nonReachabilityGrouper struct {
-	baseGrouper // "inherit" what we want, and reimplement the rest
-}
-
-func (ag *nonReachabilityGrouper) name() string {
-	return "nonReachabilityGrouper"
-}
-
-// this algorithm relies on the observation that if there's a path from a to b,
-// then they *can't* be merged (b/c of the existing dependency) so therefore we
-// merge anything that *doesn't* satisfy this condition or that of the reverse!
-func (ag *nonReachabilityGrouper) vertexNext() (v1, v2 *Vertex, err error) {
-	for {
-		v1, v2, err = ag.baseGrouper.vertexNext() // get all iterable pairs
-		if err != nil {
-			log.Fatalf("Error running autoGroup(vertexNext): %v", err)
-		}
-
-		if v1 != v2 { // ignore self cmp early (perf optimization)
-			// if NOT reachable, they're viable...
-			out1 := ag.graph.Reachability(v1, v2)
-			out2 := ag.graph.Reachability(v2, v1)
-			if len(out1) == 0 && len(out2) == 0 {
-				return // return v1 and v2, they're viable
-			}
-		}
-
-		// if we got here, it means we're skipping over this candidate!
-		if ok, err := ag.baseGrouper.vertexTest(false); err != nil {
-			log.Fatalf("Error running autoGroup(vertexTest): %v", err)
-		} else if !ok {
-			return nil, nil, nil // done!
-		}
-
-		// the vertexTest passed, so loop and try with a new pair...
-	}
-}
-
-// autoGroup is the mechanical auto group "runner" that runs the interface spec
-func (g *Graph) autoGroup(ag AutoGrouper) chan string {
-	strch := make(chan string) // output log messages here
-	go func(strch chan string) {
-		strch <- fmt.Sprintf("Compile: Grouping: Algorithm: %v...", ag.name())
-		if err := ag.init(g); err != nil {
-			log.Fatalf("Error running autoGroup(init): %v", err)
-		}
-
-		for {
-			var v, w *Vertex
-			v, w, err := ag.vertexNext() // get pair to compare
-			if err != nil {
-				log.Fatalf("Error running autoGroup(vertexNext): %v", err)
-			}
-			merged := false
-			// save names since they change during the runs
-			vStr := fmt.Sprintf("%s", v) // valid even if it is nil
-			wStr := fmt.Sprintf("%s", w)
-
-			if err := ag.vertexCmp(v, w); err != nil { // cmp ?
-				if DEBUG {
-					strch <- fmt.Sprintf("Compile: Grouping: !GroupCmp for: %s into %s", wStr, vStr)
-				}
-
-				// remove grouped vertex and merge edges (res is safe)
-			} else if err := g.VertexMerge(v, w, ag.vertexMerge, ag.edgeMerge); err != nil { // merge...
-				strch <- fmt.Sprintf("Compile: Grouping: !VertexMerge for: %s into %s", wStr, vStr)
-
-			} else { // success!
-				strch <- fmt.Sprintf("Compile: Grouping: Success for: %s into %s", wStr, vStr)
-				merged = true // woo
-			}
-
-			// did these get used?
-			if ok, err := ag.vertexTest(merged); err != nil {
-				log.Fatalf("Error running autoGroup(vertexTest): %v", err)
-			} else if !ok {
-				break // done!
-			}
-		}
-
-		close(strch)
-		return
-	}(strch) // call function
-	return strch
-}
-
-// AutoGroup runs the auto grouping on the graph and prints out log messages
-func (g *Graph) AutoGroup() {
-	// receive log messages from channel...
-	// this allows test cases to avoid printing them when they're unwanted!
-	// TODO: this algorithm may not be correct in all cases. replace if needed!
-	for str := range g.autoGroup(&nonReachabilityGrouper{}) {
-		log.Println(str)
-	}
-}

configwatch.go

@@ -1,155 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-import (
-	"gopkg.in/fsnotify.v1"
-	//"github.com/go-fsnotify/fsnotify" // git master of "gopkg.in/fsnotify.v1"
-	"log"
-	"math"
-	"path"
-	"strings"
-	"syscall"
-)
-
-// XXX: it would be great if we could reuse code between this and the file resource
-// XXX: patch this to submit it as part of go-fsnotify if they're interested...
-func ConfigWatch(file string) chan bool {
-	ch := make(chan bool)
-	go func() {
-		var safename = path.Clean(file) // no trailing slash
-
-		watcher, err := fsnotify.NewWatcher()
-		if err != nil {
-			log.Fatal(err)
-		}
-		defer watcher.Close()
-
-		patharray := PathSplit(safename) // tokenize the path
-		var index = len(patharray)       // starting index
-		var current string               // current "watcher" location
-		var deltaDepth int               // depth delta between watcher and event
-		var send = false                 // send event?
-
-		for {
-			current = strings.Join(patharray[0:index], "/")
-			if current == "" { // the empty string top is the root dir ("/")
-				current = "/"
-			}
-			log.Printf("Watching: %v", current) // attempting to watch...
-
-			// initialize in the loop so that we can reset on rm-ed handles
-			err = watcher.Add(current)
-			if err != nil {
-				if err == syscall.ENOENT {
-					index-- // usually not found, move up one dir
-				} else if err == syscall.ENOSPC {
-					// XXX: occasionally: no space left on device,
-					// XXX: probably due to lack of inotify watches
-					log.Printf("Out of inotify watches for config(%v)", file)
-					log.Fatal(err)
-				} else {
-					log.Printf("Unknown config(%v) error:", file)
-					log.Fatal(err)
-				}
-				index = int(math.Max(1, float64(index)))
-				continue
-			}
-
-			select {
-			case event := <-watcher.Events:
-				// the deeper you go, the bigger the deltaDepth is...
-				// this is the difference between what we're watching,
-				// and the event... doesn't mean we can't watch deeper
-				if current == event.Name {
-					deltaDepth = 0 // i was watching what i was looking for
-
-				} else if HasPathPrefix(event.Name, current) {
-					deltaDepth = len(PathSplit(current)) - len(PathSplit(event.Name)) // -1 or less
-
-				} else if HasPathPrefix(current, event.Name) {
-					deltaDepth = len(PathSplit(event.Name)) - len(PathSplit(current)) // +1 or more
-
-				} else {
-					// TODO different watchers get each others events!
-					// https://github.com/go-fsnotify/fsnotify/issues/95
-					// this happened with two values such as:
-					// event.Name: /tmp/mgmt/f3 and current: /tmp/mgmt/f2
-					continue
-				}
-				//log.Printf("The delta depth is: %v", deltaDepth)
-
-				// if we have what we wanted, awesome, send an event...
-				if event.Name == safename {
-					//log.Println("Event!")
-					send = true
-
-					// file removed, move the watch upwards
-					if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
-						//log.Println("Removal!")
-						watcher.Remove(current)
-						index--
-					}
-
-					// we must be a parent watcher, so descend in
-					if deltaDepth < 0 {
-						watcher.Remove(current)
-						index++
-					}
-
-					// if safename starts with event.Name, we're above, and no event should be sent
-				} else if HasPathPrefix(safename, event.Name) {
-					//log.Println("Above!")
-
-					if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
-						log.Println("Removal!")
-						watcher.Remove(current)
-						index--
-					}
-
-					if deltaDepth < 0 {
-						log.Println("Parent!")
-						if PathPrefixDelta(safename, event.Name) == 1 { // we're the parent dir
-							//send = true
-						}
-						watcher.Remove(current)
-						index++
-					}
-
-					// if event.Name startswith safename, send event, we're already deeper
-				} else if HasPathPrefix(event.Name, safename) {
-					//log.Println("Event2!")
-					//send = true
-				}
-
-			case err := <-watcher.Errors:
-				log.Printf("error: %v", err)
-				log.Fatal(err)
-
-			}
-
-			// do our event sending all together to avoid duplicate msgs
-			if send {
-				send = false
-				ch <- true
-			}
-		}
-		//close(ch)
-	}()
-	return ch
-}

converger/converger.go

@@ -0,0 +1,382 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package converger is a facility for reporting the converged state.
+package converger
+
+import (
+	"fmt"
+	"sync"
+	"time"
+
+	"github.com/purpleidea/mgmt/util"
+)
+
+// TODO: we could make a new function that masks out the state of certain
+// UID's, but at the moment the new Timer code has obsoleted the need...
+
+// Converger is the general interface for implementing a convergence watcher
+type Converger interface { // TODO: need a better name
+	Register() ConvergerUID
+	IsConverged(ConvergerUID) bool         // is the UID converged ?
+	SetConverged(ConvergerUID, bool) error // set the converged state of the UID
+	Unregister(ConvergerUID)
+	Start()
+	Pause()
+	Loop(bool)
+	ConvergedTimer(ConvergerUID) <-chan time.Time
+	Status() map[uint64]bool
+	Timeout() int                // returns the timeout that this was created with
+	SetStateFn(func(bool) error) // sets the stateFn
+}
+
+// ConvergerUID is the interface resources can use to notify with if converged
+// you'll need to use part of the Converger interface to Register initially too
+type ConvergerUID interface {
+	ID() uint64   // get Id
+	Name() string // get a friendly name
+	SetName(string)
+	IsValid() bool // has Id been initialized ?
+	InvalidateID() // set Id to nil
+	IsConverged() bool
+	SetConverged(bool) error
+	Unregister()
+	ConvergedTimer() <-chan time.Time
+	StartTimer() (func() error, error) // cancellable is the same as StopTimer()
+	ResetTimer() error                 // resets counter to zero
+	StopTimer() error
+}
+
+// converger is an implementation of the Converger interface
+type converger struct {
+	timeout   int              // must be zero (instant) or greater seconds to run
+	stateFn   func(bool) error // run on converged state changes with state bool
+	converged bool             // did we converge (state changes of this run Fn)
+	channel   chan struct{}    // signal here to run an isConverged check
+	control   chan bool        // control channel for start/pause
+	mutex     sync.RWMutex     // used for controlling access to status and lastid
+	lastid    uint64
+	status    map[uint64]bool
+}
+
+// convergerUID is an implementation of the ConvergerUID interface
+type convergerUID struct {
+	converger Converger
+	id        uint64
+	name      string // user defined, friendly name
+	mutex     sync.Mutex
+	timer     chan struct{}
+	running   bool // is the above timer running?
+}
+
+// NewConverger builds a new converger struct
+func NewConverger(timeout int, stateFn func(bool) error) *converger {
+	return &converger{
+		timeout: timeout,
+		stateFn: stateFn,
+		channel: make(chan struct{}),
+		control: make(chan bool),
+		lastid:  0,
+		status:  make(map[uint64]bool),
+	}
+}
+
+// Register assigns a ConvergerUID to the caller
+func (obj *converger) Register() ConvergerUID {
+	obj.mutex.Lock()
+	defer obj.mutex.Unlock()
+	obj.lastid++
+	obj.status[obj.lastid] = false // initialize as not converged
+	return &convergerUID{
+		converger: obj,
+		id:        obj.lastid,
+		name:      fmt.Sprintf("%d", obj.lastid), // some default
+		timer:     nil,
+		running:   false,
+	}
+}
+
+// IsConverged gets the converged status of a uid
+func (obj *converger) IsConverged(uid ConvergerUID) bool {
+	if !uid.IsValid() {
+		panic(fmt.Sprintf("Id of ConvergerUID(%s) is nil!", uid.Name()))
+	}
+	obj.mutex.RLock()
+	isConverged, found := obj.status[uid.ID()] // lookup
+	obj.mutex.RUnlock()
+	if !found {
+		panic("Id of ConvergerUID is unregistered!")
+	}
+	return isConverged
+}
+
+// SetConverged updates the converger with the converged state of the UID
+func (obj *converger) SetConverged(uid ConvergerUID, isConverged bool) error {
+	if !uid.IsValid() {
+		return fmt.Errorf("Id of ConvergerUID(%s) is nil!", uid.Name())
+	}
+	obj.mutex.Lock()
+	if _, found := obj.status[uid.ID()]; !found {
+		panic("Id of ConvergerUID is unregistered!")
+	}
+	obj.status[uid.ID()] = isConverged // set
+	obj.mutex.Unlock()                 // unlock *before* poke or deadlock!
+	if isConverged != obj.converged {  // only poke if it would be helpful
+		// run in a go routine so that we never block... just queue up!
+		// this allows us to send events, even if we haven't started...
+		go func() { obj.channel <- struct{}{} }()
+	}
+	return nil
+}
+
+// isConverged returns true if *every* registered uid has converged
+func (obj *converger) isConverged() bool {
+	obj.mutex.RLock() // take a read lock
+	defer obj.mutex.RUnlock()
+	for _, v := range obj.status {
+		if !v { // everyone must be converged for this to be true
+			return false
+		}
+	}
+	return true
+}
+
+// Unregister dissociates the ConvergedUID from the converged checking
+func (obj *converger) Unregister(uid ConvergerUID) {
+	if !uid.IsValid() {
+		panic(fmt.Sprintf("Id of ConvergerUID(%s) is nil!", uid.Name()))
+	}
+	obj.mutex.Lock()
+	uid.StopTimer() // ignore any errors
+	delete(obj.status, uid.ID())
+	obj.mutex.Unlock()
+	uid.InvalidateID()
+}
+
+// Start causes a Converger object to start or resume running
+func (obj *converger) Start() {
+	obj.control <- true
+}
+
+// Pause causes a Converger object to stop running temporarily
+func (obj *converger) Pause() { // FIXME: add a sync ACK on pause before return
+	obj.control <- false
+}
+
+// Loop is the main loop for a Converger object; it usually runs in a goroutine
+// TODO: we could eventually have each resource tell us as soon as it converges
+// and then keep track of the time delays here, to avoid callers needing select
+// NOTE: when we have very short timeouts, if we start before all the resources
+// have joined the map, then it might appears as if we converged before we did!
+func (obj *converger) Loop(startPaused bool) {
+	if obj.control == nil {
+		panic("Converger not initialized correctly")
+	}
+	if startPaused { // start paused without racing
+		select {
+		case e := <-obj.control:
+			if !e {
+				panic("Converger expected true!")
+			}
+		}
+	}
+	for {
+		select {
+		case e := <-obj.control: // expecting "false" which means pause!
+			if e {
+				panic("Converger expected false!")
+			}
+			// now i'm paused...
+			select {
+			case e := <-obj.control:
+				if !e {
+					panic("Converger expected true!")
+				}
+				// restart
+				// kick once to refresh the check...
+				go func() { obj.channel <- struct{}{} }()
+				continue
+			}
+
+		case <-obj.channel:
+			if !obj.isConverged() {
+				if obj.converged { // we're doing a state change
+					if obj.stateFn != nil {
+						// call an arbitrary function
+						if err := obj.stateFn(false); err != nil {
+							// FIXME: what to do on error ?
+						}
+					}
+				}
+				obj.converged = false
+				continue
+			}
+
+			// we have converged!
+			if obj.timeout >= 0 { // only run if timeout is valid
+				if !obj.converged { // we're doing a state change
+					if obj.stateFn != nil {
+						// call an arbitrary function
+						if err := obj.stateFn(true); err != nil {
+							// FIXME: what to do on error ?
+						}
+					}
+				}
+			}
+			obj.converged = true
+			// loop and wait again...
+		}
+	}
+}
+
+// ConvergedTimer adds a timeout to a select call and blocks until then
+// TODO: this means we could eventually have per resource converged timeouts
+func (obj *converger) ConvergedTimer(uid ConvergerUID) <-chan time.Time {
+	// be clever: if i'm already converged, this timeout should block which
+	// avoids unnecessary new signals being sent! this avoids fast loops if
+	// we have a low timeout, or in particular a timeout == 0
+	if uid.IsConverged() {
+		// blocks the case statement in select forever!
+		return util.TimeAfterOrBlock(-1)
+	}
+	return util.TimeAfterOrBlock(obj.timeout)
+}
+
+// Status returns a map of the converged status of each UID.
+func (obj *converger) Status() map[uint64]bool {
+	status := make(map[uint64]bool)
+	obj.mutex.RLock() // take a read lock
+	defer obj.mutex.RUnlock()
+	for k, v := range obj.status { // make a copy to avoid the mutex
+		status[k] = v
+	}
+	return status
+}
+
+// Timeout returns the timeout in seconds that converger was created with. This
+// is useful to avoid passing in the timeout value separately when you're
+// already passing in the Converger struct.
+func (obj *converger) Timeout() int {
+	return obj.timeout
+}
+
+// SetStateFn sets the state function to be run on change of converged state.
+func (obj *converger) SetStateFn(stateFn func(bool) error) {
+	obj.stateFn = stateFn
+}
+
+// Id returns the unique id of this UID object
+func (obj *convergerUID) ID() uint64 {
+	return obj.id
+}
+
+// Name returns a user defined name for the specific convergerUID.
+func (obj *convergerUID) Name() string {
+	return obj.name
+}
+
+// SetName sets a user defined name for the specific convergerUID.
+func (obj *convergerUID) SetName(name string) {
+	obj.name = name
+}
+
+// IsValid tells us if the id is valid or has already been destroyed
+func (obj *convergerUID) IsValid() bool {
+	return obj.id != 0 // an id of 0 is invalid
+}
+
+// InvalidateID marks the id as no longer valid
+func (obj *convergerUID) InvalidateID() {
+	obj.id = 0 // an id of 0 is invalid
+}
+
+// IsConverged is a helper function to the regular IsConverged method
+func (obj *convergerUID) IsConverged() bool {
+	return obj.converger.IsConverged(obj)
+}
+
+// SetConverged is a helper function to the regular SetConverged notification
+func (obj *convergerUID) SetConverged(isConverged bool) error {
+	return obj.converger.SetConverged(obj, isConverged)
+}
+
+// Unregister is a helper function to unregister myself
+func (obj *convergerUID) Unregister() {
+	obj.converger.Unregister(obj)
+}
+
+// ConvergedTimer is a helper around the regular ConvergedTimer method
+func (obj *convergerUID) ConvergedTimer() <-chan time.Time {
+	return obj.converger.ConvergedTimer(obj)
+}
+
+// StartTimer runs an invisible timer that automatically converges on timeout.
+func (obj *convergerUID) StartTimer() (func() error, error) {
+	obj.mutex.Lock()
+	if !obj.running {
+		obj.timer = make(chan struct{})
+		obj.running = true
+	} else {
+		obj.mutex.Unlock()
+		return obj.StopTimer, fmt.Errorf("Timer already started!")
+	}
+	obj.mutex.Unlock()
+	go func() {
+		for {
+			select {
+			case _, ok := <-obj.timer: // reset signal channel
+				if !ok { // channel is closed
+					return // false to exit
+				}
+				obj.SetConverged(false)
+
+			case <-obj.ConvergedTimer():
+				obj.SetConverged(true) // converged!
+				select {
+				case _, ok := <-obj.timer: // reset signal channel
+					if !ok { // channel is closed
+						return // false to exit
+					}
+				}
+			}
+		}
+	}()
+	return obj.StopTimer, nil
+}
+
+// ResetTimer resets the counter to zero if using a StartTimer internally.
+func (obj *convergerUID) ResetTimer() error {
+	obj.mutex.Lock()
+	defer obj.mutex.Unlock()
+	if obj.running {
+		obj.timer <- struct{}{} // send the reset message
+		return nil
+	}
+	return fmt.Errorf("Timer hasn't been started!")
+}
+
+// StopTimer stops the running timer permanently until a StartTimer is run.
+func (obj *convergerUID) StopTimer() error {
+	obj.mutex.Lock()
+	defer obj.mutex.Unlock()
+	if !obj.running {
+		return fmt.Errorf("Timer isn't running!")
+	}
+	close(obj.timer)
+	obj.running = false
+	return nil
+}

doc.go

@@ -15,12 +15,5 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
+// Package main provides the main entrypoint for using the `mgmt` software.
 package main
-
-import (
-//"testing"
-)
-
-//func TestT1(t *testing.T) {
-
-//}

docker/Dockerfile

@@ -0,0 +1,22 @@
+FROM golang:1.6.2
+
+MAINTAINER Michał Czeraszkiewicz <contact@czerasz.com>
+
+# Set the reset cache variable
+# Read more here: http://czerasz.com/2014/11/13/docker-tip-and-tricks/#use-refreshedat-variable-for-better-cache-control
+ENV REFRESHED_AT 2016-05-10
+
+# Update the package list to be able to use required packages
+RUN apt-get update
+
+# Change the working directory
+WORKDIR /go/src/mgmt
+
+# Copy all the files to the working directory
+COPY . /go/src/mgmt
+
+# Install dependencies
+RUN make deps
+
+# Build the binary
+RUN make build

docker/Dockerfile.development

@@ -0,0 +1,31 @@
+FROM golang:1.6.2
+
+MAINTAINER Michał Czeraszkiewicz <contact@czerasz.com>
+
+# Set the reset cache variable
+# Read more here: http://czerasz.com/2014/11/13/docker-tip-and-tricks/#use-refreshedat-variable-for-better-cache-control
+ENV REFRESHED_AT 2016-05-14
+
+RUN apt-get update
+
+# Setup User to match Host User
+# Give the nre user superuser permissions
+ARG USER_ID=1000
+ARG GROUP_ID=1000
+ARG USER_NAME=mgmt
+ARG GROUP_NAME=$USER_NAME
+RUN groupadd --gid $GROUP_ID $GROUP_NAME && \
+	useradd --create-home --home /home/$USER_NAME --uid ${USER_ID} --gid $GROUP_NAME --groups sudo $USER_NAME && \
+	echo "$USER_NAME ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers
+
+# Copy all the files to the working directory
+COPY . /home/$USER_NAME/mgmt
+
+# Change working directory
+WORKDIR /home/$USER_NAME/mgmt
+
+# Install dependencies
+RUN make deps
+
+# Change user
+USER ${USER_NAME}

docker/scripts/build

@@ -0,0 +1,26 @@
+#!/bin/bash
+
+script_directory="$( cd "$( dirname "$0" )" && pwd )"
+project_directory=$script_directory/../..
+
+# Specify the Docker image name
+image_name='purpleidea/mgmt'
+
+# Build the image which contains the compiled binary
+docker build -t $image_name \
+	--file=$project_directory/docker/Dockerfile $project_directory
+
+# Remove the container if it already exists
+docker rm -f mgmt-export 2> /dev/null
+
+# Start the container in background so we can "copy out" the binary
+docker run -d --name=mgmt-export $image_name bash -c 'while true; sleep 1000; done'
+
+# Remove the current binary
+rm $project_directory/mgmt 2> /dev/null
+
+# Get the binary from the container
+docker cp mgmt-export:/go/src/mgmt/mgmt $project_directory/mgmt
+
+# Remove the container
+docker rm -f mgmt-export 2> /dev/null

docker/scripts/build-development

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+# Stop on any error
+set -e
+
+script_directory="$( cd "$( dirname "$0" )" && pwd )"
+project_directory=$script_directory/../..
+
+# Specify the Docker image name
+image_name='purpleidea/mgmt:development'
+
+# Build the image
+docker build -t $image_name \
+	--file=$project_directory/docker/Dockerfile.development $project_directory

docker/scripts/run-development

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+# Stop on any error
+set -e
+
+script_directory="$( cd "$( dirname "$0" )" && pwd )"
+project_directory=$script_directory/../..
+
+# Specify the Docker image name
+image_name='purpleidea/mgmt:development'
+
+# Run container in development mode
+docker run --rm --name=mgm_development --user=mgmt \
+	-v $project_directory:/home/mgmt/mgmt \
+	-it $image_name bash

etcd.go

@@ -1,294 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-import (
-	"fmt"
-	etcd "github.com/coreos/etcd/client"
-	etcd_context "golang.org/x/net/context"
-	"log"
-	"math"
-	"strings"
-	"time"
-)
-
-//go:generate stringer -type=etcdMsg -output=etcdmsg_stringer.go
-type etcdMsg int
-
-const (
-	etcdStart etcdMsg = iota
-	etcdEvent
-	etcdFoo
-	etcdBar
-)
-
-//go:generate stringer -type=etcdConvergedState -output=etcdconvergedstate_stringer.go
-type etcdConvergedState int
-
-const (
-	etcdConvergedNil etcdConvergedState = iota
-	//etcdConverged
-	etcdConvergedTimeout
-)
-
-type EtcdWObject struct { // etcd wrapper object
-	seed           string
-	ctimeout       int
-	converged      chan bool
-	kapi           etcd.KeysAPI
-	convergedState etcdConvergedState
-}
-
-func (etcdO *EtcdWObject) GetConvergedState() etcdConvergedState {
-	return etcdO.convergedState
-}
-
-func (etcdO *EtcdWObject) SetConvergedState(state etcdConvergedState) {
-	etcdO.convergedState = state
-}
-
-func (etcdO *EtcdWObject) GetKAPI() etcd.KeysAPI {
-	if etcdO.kapi != nil { // memoize
-		return etcdO.kapi
-	}
-
-	cfg := etcd.Config{
-		Endpoints: []string{etcdO.seed},
-		Transport: etcd.DefaultTransport,
-		// set timeout per request to fail fast when the target endpoint is unavailable
-		HeaderTimeoutPerRequest: time.Second,
-	}
-
-	var c etcd.Client
-	var err error
-
-	c, err = etcd.New(cfg)
-	if err != nil {
-		// XXX: not sure if this ever errors
-		if cerr, ok := err.(*etcd.ClusterError); ok {
-			// XXX: not sure if this part ever matches
-			// not running or disconnected
-			if cerr == etcd.ErrClusterUnavailable {
-				log.Fatal("XXX: etcd: ErrClusterUnavailable")
-			} else {
-				log.Fatal("XXX: etcd: Unknown")
-			}
-		}
-		log.Fatal(err) // some unhandled error
-	}
-	etcdO.kapi = etcd.NewKeysAPI(c)
-	return etcdO.kapi
-}
-
-type EtcdChannelWatchResponse struct {
-	resp *etcd.Response
-	err  error
-}
-
-// wrap the etcd watcher.Next blocking function inside of a channel
-func (etcdO *EtcdWObject) EtcdChannelWatch(watcher etcd.Watcher, context etcd_context.Context) chan *EtcdChannelWatchResponse {
-	ch := make(chan *EtcdChannelWatchResponse)
-	go func() {
-		for {
-			resp, err := watcher.Next(context) // blocks here
-			ch <- &EtcdChannelWatchResponse{resp, err}
-		}
-	}()
-	return ch
-}
-
-func (etcdO *EtcdWObject) EtcdWatch() chan etcdMsg {
-	kapi := etcdO.GetKAPI()
-	ctimeout := etcdO.ctimeout
-	converged := etcdO.converged
-	// XXX: i think we need this buffered so that when we're hanging on the
-	// channel, which is inside the EtcdWatch main loop, we still want the
-	// calls to Get/Set on etcd to succeed, so blocking them here would
-	// kill the whole thing
-	ch := make(chan etcdMsg, 1) // XXX: buffer of at least 1 is required
-	go func(ch chan etcdMsg) {
-		tmin := 500   // initial (min) delay in ms
-		t := tmin     // current time
-		tmult := 2    // multiplier for exponential delay
-		tmax := 16000 // max delay
-		watcher := kapi.Watcher("/exported/", &etcd.WatcherOptions{Recursive: true})
-		etcdch := etcdO.EtcdChannelWatch(watcher, etcd_context.Background())
-		for {
-			log.Printf("Etcd: Watching...")
-			var resp *etcd.Response // = nil by default
-			var err error
-			select {
-			case out := <-etcdch:
-				etcdO.SetConvergedState(etcdConvergedNil)
-				resp, err = out.resp, out.err
-
-			case _ = <-TimeAfterOrBlock(ctimeout):
-				etcdO.SetConvergedState(etcdConvergedTimeout)
-				converged <- true
-				continue
-			}
-
-			if err != nil {
-				if err == etcd_context.Canceled {
-					// ctx is canceled by another routine
-					log.Fatal("Canceled")
-				} else if err == etcd_context.DeadlineExceeded {
-					// ctx is attached with a deadline and it exceeded
-					log.Fatal("Deadline")
-				} else if cerr, ok := err.(*etcd.ClusterError); ok {
-					// not running or disconnected
-					// TODO: is there a better way to parse errors?
-					for _, e := range cerr.Errors {
-						if strings.HasSuffix(e.Error(), "getsockopt: connection refused") {
-							t = int(math.Min(float64(t*tmult), float64(tmax)))
-							log.Printf("Etcd: Waiting %d ms for connection...", t)
-							time.Sleep(time.Duration(t) * time.Millisecond) // sleep for t ms
-
-						} else if e.Error() == "unexpected EOF" {
-							log.Printf("Etcd: Unexpected disconnect...")
-
-						} else if e.Error() == "EOF" {
-							log.Printf("Etcd: Disconnected...")
-
-						} else if strings.HasPrefix(e.Error(), "unsupported protocol scheme") {
-							// usually a bad peer endpoint value
-							log.Fatal("Bad peer endpoint value?")
-
-						} else {
-							log.Fatal("Woops: ", e.Error())
-						}
-					}
-				} else {
-					// bad cluster endpoints, which are not etcd servers
-					log.Fatal("Woops: ", err)
-				}
-			} else {
-				//log.Print(resp)
-				//log.Printf("Watcher().Node.Value(%v): %+v", key, resp.Node.Value)
-				// FIXME: we should actually reset when the server comes back, not here on msg!
-				//XXX: can we fix this with one of these patterns?: https://blog.golang.org/go-concurrency-patterns-timing-out-and
-				t = tmin // reset timer
-
-				// don't trigger event if nothing changed
-				if n, p := resp.Node, resp.PrevNode; resp.Action == "set" && p != nil {
-					if n.Key == p.Key && n.Value == p.Value {
-						continue
-					}
-				}
-
-				// FIXME: we get events on key/res/value changes for
-				// each res directory... ignore the non final ones...
-				// IOW, ignore everything except for the value or some
-				// field which gets set last... this could be the max count field thing...
-
-				log.Printf("Etcd: Value: %v", resp.Node.Value) // event
-				ch <- etcdEvent                                // event
-			}
-
-		} // end for loop
-		//close(ch)
-	}(ch) // call go routine
-	return ch
-}
-
-// helper function to store our data in etcd
-func (etcdO *EtcdWObject) EtcdPut(hostname, key, res string, data string) bool {
-	kapi := etcdO.GetKAPI()
-	path := fmt.Sprintf("/exported/%s/resources/%s/res", hostname, key)
-	_, err := kapi.Set(etcd_context.Background(), path, res, nil)
-	// XXX validate...
-
-	path = fmt.Sprintf("/exported/%s/resources/%s/value", hostname, key)
-	resp, err := kapi.Set(etcd_context.Background(), path, data, nil)
-	if err != nil {
-		if cerr, ok := err.(*etcd.ClusterError); ok {
-			// not running or disconnected
-			for _, e := range cerr.Errors {
-				if strings.HasSuffix(e.Error(), "getsockopt: connection refused") {
-				}
-				//if e == etcd.ErrClusterUnavailable
-			}
-		}
-		log.Printf("Etcd: Could not store %v key.", key)
-		return false
-	}
-	log.Print("Etcd: ", resp) // w00t... bonus
-	return true
-}
-
-// lookup /exported/ node hierarchy
-func (etcdO *EtcdWObject) EtcdGet() (etcd.Nodes, bool) {
-	kapi := etcdO.GetKAPI()
-	// key structure is /exported/<hostname>/resources/...
-	resp, err := kapi.Get(etcd_context.Background(), "/exported/", &etcd.GetOptions{Recursive: true})
-	if err != nil {
-		return nil, false // not found
-	}
-	return resp.Node.Nodes, true
-}
-
-func (etcdO *EtcdWObject) EtcdGetProcess(nodes etcd.Nodes, res string) []string {
-	//path := fmt.Sprintf("/exported/%s/resources/", h)
-	top := "/exported/"
-	log.Printf("Etcd: Get: %+v", nodes) // Get().Nodes.Nodes
-	var output []string
-
-	for _, x := range nodes { // loop through hosts
-		if !strings.HasPrefix(x.Key, top) {
-			log.Fatal("Error!")
-		}
-		host := x.Key[len(top):]
-		//log.Printf("Get().Nodes[%v]: %+v ==> %+v", -1, host, x.Nodes)
-		//log.Printf("Get().Nodes[%v]: %+v ==> %+v", i, x.Key, x.Nodes)
-		resources, ok := EtcdGetChildNodeByKey(x, "resources")
-		if !ok {
-			continue
-		}
-		for _, y := range resources.Nodes { // loop through resources
-			//key := y.Key # UUID?
-			//log.Printf("Get(%v): RES[%v]", host, y.Key)
-			t, ok := EtcdGetChildNodeByKey(y, "res")
-			if !ok {
-				continue
-			}
-			if res != "" && res != t.Value {
-				continue
-			} // filter based on res
-
-			v, ok := EtcdGetChildNodeByKey(y, "value") // B64ToObj this
-			if !ok {
-				continue
-			}
-			log.Printf("Etcd: Hostname: %v; Get: %v", host, t.Value)
-
-			output = append(output, v.Value)
-		}
-	}
-	return output
-}
-
-// TODO: wrap this somehow so it's a method of *etcd.Node
-// helper function that returns the node for a particular key under a node
-func EtcdGetChildNodeByKey(node *etcd.Node, key string) (*etcd.Node, bool) {
-	for _, x := range node.Nodes {
-		if x.Key == fmt.Sprintf("%s/%s", node.Key, key) {
-			return x, true
-		}
-	}
-	return nil, false // not found
-}

event.go

@@ -1,75 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-//go:generate stringer -type=eventName -output=eventname_stringer.go
-type eventName int
-
-const (
-	eventNil eventName = iota
-	eventExit
-	eventStart
-	eventPause
-	eventPoke
-	eventBackPoke
-)
-
-type Resp chan bool
-
-type Event struct {
-	Name eventName
-	Resp Resp // channel to send an ack response on, nil to skip
-	//Wg   *sync.WaitGroup // receiver barrier to Wait() for everyone else on
-	Msg      string // some words for fun
-	Activity bool   // did something interesting happen?
-}
-
-// send a single acknowledgement on the channel if one was requested
-func (event *Event) ACK() {
-	if event.Resp != nil { // if they've requested an ACK
-		event.Resp <- true // send ACK
-	}
-}
-
-func (event *Event) NACK() {
-	if event.Resp != nil { // if they've requested an ACK
-		event.Resp <- false // send NACK
-	}
-}
-
-// Resp is just a helper to return the right type of response channel
-func NewResp() Resp {
-	resp := make(chan bool)
-	return resp
-}
-
-// ACKWait waits for a +ive Ack from a Resp channel
-func (resp Resp) ACKWait() {
-	for {
-		value := <-resp
-		// wait until true value
-		if value {
-			return
-		}
-	}
-}
-
-// get the activity value
-func (event *Event) GetActivity() bool {
-	return event.Activity
-}

event/event.go

@@ -0,0 +1,120 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package event provides some primitives that are used for message passing.
+package event
+
+import (
+	"fmt"
+)
+
+//go:generate stringer -type=EventName -output=eventname_stringer.go
+
+// EventName represents the type of event being passed.
+type EventName int
+
+// The different event names are used in different contexts.
+const (
+	EventNil EventName = iota
+	EventExit
+	EventStart
+	EventPause
+	EventPoke
+	EventBackPoke
+)
+
+// Resp is a channel to be used for boolean responses. A nil represents an ACK,
+// and a non-nil represents a NACK (false). This also lets us use custom errors.
+type Resp chan error
+
+// Event is the main struct that stores event information and responses.
+type Event struct {
+	Name EventName
+	Resp Resp // channel to send an ack response on, nil to skip
+	//Wg   *sync.WaitGroup // receiver barrier to Wait() for everyone else on
+	Msg      string // some words for fun
+	Activity bool   // did something interesting happen?
+}
+
+// ACK sends a single acknowledgement on the channel if one was requested.
+func (event *Event) ACK() {
+	if event.Resp != nil { // if they've requested an ACK
+		event.Resp.ACK()
+	}
+}
+
+// NACK sends a negative acknowledgement message on the channel if one was requested.
+func (event *Event) NACK() {
+	if event.Resp != nil { // if they've requested a NACK
+		event.Resp.NACK()
+	}
+}
+
+// ACKNACK sends a custom ACK or NACK message on the channel if one was requested.
+func (event *Event) ACKNACK(err error) {
+	if event.Resp != nil { // if they've requested a NACK
+		event.Resp.ACKNACK(err)
+	}
+}
+
+// NewResp is just a helper to return the right type of response channel.
+func NewResp() Resp {
+	resp := make(chan error)
+	return resp
+}
+
+// ACK sends a true value to resp.
+func (resp Resp) ACK() {
+	if resp != nil {
+		resp <- nil
+	}
+}
+
+// NACK sends a false value to resp.
+func (resp Resp) NACK() {
+	if resp != nil {
+		resp <- fmt.Errorf("NACK")
+	}
+}
+
+// ACKNACK sends a custom ACK or NACK. The ACK value is always nil, the NACK can
+// be any non-nil error value.
+func (resp Resp) ACKNACK(err error) {
+	if resp != nil {
+		resp <- err
+	}
+}
+
+// Wait waits for any response from a Resp channel and returns it.
+func (resp Resp) Wait() error {
+	return <-resp
+}
+
+// ACKWait waits for a +ive Ack from a Resp channel.
+func (resp Resp) ACKWait() {
+	for {
+		// wait until true value
+		if resp.Wait() == nil {
+			return
+		}
+	}
+}
+
+// GetActivity returns the activity value.
+func (event *Event) GetActivity() bool {
+	return event.Activity
+}

examples/autogroup2.yaml

@@ -0,0 +1,17 @@
+---
+graph: mygraph
+resources:
+  pkg:
+  - name: powertop
+    meta:
+      autogroup: true
+    state: installed
+  - name: sl
+    meta:
+      autogroup: true
+    state: installed
+  - name: cowsay
+    meta:
+      autogroup: true
+    state: installed
+edges: []

examples/etcd1d.yaml

@@ -0,0 +1,18 @@
+---
+graph: mygraph
+resources:
+  file:
+  - name: file1d
+    path: "/tmp/mgmtD/f1d"
+    content: |
+      i am f1
+    state: exists
+  - name: "@@file2d"
+    path: "/tmp/mgmtD/f2d"
+    content: |
+      i am f2, exported from host D
+    state: exists
+collect:
+- kind: file
+  pattern: "/tmp/mgmtD/"
+edges: []

examples/exec3.yaml

@@ -0,0 +1,59 @@
+---
+graph: parallel
+resources:
+  exec:
+  - name: pkg10
+    cmd: sleep 10s
+    shell: ''
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+  - name: svc10
+    cmd: sleep 10s
+    shell: ''
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+  - name: exec10
+    cmd: sleep 10s
+    shell: ''
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+  - name: pkg15
+    cmd: sleep 15s
+    shell: ''
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+edges:
+- name: e1
+  from:
+    kind: exec
+    name: pkg10
+  to:
+    kind: exec
+    name: svc10
+- name: e2
+  from:
+    kind: exec
+    name: svc10
+  to:
+    kind: exec
+    name: exec10

examples/file2.yaml

@@ -0,0 +1,13 @@
+---
+graph: mygraph
+resources:
+  noop:
+  - name: noop1
+  file:
+  - name: file1
+    path: "/tmp/mgmt/hello/"
+    source: "/var/lib/mgmt/files/some_dir/"
+    recurse: true
+    force: true
+    state: exists
+edges: []

examples/file3.yaml

@@ -0,0 +1,14 @@
+---
+graph: mygraph
+comment: You can test Watch and CheckApply failures with chmod ugo-r and chmod ugo-w.
+resources:
+  file:
+  - name: file1
+    path: "/tmp/mgmt/f1"
+    meta:
+      retry: 3
+      delay: 5000
+    content: |
+      i am f1
+    state: exists
+edges: []

examples/lib/libmgmt1.go

@@ -0,0 +1,188 @@
+// libmgmt example
+package main
+
+import (
+	"fmt"
+	"log"
+	"os"
+	"os/signal"
+	"sync"
+	"syscall"
+	"time"
+
+	"github.com/purpleidea/mgmt/gapi"
+	mgmt "github.com/purpleidea/mgmt/mgmtmain"
+	"github.com/purpleidea/mgmt/pgraph"
+	"github.com/purpleidea/mgmt/resources"
+	"github.com/purpleidea/mgmt/yamlgraph"
+)
+
+// MyGAPI implements the main GAPI interface.
+type MyGAPI struct {
+	Name     string // graph name
+	Interval uint   // refresh interval, 0 to never refresh
+
+	data        gapi.Data
+	initialized bool
+	closeChan   chan struct{}
+	wg          sync.WaitGroup // sync group for tunnel go routines
+}
+
+// NewMyGAPI creates a new MyGAPI struct and calls Init().
+func NewMyGAPI(data gapi.Data, name string, interval uint) (*MyGAPI, error) {
+	obj := &MyGAPI{
+		Name:     name,
+		Interval: interval,
+	}
+	return obj, obj.Init(data)
+}
+
+// Init initializes the MyGAPI struct.
+func (obj *MyGAPI) Init(data gapi.Data) error {
+	if obj.initialized {
+		return fmt.Errorf("Already initialized!")
+	}
+	if obj.Name == "" {
+		return fmt.Errorf("The graph name must be specified!")
+	}
+	obj.data = data // store for later
+	obj.closeChan = make(chan struct{})
+	obj.initialized = true
+	return nil
+}
+
+// Graph returns a current Graph.
+func (obj *MyGAPI) Graph() (*pgraph.Graph, error) {
+	if !obj.initialized {
+		return nil, fmt.Errorf("libmgmt: MyGAPI is not initialized")
+	}
+
+	n1, err := resources.NewNoopRes("noop1")
+	if err != nil {
+		return nil, fmt.Errorf("Can't create resource: %v", err)
+	}
+
+	// we can still build a graph via the yaml method
+	gc := &yamlgraph.GraphConfig{
+		Graph: obj.Name,
+		Resources: yamlgraph.Resources{ // must redefine anonymous struct :(
+			// in alphabetical order
+			Exec:  []*resources.ExecRes{},
+			File:  []*resources.FileRes{},
+			Msg:   []*resources.MsgRes{},
+			Noop:  []*resources.NoopRes{n1},
+			Pkg:   []*resources.PkgRes{},
+			Svc:   []*resources.SvcRes{},
+			Timer: []*resources.TimerRes{},
+			Virt:  []*resources.VirtRes{},
+		},
+		//Collector: []collectorResConfig{},
+		//Edges:     []Edge{},
+		Comment: "comment!",
+	}
+
+	g, err := gc.NewGraphFromConfig(obj.data.Hostname, obj.data.EmbdEtcd, obj.data.Noop)
+	return g, err
+}
+
+// SwitchStream returns nil errors every time there could be a new graph.
+func (obj *MyGAPI) SwitchStream() chan error {
+	if obj.data.NoWatch || obj.Interval <= 0 {
+		return nil
+	}
+	ch := make(chan error)
+	obj.wg.Add(1)
+	go func() {
+		defer obj.wg.Done()
+		defer close(ch) // this will run before the obj.wg.Done()
+		if !obj.initialized {
+			ch <- fmt.Errorf("libmgmt: MyGAPI is not initialized")
+			return
+		}
+
+		// arbitrarily change graph every interval seconds
+		ticker := time.NewTicker(time.Duration(obj.Interval) * time.Second)
+		defer ticker.Stop()
+		for {
+			select {
+			case <-ticker.C:
+				log.Printf("libmgmt: Generating new graph...")
+				ch <- nil // trigger a run
+			case <-obj.closeChan:
+				return
+			}
+		}
+	}()
+	return ch
+}
+
+// Close shuts down the MyGAPI.
+func (obj *MyGAPI) Close() error {
+	if !obj.initialized {
+		return fmt.Errorf("libmgmt: MyGAPI is not initialized")
+	}
+	close(obj.closeChan)
+	obj.wg.Wait()
+	obj.initialized = false // closed = true
+	return nil
+}
+
+// Run runs an embedded mgmt server.
+func Run() error {
+
+	obj := &mgmt.Main{}
+	obj.Program = "libmgmt" // TODO: set on compilation
+	obj.Version = "0.0.1"   // TODO: set on compilation
+	obj.TmpPrefix = true
+	obj.IdealClusterSize = -1
+	obj.ConvergedTimeout = -1
+	obj.Noop = true
+
+	obj.GAPI = &MyGAPI{ // graph API
+		Name:     "libmgmt", // TODO: set on compilation
+		Interval: 15,        // arbitrarily change graph every 15 seconds
+	}
+
+	if err := obj.Init(); err != nil {
+		return err
+	}
+
+	// install the exit signal handler
+	exit := make(chan struct{})
+	defer close(exit)
+	go func() {
+		signals := make(chan os.Signal, 1)
+		signal.Notify(signals, os.Interrupt) // catch ^C
+		//signal.Notify(signals, os.Kill) // catch signals
+		signal.Notify(signals, syscall.SIGTERM)
+
+		select {
+		case sig := <-signals: // any signal will do
+			if sig == os.Interrupt {
+				log.Println("Interrupted by ^C")
+				obj.Exit(nil)
+				return
+			}
+			log.Println("Interrupted by signal")
+			obj.Exit(fmt.Errorf("Killed by %v", sig))
+			return
+		case <-exit:
+			return
+		}
+	}()
+
+	if err := obj.Run(); err != nil {
+		return err
+	}
+	return nil
+}
+
+func main() {
+	log.Printf("Hello!")
+	if err := Run(); err != nil {
+		fmt.Println(err)
+		os.Exit(1)
+		return
+	}
+	log.Printf("Goodbye!")
+}

examples/lib/libmgmt2.go

@@ -0,0 +1,188 @@
+// libmgmt example
+package main
+
+import (
+	"fmt"
+	"log"
+	"os"
+	"os/signal"
+	"strconv"
+	"sync"
+	"syscall"
+	"time"
+
+	"github.com/purpleidea/mgmt/gapi"
+	mgmt "github.com/purpleidea/mgmt/mgmtmain"
+	"github.com/purpleidea/mgmt/pgraph"
+	"github.com/purpleidea/mgmt/resources"
+)
+
+// MyGAPI implements the main GAPI interface.
+type MyGAPI struct {
+	Name     string // graph name
+	Count    uint   // number of resources to create
+	Interval uint   // refresh interval, 0 to never refresh
+
+	data        gapi.Data
+	initialized bool
+	closeChan   chan struct{}
+	wg          sync.WaitGroup // sync group for tunnel go routines
+}
+
+// NewMyGAPI creates a new MyGAPI struct and calls Init().
+func NewMyGAPI(data gapi.Data, name string, interval uint, count uint) (*MyGAPI, error) {
+	obj := &MyGAPI{
+		Name:     name,
+		Count:    count,
+		Interval: interval,
+	}
+	return obj, obj.Init(data)
+}
+
+// Init initializes the MyGAPI struct.
+func (obj *MyGAPI) Init(data gapi.Data) error {
+	if obj.initialized {
+		return fmt.Errorf("Already initialized!")
+	}
+	if obj.Name == "" {
+		return fmt.Errorf("The graph name must be specified!")
+	}
+	obj.data = data // store for later
+	obj.closeChan = make(chan struct{})
+	obj.initialized = true
+	return nil
+}
+
+// Graph returns a current Graph.
+func (obj *MyGAPI) Graph() (*pgraph.Graph, error) {
+	if !obj.initialized {
+		return nil, fmt.Errorf("libmgmt: MyGAPI is not initialized")
+	}
+
+	g := pgraph.NewGraph(obj.Name)
+	var vertex *pgraph.Vertex
+	for i := uint(0); i < obj.Count; i++ {
+		n, err := resources.NewNoopRes(fmt.Sprintf("noop%d", i))
+		if err != nil {
+			return nil, fmt.Errorf("Can't create resource: %v", err)
+		}
+		v := pgraph.NewVertex(n)
+		g.AddVertex(v)
+		if i > 0 {
+			g.AddEdge(vertex, v, pgraph.NewEdge(fmt.Sprintf("e%d", i)))
+		}
+		vertex = v // save
+	}
+
+	//g, err := config.NewGraphFromConfig(obj.data.Hostname, obj.data.EmbdEtcd, obj.data.Noop)
+	return g, nil
+}
+
+// SwitchStream returns nil errors every time there could be a new graph.
+func (obj *MyGAPI) SwitchStream() chan error {
+	if obj.data.NoWatch || obj.Interval <= 0 {
+		return nil
+	}
+	ch := make(chan error)
+	obj.wg.Add(1)
+	go func() {
+		defer obj.wg.Done()
+		defer close(ch) // this will run before the obj.wg.Done()
+		if !obj.initialized {
+			ch <- fmt.Errorf("libmgmt: MyGAPI is not initialized")
+			return
+		}
+
+		// arbitrarily change graph every interval seconds
+		ticker := time.NewTicker(time.Duration(obj.Interval) * time.Second)
+		defer ticker.Stop()
+		for {
+			select {
+			case <-ticker.C:
+				log.Printf("libmgmt: Generating new graph...")
+				ch <- nil // trigger a run
+			case <-obj.closeChan:
+				return
+			}
+		}
+	}()
+	return ch
+}
+
+// Close shuts down the MyGAPI.
+func (obj *MyGAPI) Close() error {
+	if !obj.initialized {
+		return fmt.Errorf("libmgmt: MyGAPI is not initialized")
+	}
+	close(obj.closeChan)
+	obj.wg.Wait()
+	obj.initialized = false // closed = true
+	return nil
+}
+
+// Run runs an embedded mgmt server.
+func Run(count uint) error {
+
+	obj := &mgmt.Main{}
+	obj.Program = "libmgmt" // TODO: set on compilation
+	obj.Version = "0.0.1"   // TODO: set on compilation
+	obj.TmpPrefix = true
+	obj.IdealClusterSize = -1
+	obj.ConvergedTimeout = -1
+	obj.Noop = true
+
+	obj.GAPI = &MyGAPI{ // graph API
+		Name:     "libmgmt", // TODO: set on compilation
+		Count:    count,     // number of vertices to add
+		Interval: 15,        // arbitrarily change graph every 15 seconds
+	}
+
+	if err := obj.Init(); err != nil {
+		return err
+	}
+
+	// install the exit signal handler
+	exit := make(chan struct{})
+	defer close(exit)
+	go func() {
+		signals := make(chan os.Signal, 1)
+		signal.Notify(signals, os.Interrupt) // catch ^C
+		//signal.Notify(signals, os.Kill) // catch signals
+		signal.Notify(signals, syscall.SIGTERM)
+
+		select {
+		case sig := <-signals: // any signal will do
+			if sig == os.Interrupt {
+				log.Println("Interrupted by ^C")
+				obj.Exit(nil)
+				return
+			}
+			log.Println("Interrupted by signal")
+			obj.Exit(fmt.Errorf("Killed by %v", sig))
+			return
+		case <-exit:
+			return
+		}
+	}()
+
+	if err := obj.Run(); err != nil {
+		return err
+	}
+	return nil
+}
+
+func main() {
+	log.Printf("Hello!")
+	var count uint = 1 // default
+	if len(os.Args) == 2 {
+		if i, err := strconv.Atoi(os.Args[1]); err == nil && i > 0 {
+			count = uint(i)
+		}
+	}
+	if err := Run(count); err != nil {
+		fmt.Println(err)
+		os.Exit(1)
+		return
+	}
+	log.Printf("Goodbye!")
+}

examples/msg1.yaml

@@ -0,0 +1,19 @@
+---
+graph: mygraph
+comment: timer example
+resources:
+  timer:
+  - name: timer1
+    interval: 30
+  msg:
+  - name: msg1
+    body: mgmt logged this message
+    journal: true
+edges:
+- name: e1
+  from:
+    kind: timer
+    name: timer1
+  to:
+    kind: msg
+    name: msg1

examples/noop1.yaml

@@ -0,0 +1,24 @@
+---
+graph: mygraph
+comment: noop example
+resources:
+  noop:
+  - name: noop1
+    meta:
+      noop: true
+  file:
+  - name: file1
+    path: "/tmp/mgmt-hello-noop"
+    content: |
+      hello world from @purpleidea
+    state: exists
+    meta:
+      noop: true
+edges:
+- name: e1
+  from:
+    kind: noop
+    name: noop1
+  to:
+    kind: file
+    name: file1

examples/remote1.yaml

@@ -0,0 +1,23 @@
+---
+graph: mygraph
+comment: remote noop example
+resources:
+  noop:
+  - name: noop1
+    meta:
+      noop: true
+  file:
+  - name: file1
+    path: "/tmp/mgmt-remote-hello"
+    content: |
+      hello world from @purpleidea
+    state: exists
+edges:
+- name: e1
+  from:
+    kind: noop
+    name: noop1
+  to:
+    kind: file
+    name: file1
+remote: "ssh://root:password@hostname:22"

examples/remote2a.yaml

@@ -0,0 +1,20 @@
+---
+graph: mygraph
+comment: remote noop example
+resources:
+  file:
+  - name: file1a
+    path: "/tmp/file1a"
+    content: |
+      i am file1a
+    state: exists
+  - name: "@@file2a"
+    path: "/tmp/file2a"
+    content: |
+      i am file2a, exported from host a
+    state: exists
+collect:
+- kind: file
+  pattern: "/tmp/"
+edges: []
+remote: ssh://root:vagrant@192.168.121.201:22

examples/remote2b.yaml

@@ -0,0 +1,20 @@
+---
+graph: mygraph
+comment: remote noop example
+resources:
+  file:
+  - name: file1b
+    path: "/tmp/file1b"
+    content: |
+      i am file1b
+    state: exists
+  - name: "@@file2b"
+    path: "/tmp/file2b"
+    content: |
+      i am file2b, exported from host b
+    state: exists
+collect:
+- kind: file
+  pattern: "/tmp/"
+edges: []
+remote: ssh://root:vagrant@192.168.121.202:22

examples/timer1.yaml

@@ -0,0 +1,25 @@
+---
+graph: mygraph
+comment: timer example
+resources:
+  timer:
+  - name: timer1
+    interval: 30
+  exec:
+  - name: exec1
+    cmd: echo hello world
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+edges:
+- name: e1
+  from:
+    kind: timer
+    name: timer1
+  to:
+    kind: exec
+    name: exec1

examples/timer2.yaml

@@ -0,0 +1,43 @@
+---
+graph: mygraph
+comment: example of multiple timers
+resources:
+  timer:
+  - name: timer1
+    interval: 30
+  - name: timer2
+    interval: 60
+  exec:
+  - name: exec1
+    cmd: echo hello world 30
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+  - name: exec2
+    cmd: echo hello world 60
+    timeout: 0
+    watchcmd: ''
+    watchshell: ''
+    ifcmd: ''
+    ifshell: ''
+    pollint: 0
+    state: present
+edges:
+- name: e1
+  from:
+    kind: timer
+    name: timer1
+  to:
+    kind: exec
+    name: exec1
+- name: e2
+  from:
+    kind: timer
+    name: timer2
+  to:
+    kind: exec
+    name: exec2

examples/virt1.yaml

@@ -0,0 +1,11 @@
+---
+graph: mygraph
+resources:
+  virt:
+  - name: mgmt1
+    uri: 'qemu:///session'
+    cpus: 1
+    memory: 524288
+    state: shutoff
+    transient: true
+edges: []

examples/virt2.yaml

@@ -0,0 +1,11 @@
+---
+graph: mygraph
+resources:
+  virt:
+  - name: mgmt2
+    uri: 'qemu:///session'
+    cpus: 1
+    memory: 524288
+    state: shutoff
+    transient: false
+edges: []

examples/virt3.yaml

@@ -0,0 +1,11 @@
+---
+graph: mygraph
+resources:
+  virt:
+  - name: mgmt3
+    uri: 'qemu:///session'
+    cpus: 1
+    memory: 524288
+    state: running
+    transient: false
+edges: []

file.go

@@ -1,502 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-import (
-	"crypto/sha256"
-	"encoding/hex"
-	"gopkg.in/fsnotify.v1"
-	//"github.com/go-fsnotify/fsnotify" // git master of "gopkg.in/fsnotify.v1"
-	"encoding/gob"
-	"io"
-	"log"
-	"math"
-	"os"
-	"path"
-	"strings"
-	"syscall"
-)
-
-func init() {
-	gob.Register(&FileRes{})
-}
-
-type FileRes struct {
-	BaseRes   `yaml:",inline"`
-	Path      string `yaml:"path"` // path variable (should default to name)
-	Dirname   string `yaml:"dirname"`
-	Basename  string `yaml:"basename"`
-	Content   string `yaml:"content"`
-	State     string `yaml:"state"` // state: exists/present?, absent, (undefined?)
-	sha256sum string
-}
-
-func NewFileRes(name, path, dirname, basename, content, state string) *FileRes {
-	// FIXME if path = nil, path = name ...
-	obj := &FileRes{
-		BaseRes: BaseRes{
-			Name: name,
-		},
-		Path:      path,
-		Dirname:   dirname,
-		Basename:  basename,
-		Content:   content,
-		State:     state,
-		sha256sum: "",
-	}
-	obj.Init()
-	return obj
-}
-
-func (obj *FileRes) Init() {
-	obj.BaseRes.kind = "File"
-	obj.BaseRes.Init() // call base init, b/c we're overriding
-}
-
-func (obj *FileRes) GetPath() string {
-	d := Dirname(obj.Path)
-	b := Basename(obj.Path)
-	if !obj.Validate() || (obj.Dirname == "" && obj.Basename == "") {
-		return obj.Path
-	} else if obj.Dirname == "" {
-		return d + obj.Basename
-	} else if obj.Basename == "" {
-		return obj.Dirname + b
-	} else { // if obj.dirname != "" && obj.basename != "" {
-		return obj.Dirname + obj.Basename
-	}
-}
-
-// validate if the params passed in are valid data
-func (obj *FileRes) Validate() bool {
-	if obj.Dirname != "" {
-		// must end with /
-		if obj.Dirname[len(obj.Dirname)-1:] != "/" {
-			return false
-		}
-	}
-	if obj.Basename != "" {
-		// must not start with /
-		if obj.Basename[0:1] == "/" {
-			return false
-		}
-	}
-	return true
-}
-
-// File watcher for files and directories
-// Modify with caution, probably important to write some test cases first!
-// obj.GetPath(): file or directory
-func (obj *FileRes) Watch(processChan chan Event) {
-	if obj.IsWatching() {
-		return
-	}
-	obj.SetWatching(true)
-	defer obj.SetWatching(false)
-
-	//var recursive bool = false
-	//var isdir = (obj.GetPath()[len(obj.GetPath())-1:] == "/") // dirs have trailing slashes
-	//log.Printf("IsDirectory: %v", isdir)
-	//vertex := obj.GetVertex()         // stored with SetVertex
-	var safename = path.Clean(obj.GetPath()) // no trailing slash
-
-	watcher, err := fsnotify.NewWatcher()
-	if err != nil {
-		log.Fatal(err)
-	}
-	defer watcher.Close()
-
-	patharray := PathSplit(safename) // tokenize the path
-	var index = len(patharray)       // starting index
-	var current string               // current "watcher" location
-	var deltaDepth int               // depth delta between watcher and event
-	var send = false                 // send event?
-	var exit = false
-	var dirty = false
-
-	for {
-		current = strings.Join(patharray[0:index], "/")
-		if current == "" { // the empty string top is the root dir ("/")
-			current = "/"
-		}
-		if DEBUG {
-			log.Printf("File[%v]: Watching: %v", obj.GetName(), current) // attempting to watch...
-		}
-		// initialize in the loop so that we can reset on rm-ed handles
-		err = watcher.Add(current)
-		if err != nil {
-			if DEBUG {
-				log.Printf("File[%v]: watcher.Add(%v): Error: %v", obj.GetName(), current, err)
-			}
-			if err == syscall.ENOENT {
-				index-- // usually not found, move up one dir
-			} else if err == syscall.ENOSPC {
-				// XXX: occasionally: no space left on device,
-				// XXX: probably due to lack of inotify watches
-				log.Printf("%v[%v]: Out of inotify watches!", obj.Kind(), obj.GetName())
-				log.Fatal(err)
-			} else {
-				log.Printf("Unknown file[%v] error:", obj.Name)
-				log.Fatal(err)
-			}
-			index = int(math.Max(1, float64(index)))
-			continue
-		}
-
-		obj.SetState(resStateWatching) // reset
-		select {
-		case event := <-watcher.Events:
-			if DEBUG {
-				log.Printf("File[%v]: Watch(%v), Event(%v): %v", obj.GetName(), current, event.Name, event.Op)
-			}
-			obj.SetConvergedState(resConvergedNil) // XXX: technically i can detect if the event is erroneous or not first
-			// the deeper you go, the bigger the deltaDepth is...
-			// this is the difference between what we're watching,
-			// and the event... doesn't mean we can't watch deeper
-			if current == event.Name {
-				deltaDepth = 0 // i was watching what i was looking for
-
-			} else if HasPathPrefix(event.Name, current) {
-				deltaDepth = len(PathSplit(current)) - len(PathSplit(event.Name)) // -1 or less
-
-			} else if HasPathPrefix(current, event.Name) {
-				deltaDepth = len(PathSplit(event.Name)) - len(PathSplit(current)) // +1 or more
-
-			} else {
-				// TODO different watchers get each others events!
-				// https://github.com/go-fsnotify/fsnotify/issues/95
-				// this happened with two values such as:
-				// event.Name: /tmp/mgmt/f3 and current: /tmp/mgmt/f2
-				continue
-			}
-			//log.Printf("The delta depth is: %v", deltaDepth)
-
-			// if we have what we wanted, awesome, send an event...
-			if event.Name == safename {
-				//log.Println("Event!")
-				// FIXME: should all these below cases trigger?
-				send = true
-				dirty = true
-
-				// file removed, move the watch upwards
-				if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
-					//log.Println("Removal!")
-					watcher.Remove(current)
-					index--
-				}
-
-				// we must be a parent watcher, so descend in
-				if deltaDepth < 0 {
-					watcher.Remove(current)
-					index++
-				}
-
-				// if safename starts with event.Name, we're above, and no event should be sent
-			} else if HasPathPrefix(safename, event.Name) {
-				//log.Println("Above!")
-
-				if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
-					log.Println("Removal!")
-					watcher.Remove(current)
-					index--
-				}
-
-				if deltaDepth < 0 {
-					log.Println("Parent!")
-					if PathPrefixDelta(safename, event.Name) == 1 { // we're the parent dir
-						send = true
-						dirty = true
-					}
-					watcher.Remove(current)
-					index++
-				}
-
-				// if event.Name startswith safename, send event, we're already deeper
-			} else if HasPathPrefix(event.Name, safename) {
-				//log.Println("Event2!")
-				send = true
-				dirty = true
-			}
-
-		case err := <-watcher.Errors:
-			obj.SetConvergedState(resConvergedNil) // XXX ?
-			log.Printf("error: %v", err)
-			log.Fatal(err)
-			//obj.events <- fmt.Sprintf("file: %v", "error") // XXX: how should we handle errors?
-
-		case event := <-obj.events:
-			obj.SetConvergedState(resConvergedNil)
-			if exit, send = obj.ReadEvent(&event); exit {
-				return // exit
-			}
-			//dirty = false // these events don't invalidate state
-
-		case _ = <-TimeAfterOrBlock(obj.ctimeout):
-			obj.SetConvergedState(resConvergedTimeout)
-			obj.converged <- true
-			continue
-		}
-
-		// do all our event sending all together to avoid duplicate msgs
-		if send {
-			send = false
-			// only invalid state on certain types of events
-			if dirty {
-				dirty = false
-				obj.isStateOK = false // something made state dirty
-			}
-			resp := NewResp()
-			processChan <- Event{eventNil, resp, "", true} // trigger process
-			resp.ACKWait()                                 // wait for the ACK()
-		}
-	}
-}
-
-func (obj *FileRes) HashSHA256fromContent() string {
-	if obj.sha256sum != "" { // return if already computed
-		return obj.sha256sum
-	}
-
-	hash := sha256.New()
-	hash.Write([]byte(obj.Content))
-	obj.sha256sum = hex.EncodeToString(hash.Sum(nil))
-	return obj.sha256sum
-}
-
-func (obj *FileRes) FileHashSHA256Check() (bool, error) {
-	if PathIsDir(obj.GetPath()) { // assert
-		log.Fatal("This should only be called on a File resource.")
-	}
-	// run a diff, and return true if it needs changing
-	hash := sha256.New()
-	f, err := os.Open(obj.GetPath())
-	if err != nil {
-		if e, ok := err.(*os.PathError); ok && (e.Err.(syscall.Errno) == syscall.ENOENT) {
-			return false, nil // no "error", file is just absent
-		}
-		return false, err
-	}
-	defer f.Close()
-	if _, err := io.Copy(hash, f); err != nil {
-		return false, err
-	}
-	sha256sum := hex.EncodeToString(hash.Sum(nil))
-	//log.Printf("sha256sum: %v", sha256sum)
-	if obj.HashSHA256fromContent() == sha256sum {
-		return true, nil
-	}
-	return false, nil
-}
-
-func (obj *FileRes) FileApply() error {
-	if PathIsDir(obj.GetPath()) {
-		log.Fatal("This should only be called on a File resource.")
-	}
-
-	if obj.State == "absent" {
-		log.Printf("About to remove: %v", obj.GetPath())
-		err := os.Remove(obj.GetPath())
-		return err // either nil or not, for success or failure
-	}
-
-	f, err := os.Create(obj.GetPath())
-	if err != nil {
-		return nil
-	}
-	defer f.Close()
-
-	_, err = io.WriteString(f, obj.Content)
-	if err != nil {
-		return err
-	}
-
-	return nil // success
-}
-
-func (obj *FileRes) CheckApply(apply bool) (stateok bool, err error) {
-	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
-
-	if obj.isStateOK { // cache the state
-		return true, nil
-	}
-
-	if _, err = os.Stat(obj.GetPath()); os.IsNotExist(err) {
-		// no such file or directory
-		if obj.State == "absent" {
-			// missing file should be missing, phew :)
-			obj.isStateOK = true
-			return true, nil
-		}
-	}
-	err = nil // reset
-
-	// FIXME: add file mode check here...
-
-	if PathIsDir(obj.GetPath()) {
-		log.Fatal("Not implemented!") // XXX
-	} else {
-		ok, err := obj.FileHashSHA256Check()
-		if err != nil {
-			return false, err
-		}
-		if ok {
-			obj.isStateOK = true
-			return true, nil
-		}
-		// if no err, but !ok, then we continue on...
-	}
-
-	// state is not okay, no work done, exit, but without error
-	if !apply {
-		return false, nil
-	}
-
-	// apply portion
-	log.Printf("%v[%v]: Apply", obj.Kind(), obj.GetName())
-	if PathIsDir(obj.GetPath()) {
-		log.Fatal("Not implemented!") // XXX
-	} else {
-		err = obj.FileApply()
-		if err != nil {
-			return false, err
-		}
-	}
-
-	obj.isStateOK = true
-	return false, nil // success
-}
-
-type FileUUID struct {
-	BaseUUID
-	path string
-}
-
-// if and only if they are equivalent, return true
-// if they are not equivalent, return false
-func (obj *FileUUID) IFF(uuid ResUUID) bool {
-	res, ok := uuid.(*FileUUID)
-	if !ok {
-		return false
-	}
-	return obj.path == res.path
-}
-
-type FileResAutoEdges struct {
-	data    []ResUUID
-	pointer int
-	found   bool
-}
-
-func (obj *FileResAutoEdges) Next() []ResUUID {
-	if obj.found {
-		log.Fatal("Shouldn't be called anymore!")
-	}
-	if len(obj.data) == 0 { // check length for rare scenarios
-		return nil
-	}
-	value := obj.data[obj.pointer]
-	obj.pointer++
-	return []ResUUID{value} // we return one, even though api supports N
-}
-
-// get results of the earlier Next() call, return if we should continue!
-func (obj *FileResAutoEdges) Test(input []bool) bool {
-	// if there aren't any more remaining
-	if len(obj.data) <= obj.pointer {
-		return false
-	}
-	if obj.found { // already found, done!
-		return false
-	}
-	if len(input) != 1 { // in case we get given bad data
-		log.Fatal("Expecting a single value!")
-	}
-	if input[0] { // if a match is found, we're done!
-		obj.found = true // no more to find!
-		return false
-	}
-	return true // keep going
-}
-
-// generate a simple linear sequence of each parent directory from bottom up!
-func (obj *FileRes) AutoEdges() AutoEdge {
-	var data []ResUUID                             // store linear result chain here...
-	values := PathSplitFullReversed(obj.GetPath()) // build it
-	_, values = values[0], values[1:]              // get rid of first value which is me!
-	for _, x := range values {
-		var reversed = true // cheat by passing a pointer
-		data = append(data, &FileUUID{
-			BaseUUID: BaseUUID{
-				name:     obj.GetName(),
-				kind:     obj.Kind(),
-				reversed: &reversed,
-			},
-			path: x, // what matters
-		}) // build list
-	}
-	return &FileResAutoEdges{
-		data:    data,
-		pointer: 0,
-		found:   false,
-	}
-}
-
-func (obj *FileRes) GetUUIDs() []ResUUID {
-	x := &FileUUID{
-		BaseUUID: BaseUUID{name: obj.GetName(), kind: obj.Kind()},
-		path:     obj.GetPath(),
-	}
-	return []ResUUID{x}
-}
-
-func (obj *FileRes) GroupCmp(r Res) bool {
-	_, ok := r.(*FileRes)
-	if !ok {
-		return false
-	}
-	// TODO: we might be able to group directory children into a single
-	// recursive watcher in the future, thus saving fanotify watches
-	return false // not possible atm
-}
-
-func (obj *FileRes) Compare(res Res) bool {
-	switch res.(type) {
-	case *FileRes:
-		res := res.(*FileRes)
-		if obj.Name != res.Name {
-			return false
-		}
-		if obj.GetPath() != res.Path {
-			return false
-		}
-		if obj.Content != res.Content {
-			return false
-		}
-		if obj.State != res.State {
-			return false
-		}
-	default:
-		return false
-	}
-	return true
-}
-
-func (obj *FileRes) CollectPattern(pattern string) {
-	// XXX: currently the pattern for files can only override the Dirname variable :P
-	obj.Dirname = pattern // XXX: simplistic for now
-}

gapi/gapi.go

@@ -0,0 +1,41 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package gapi defines the interface that graph API generators must meet.
+package gapi
+
+import (
+	"github.com/purpleidea/mgmt/etcd"
+	"github.com/purpleidea/mgmt/pgraph"
+)
+
+// Data is the set of input values passed into the GAPI structs via Init.
+type Data struct {
+	Hostname string // uuid for the host, required for GAPI
+	EmbdEtcd *etcd.EmbdEtcd
+	Noop     bool
+	NoWatch  bool
+	// NOTE: we can add more fields here if needed by GAPI endpoints
+}
+
+// GAPI is a Graph API that represents incoming graphs and change streams.
+type GAPI interface {
+	Init(Data) error               // initializes the GAPI and passes in useful data
+	Graph() (*pgraph.Graph, error) // returns the most recent pgraph
+	SwitchStream() chan error      // returns a stream of switch events
+	Close() error                  // shutdown the GAPI
+}

global/global.go

@@ -0,0 +1,26 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package global holds some global variables that are used throughout the code.
+package global
+
+// These constants are used throughout the program.
+const (
+	DEBUG   = false // add additional log messages
+	TRACE   = false // add execution flow log messages
+	VERBOSE = false // add extra log message output
+)

gopath/.gitignore

@@ -0,0 +1,2 @@
+bin/
+pkg/

gopath/src

@@ -0,0 +1 @@
+../vendor

main.go

@@ -18,13 +18,10 @@
 package main
 
 import (
-	"github.com/codegangsta/cli"
+	"fmt"
-	"log"
 	"os"
-	"os/signal"
+
-	"sync"
+	"github.com/purpleidea/mgmt/mgmtmain"
-	"syscall"
-	"time"
 )
 
 // set at compile time
@@ -33,262 +30,10 @@ var (
 	version string
 )
 
-const (
-	DEBUG = false
-)
-
-// signal handler
-func waitForSignal(exit chan bool) {
-	signals := make(chan os.Signal, 1)
-	signal.Notify(signals, os.Interrupt) // catch ^C
-	//signal.Notify(signals, os.Kill) // catch signals
-	signal.Notify(signals, syscall.SIGTERM)
-
-	select {
-	case e := <-signals: // any signal will do
-		if e == os.Interrupt {
-			log.Println("Interrupted by ^C")
-		} else {
-			log.Println("Interrupted by signal")
-		}
-	case <-exit: // or a manual signal
-		log.Println("Interrupted by exit signal")
-	}
-}
-
-func run(c *cli.Context) {
-	var start = time.Now().UnixNano()
-	var wg sync.WaitGroup
-	exit := make(chan bool)      // exit signal
-	converged := make(chan bool) // converged signal
-	log.Printf("This is: %v, version: %v", program, version)
-	log.Printf("Main: Start: %v", start)
-	var G, fullGraph *Graph
-
-	// exit after `max-runtime` seconds for no reason at all...
-	if i := c.Int("max-runtime"); i > 0 {
-		go func() {
-			time.Sleep(time.Duration(i) * time.Second)
-			exit <- true
-		}()
-	}
-
-	// initial etcd peer endpoint
-	seed := c.String("seed")
-	if seed == "" {
-		// XXX: start up etcd server, others will join me!
-		seed = "http://127.0.0.1:2379" // thus we use the local server!
-	}
-	// then, connect to `seed` as a client
-
-	// FIXME: validate seed, or wait for it to fail in etcd init?
-
-	// etcd
-	etcdO := &EtcdWObject{
-		seed:      seed,
-		ctimeout:  c.Int("converged-timeout"),
-		converged: converged,
-	}
-
-	hostname := c.String("hostname")
-	if hostname == "" {
-		hostname, _ = os.Hostname() // etcd watch key // XXX: this is not the correct key name this is the set key name... WOOPS
-	}
-	go func() {
-		startchan := make(chan struct{}) // start signal
-		go func() { startchan <- struct{}{} }()
-		file := c.String("file")
-		configchan := make(chan bool)
-		if !c.Bool("no-watch") {
-			configchan = ConfigWatch(file)
-		}
-		log.Println("Etcd: Starting...")
-		etcdchan := etcdO.EtcdWatch()
-		first := true // first loop or not
-		for {
-			log.Println("Main: Waiting...")
-			select {
-			case _ = <-startchan: // kick the loop once at start
-				// pass
-			case msg := <-etcdchan:
-				switch msg {
-				// some types of messages we ignore...
-				case etcdFoo, etcdBar:
-					continue
-				// while others passthrough and cause a compile!
-				case etcdStart, etcdEvent:
-					// pass
-				default:
-					log.Fatal("Etcd: Unhandled message: ", msg)
-				}
-			case msg := <-configchan:
-				if c.Bool("no-watch") || !msg {
-					continue // not ready to read config
-				}
-				//case compile_event: XXX
-			}
-
-			config := ParseConfigFromFile(file)
-			if config == nil {
-				log.Printf("Config parse failure")
-				continue
-			}
-
-			// run graph vertex LOCK...
-			if !first { // TODO: we can flatten this check out I think
-				G.Pause() // sync
-			}
-
-			// build graph from yaml file on events (eg: from etcd)
-			// we need the vertices to be paused to work on them
-			if newFullgraph, err := fullGraph.NewGraphFromConfig(config, etcdO, hostname); err == nil { // keep references to all original elements
-				fullGraph = newFullgraph
-			} else {
-				log.Printf("Config: Error making new graph from config: %v", err)
-				// unpause!
-				if !first {
-					G.Start(&wg, first) // sync
-				}
-				continue
-			}
-
-			G = fullGraph.Copy() // copy to active graph
-			// XXX: do etcd transaction out here...
-			G.AutoEdges() // add autoedges; modifies the graph
-			G.AutoGroup() // run autogroup; modifies the graph
-			// TODO: do we want to do a transitive reduction?
-
-			log.Printf("Graph: %v", G) // show graph
-			err := G.ExecGraphviz(c.String("graphviz-filter"), c.String("graphviz"))
-			if err != nil {
-				log.Printf("Graphviz: %v", err)
-			} else {
-				log.Printf("Graphviz: Successfully generated graph!")
-			}
-			G.SetVertex()
-			G.SetConvergedCallback(c.Int("converged-timeout"), converged)
-			// G.Start(...) needs to be synchronous or wait,
-			// because if half of the nodes are started and
-			// some are not ready yet and the EtcdWatch
-			// loops, we'll cause G.Pause(...) before we
-			// even got going, thus causing nil pointer errors
-			G.Start(&wg, first) // sync
-			first = false
-		}
-	}()
-
-	if i := c.Int("converged-timeout"); i >= 0 {
-		go func() {
-		ConvergedLoop:
-			for {
-				<-converged // when anyone says they have converged
-
-				if etcdO.GetConvergedState() != etcdConvergedTimeout {
-					continue
-				}
-				for v := range G.GetVerticesChan() {
-					if v.Res.GetConvergedState() != resConvergedTimeout {
-						continue ConvergedLoop
-					}
-				}
-
-				// if all have converged, exit
-				log.Printf("Converged for %d seconds, exiting!", i)
-				exit <- true
-				for {
-					<-converged
-				} // unblock/drain
-				//return
-			}
-		}()
-	}
-
-	log.Println("Main: Running...")
-
-	waitForSignal(exit) // pass in exit channel to watch
-
-	G.Exit() // tell all the children to exit
-
-	if DEBUG {
-		log.Printf("Graph: %v", G)
-	}
-
-	wg.Wait() // wait for primary go routines to exit
-
-	// TODO: wait for each vertex to exit...
-	log.Println("Goodbye!")
-}
-
 func main() {
-	//if DEBUG {
+	if err := mgmtmain.CLI(program, version); err != nil {
-	log.SetFlags(log.LstdFlags | log.Lshortfile)
+		fmt.Println(err)
-	//}
+		os.Exit(1)
-	log.SetFlags(log.Flags() - log.Ldate) // remove the date for now
+		return
-	if program == "" || version == "" {
-		log.Fatal("Program was not compiled correctly. Please see Makefile.")
 	}
-	app := cli.NewApp()
-	app.Name = program
-	app.Usage = "next generation config management"
-	app.Version = version
-	//app.Action = ... // without a default action, help runs
-
-	app.Commands = []cli.Command{
-		{
-			Name:    "run",
-			Aliases: []string{"r"},
-			Usage:   "run",
-			Action:  run,
-			Flags: []cli.Flag{
-				cli.StringFlag{
-					Name:  "file, f",
-					Value: "",
-					Usage: "graph definition to run",
-				},
-				cli.BoolFlag{
-					Name:  "no-watch",
-					Usage: "do not update graph on watched graph definition file changes",
-				},
-				cli.StringFlag{
-					Name:  "code, c",
-					Value: "",
-					Usage: "code definition to run",
-				},
-				cli.StringFlag{
-					Name:  "graphviz, g",
-					Value: "",
-					Usage: "output file for graphviz data",
-				},
-				cli.StringFlag{
-					Name:  "graphviz-filter, gf",
-					Value: "dot", // directed graph default
-					Usage: "graphviz filter to use",
-				},
-				// useful for testing multiple instances on same machine
-				cli.StringFlag{
-					Name:  "hostname",
-					Value: "",
-					Usage: "hostname to use",
-				},
-				// if empty, it will startup a new server
-				cli.StringFlag{
-					Name:  "seed, s",
-					Value: "",
-					Usage: "default etc peer endpoint",
-				},
-				cli.IntFlag{
-					Name:  "converged-timeout, t",
-					Value: -1,
-					Usage: "exit after approximately this many seconds in a converged state",
-				},
-				cli.IntFlag{
-					Name:  "max-runtime",
-					Value: 0,
-					Usage: "exit after a maximum of approximately this many seconds",
-				},
-			},
-		},
-	}
-	app.EnableBashCompletion = true
-	app.Run(os.Args)
 }

mgmtmain/cli.go

@@ -0,0 +1,296 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package mgmtmain
+
+import (
+	"fmt"
+	"log"
+	"os"
+	"os/signal"
+	"syscall"
+
+	"github.com/purpleidea/mgmt/puppet"
+	"github.com/purpleidea/mgmt/yamlgraph"
+
+	"github.com/urfave/cli"
+)
+
+// run is the main run target.
+func run(c *cli.Context) error {
+
+	obj := &Main{}
+
+	obj.Program = c.App.Name
+	obj.Version = c.App.Version
+
+	if h := c.String("hostname"); c.IsSet("hostname") && h != "" {
+		obj.Hostname = &h
+	}
+
+	if s := c.String("prefix"); c.IsSet("prefix") && s != "" {
+		obj.Prefix = &s
+	}
+	obj.TmpPrefix = c.Bool("tmp-prefix")
+	obj.AllowTmpPrefix = c.Bool("allow-tmp-prefix")
+
+	if _ = c.String("code"); c.IsSet("code") {
+		if obj.GAPI != nil {
+			return fmt.Errorf("Can't combine code GAPI with existing GAPI.")
+		}
+		// TODO: implement DSL GAPI
+		//obj.GAPI = &dsl.GAPI{
+		//	Code: &s,
+		//}
+		return fmt.Errorf("The Code GAPI is not implemented yet!") // TODO: DSL
+	}
+	if y := c.String("yaml"); c.IsSet("yaml") {
+		if obj.GAPI != nil {
+			return fmt.Errorf("Can't combine YAML GAPI with existing GAPI.")
+		}
+		obj.GAPI = &yamlgraph.GAPI{
+			File: &y,
+		}
+	}
+	if p := c.String("puppet"); c.IsSet("puppet") {
+		if obj.GAPI != nil {
+			return fmt.Errorf("Can't combine puppet GAPI with existing GAPI.")
+		}
+		obj.GAPI = &puppet.GAPI{
+			PuppetParam: &p,
+			PuppetConf:  c.String("puppet-conf"),
+		}
+	}
+	obj.Remotes = c.StringSlice("remote") // FIXME: GAPI-ify somehow?
+
+	obj.NoWatch = c.Bool("no-watch")
+	obj.Noop = c.Bool("noop")
+	obj.Graphviz = c.String("graphviz")
+	obj.GraphvizFilter = c.String("graphviz-filter")
+	obj.ConvergedTimeout = c.Int("converged-timeout")
+	obj.MaxRuntime = uint(c.Int("max-runtime"))
+
+	obj.Seeds = c.StringSlice("seeds")
+	obj.ClientURLs = c.StringSlice("client-urls")
+	obj.ServerURLs = c.StringSlice("server-urls")
+	obj.IdealClusterSize = c.Int("ideal-cluster-size")
+	obj.NoServer = c.Bool("no-server")
+
+	obj.CConns = uint16(c.Int("cconns"))
+	obj.AllowInteractive = c.Bool("allow-interactive")
+	obj.SSHPrivIDRsa = c.String("ssh-priv-id-rsa")
+	obj.NoCaching = c.Bool("no-caching")
+	obj.Depth = uint16(c.Int("depth"))
+
+	if err := obj.Init(); err != nil {
+		return err
+	}
+
+	// install the exit signal handler
+	exit := make(chan struct{})
+	defer close(exit)
+	go func() {
+		signals := make(chan os.Signal, 1)
+		signal.Notify(signals, os.Interrupt) // catch ^C
+		//signal.Notify(signals, os.Kill) // catch signals
+		signal.Notify(signals, syscall.SIGTERM)
+
+		select {
+		case sig := <-signals: // any signal will do
+			if sig == os.Interrupt {
+				log.Println("Interrupted by ^C")
+				obj.Exit(nil)
+				return
+			}
+			log.Println("Interrupted by signal")
+			obj.Exit(fmt.Errorf("Killed by %v", sig))
+			return
+		case <-exit:
+			return
+		}
+	}()
+
+	if err := obj.Run(); err != nil {
+		return err
+		//return cli.NewExitError(err.Error(), 1) // TODO: ?
+		//return cli.NewExitError("", 1) // TODO: ?
+	}
+	return nil
+}
+
+// CLI is the entry point for using mgmt normally from the CLI.
+func CLI(program, version string) error {
+
+	// test for sanity
+	if program == "" || version == "" {
+		return fmt.Errorf("Program was not compiled correctly. Please see Makefile.")
+	}
+	app := cli.NewApp()
+	app.Name = program // App.name and App.version pass these values through
+	app.Version = version
+	app.Usage = "next generation config management"
+	//app.Action = ... // without a default action, help runs
+
+	app.Commands = []cli.Command{
+		{
+			Name:    "run",
+			Aliases: []string{"r"},
+			Usage:   "run",
+			Action:  run,
+			Flags: []cli.Flag{
+				// useful for testing multiple instances on same machine
+				cli.StringFlag{
+					Name:  "hostname",
+					Value: "",
+					Usage: "hostname to use",
+				},
+
+				cli.StringFlag{
+					Name:   "prefix",
+					Usage:  "specify a path to the working prefix directory",
+					EnvVar: "MGMT_PREFIX",
+				},
+				cli.BoolFlag{
+					Name:  "tmp-prefix",
+					Usage: "request a pseudo-random, temporary prefix to be used",
+				},
+				cli.BoolFlag{
+					Name:  "allow-tmp-prefix",
+					Usage: "allow creation of a new temporary prefix if main prefix is unavailable",
+				},
+
+				cli.StringFlag{
+					Name:  "code, c",
+					Value: "",
+					Usage: "code definition to run",
+				},
+				cli.StringFlag{
+					Name:  "yaml",
+					Value: "",
+					Usage: "yaml graph definition to run",
+				},
+				cli.StringFlag{
+					Name:  "puppet, p",
+					Value: "",
+					Usage: "load graph from puppet, optionally takes a manifest or path to manifest file",
+				},
+				cli.StringFlag{
+					Name:  "puppet-conf",
+					Value: "",
+					Usage: "the path to an alternate puppet.conf file",
+				},
+				cli.StringSliceFlag{
+					Name:  "remote",
+					Value: &cli.StringSlice{},
+					Usage: "list of remote graph definitions to run",
+				},
+
+				cli.BoolFlag{
+					Name:  "no-watch",
+					Usage: "do not update graph on stream switch events",
+				},
+				cli.BoolFlag{
+					Name:  "noop",
+					Usage: "globally force all resources into no-op mode",
+				},
+				cli.StringFlag{
+					Name:  "graphviz, g",
+					Value: "",
+					Usage: "output file for graphviz data",
+				},
+				cli.StringFlag{
+					Name:  "graphviz-filter, gf",
+					Value: "dot", // directed graph default
+					Usage: "graphviz filter to use",
+				},
+				cli.IntFlag{
+					Name:   "converged-timeout, t",
+					Value:  -1,
+					Usage:  "exit after approximately this many seconds in a converged state",
+					EnvVar: "MGMT_CONVERGED_TIMEOUT",
+				},
+				cli.IntFlag{
+					Name:   "max-runtime",
+					Value:  0,
+					Usage:  "exit after a maximum of approximately this many seconds",
+					EnvVar: "MGMT_MAX_RUNTIME",
+				},
+
+				// if empty, it will startup a new server
+				cli.StringSliceFlag{
+					Name:   "seeds, s",
+					Value:  &cli.StringSlice{}, // empty slice
+					Usage:  "default etc client endpoint",
+					EnvVar: "MGMT_SEEDS",
+				},
+				// port 2379 and 4001 are common
+				cli.StringSliceFlag{
+					Name:   "client-urls",
+					Value:  &cli.StringSlice{},
+					Usage:  "list of URLs to listen on for client traffic",
+					EnvVar: "MGMT_CLIENT_URLS",
+				},
+				// port 2380 and 7001 are common
+				cli.StringSliceFlag{
+					Name:   "server-urls, peer-urls",
+					Value:  &cli.StringSlice{},
+					Usage:  "list of URLs to listen on for server (peer) traffic",
+					EnvVar: "MGMT_SERVER_URLS",
+				},
+				cli.IntFlag{
+					Name:   "ideal-cluster-size",
+					Value:  -1,
+					Usage:  "ideal number of server peers in cluster; only read by initial server",
+					EnvVar: "MGMT_IDEAL_CLUSTER_SIZE",
+				},
+				cli.BoolFlag{
+					Name:  "no-server",
+					Usage: "do not let other servers peer with me",
+				},
+
+				cli.IntFlag{
+					Name:   "cconns",
+					Value:  0,
+					Usage:  "number of maximum concurrent remote ssh connections to run; 0 for unlimited",
+					EnvVar: "MGMT_CCONNS",
+				},
+				cli.BoolFlag{
+					Name:  "allow-interactive",
+					Usage: "allow interactive prompting, such as for remote passwords",
+				},
+				cli.StringFlag{
+					Name:   "ssh-priv-id-rsa",
+					Value:  "~/.ssh/id_rsa",
+					Usage:  "default path to ssh key file, set empty to never touch",
+					EnvVar: "MGMT_SSH_PRIV_ID_RSA",
+				},
+				cli.BoolFlag{
+					Name:  "no-caching",
+					Usage: "don't allow remote caching of remote execution binary",
+				},
+				cli.IntFlag{
+					Name:   "depth",
+					Hidden: true, // internal use only
+					Value:  0,
+					Usage:  "specify depth in remote hierarchy",
+				},
+			},
+		},
+	}
+	app.EnableBashCompletion = true
+	return app.Run(os.Args)
+}

mgmtmain/main.go

@@ -0,0 +1,480 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package mgmtmain
+
+import (
+	"fmt"
+	"io/ioutil"
+	"log"
+	"os"
+	"sync"
+	"time"
+
+	"github.com/purpleidea/mgmt/converger"
+	"github.com/purpleidea/mgmt/etcd"
+	"github.com/purpleidea/mgmt/gapi"
+	"github.com/purpleidea/mgmt/pgraph"
+	"github.com/purpleidea/mgmt/recwatch"
+	"github.com/purpleidea/mgmt/remote"
+	"github.com/purpleidea/mgmt/util"
+
+	etcdtypes "github.com/coreos/etcd/pkg/types"
+	"github.com/coreos/pkg/capnslog"
+	multierr "github.com/hashicorp/go-multierror"
+	errwrap "github.com/pkg/errors"
+)
+
+// Main is the main struct for running the mgmt logic.
+type Main struct {
+	Program string // the name of this program, usually set at compile time
+	Version string // the version of this program, usually set at compile time
+
+	Hostname *string // hostname to use; nil if undefined
+
+	Prefix         *string // prefix passed in; nil if undefined
+	TmpPrefix      bool    // request a pseudo-random, temporary prefix to be used
+	AllowTmpPrefix bool    // allow creation of a new temporary prefix if main prefix is unavailable
+
+	GAPI    gapi.GAPI // graph API interface struct
+	Remotes []string  // list of remote graph definitions to run
+
+	NoWatch          bool   // do not update graph on watched graph definition file changes
+	Noop             bool   // globally force all resources into no-op mode
+	Graphviz         string // output file for graphviz data
+	GraphvizFilter   string // graphviz filter to use
+	ConvergedTimeout int    // exit after approximately this many seconds in a converged state; -1 to disable
+	MaxRuntime       uint   // exit after a maximum of approximately this many seconds
+
+	Seeds            []string // default etc client endpoint
+	ClientURLs       []string // list of URLs to listen on for client traffic
+	ServerURLs       []string // list of URLs to listen on for server (peer) traffic
+	IdealClusterSize int      // ideal number of server peers in cluster; only read by initial server
+	NoServer         bool     // do not let other servers peer with me
+
+	CConns           uint16 // number of maximum concurrent remote ssh connections to run, 0 for unlimited
+	AllowInteractive bool   // allow interactive prompting, such as for remote passwords
+	SSHPrivIDRsa     string // default path to ssh key file, set empty to never touch
+	NoCaching        bool   // don't allow remote caching of remote execution binary
+	Depth            uint16 // depth in remote hierarchy; for internal use only
+
+	DEBUG   bool
+	VERBOSE bool
+
+	seeds            etcdtypes.URLs // processed seeds value
+	clientURLs       etcdtypes.URLs // processed client urls value
+	serverURLs       etcdtypes.URLs // processed server urls value
+	idealClusterSize uint16         // processed ideal cluster size value
+
+	exit chan error // exit signal
+}
+
+// Init initializes the main struct after it performs some validation.
+func (obj *Main) Init() error {
+
+	if obj.Program == "" || obj.Version == "" {
+		return fmt.Errorf("You must set the Program and Version strings!")
+	}
+
+	if obj.Prefix != nil && obj.TmpPrefix {
+		return fmt.Errorf("Choosing a prefix and the request for a tmp prefix is illogical!")
+	}
+
+	obj.idealClusterSize = uint16(obj.IdealClusterSize)
+	if obj.IdealClusterSize < 0 { // value is undefined, set to the default
+		obj.idealClusterSize = etcd.DefaultIdealClusterSize
+	}
+
+	if obj.idealClusterSize < 1 {
+		return fmt.Errorf("IdealClusterSize should be at least one!")
+	}
+
+	if obj.NoServer && len(obj.Remotes) > 0 {
+		// TODO: in this case, we won't be able to tunnel stuff back to
+		// here, so if we're okay with every remote graph running in an
+		// isolated mode, then this is okay. Improve on this if there's
+		// someone who really wants to be able to do this.
+		return fmt.Errorf("The Server is required when using Remotes!")
+	}
+
+	if obj.CConns < 0 {
+		return fmt.Errorf("The CConns value should be at least zero!")
+	}
+
+	if obj.ConvergedTimeout >= 0 && obj.CConns > 0 && len(obj.Remotes) > int(obj.CConns) {
+		return fmt.Errorf("You can't converge if you have more remotes than available connections!")
+	}
+
+	if obj.Depth < 0 { // user should not be using this argument manually
+		return fmt.Errorf("Negative values for Depth are not permitted!")
+	}
+
+	// transform the url list inputs into etcd typed lists
+	var err error
+	obj.seeds, err = etcdtypes.NewURLs(
+		util.FlattenListWithSplit(obj.Seeds, []string{",", ";", " "}),
+	)
+	if err != nil && len(obj.Seeds) > 0 {
+		return fmt.Errorf("Seeds didn't parse correctly!")
+	}
+	obj.clientURLs, err = etcdtypes.NewURLs(
+		util.FlattenListWithSplit(obj.ClientURLs, []string{",", ";", " "}),
+	)
+	if err != nil && len(obj.ClientURLs) > 0 {
+		return fmt.Errorf("ClientURLs didn't parse correctly!")
+	}
+	obj.serverURLs, err = etcdtypes.NewURLs(
+		util.FlattenListWithSplit(obj.ServerURLs, []string{",", ";", " "}),
+	)
+	if err != nil && len(obj.ServerURLs) > 0 {
+		return fmt.Errorf("ServerURLs didn't parse correctly!")
+	}
+
+	obj.exit = make(chan error)
+	return nil
+}
+
+// Exit causes a safe shutdown. This is often attached to the ^C signal handler.
+func (obj *Main) Exit(err error) {
+	obj.exit <- err // trigger an exit!
+}
+
+// Run is the main execution entrypoint to run mgmt.
+func (obj *Main) Run() error {
+
+	var start = time.Now().UnixNano()
+
+	var flags int
+	if obj.DEBUG || true { // TODO: remove || true
+		flags = log.LstdFlags | log.Lshortfile
+	}
+	flags = (flags - log.Ldate) // remove the date for now
+	log.SetFlags(flags)
+
+	// un-hijack from capnslog...
+	log.SetOutput(os.Stderr)
+	if obj.VERBOSE {
+		capnslog.SetFormatter(capnslog.NewLogFormatter(os.Stderr, "(etcd) ", flags))
+	} else {
+		capnslog.SetFormatter(capnslog.NewNilFormatter())
+	}
+
+	log.Printf("This is: %s, version: %s", obj.Program, obj.Version)
+	log.Printf("Main: Start: %v", start)
+
+	hostname, err := os.Hostname() // a sensible default
+	// allow passing in the hostname, instead of using the system setting
+	if h := obj.Hostname; h != nil && *h != "" { // override by cli
+		hostname = *h
+	} else if err != nil {
+		return errwrap.Wrapf(err, "Can't get default hostname!")
+	}
+	if hostname == "" { // safety check
+		return fmt.Errorf("Hostname cannot be empty!")
+	}
+
+	var prefix = fmt.Sprintf("/var/lib/%s/", obj.Program) // default prefix
+	if p := obj.Prefix; p != nil {
+		prefix = *p
+	}
+	// make sure the working directory prefix exists
+	if obj.TmpPrefix || os.MkdirAll(prefix, 0770) != nil {
+		if obj.TmpPrefix || obj.AllowTmpPrefix {
+			var err error
+			if prefix, err = ioutil.TempDir("", obj.Program+"-"+hostname+"-"); err != nil {
+				return fmt.Errorf("Main: Error: Can't create temporary prefix!")
+			}
+			log.Println("Main: Warning: Working prefix directory is temporary!")
+
+		} else {
+			return fmt.Errorf("Main: Error: Can't create prefix!")
+		}
+	}
+	log.Printf("Main: Working prefix is: %s", prefix)
+
+	var wg sync.WaitGroup
+	var G, oldGraph *pgraph.Graph
+
+	// exit after `max-runtime` seconds for no reason at all...
+	if i := obj.MaxRuntime; i > 0 {
+		go func() {
+			time.Sleep(time.Duration(i) * time.Second)
+			obj.Exit(nil)
+		}()
+	}
+
+	// setup converger
+	converger := converger.NewConverger(
+		obj.ConvergedTimeout,
+		nil, // stateFn gets added in by EmbdEtcd
+	)
+	go converger.Loop(true) // main loop for converger, true to start paused
+
+	// embedded etcd
+	if len(obj.seeds) == 0 {
+		log.Printf("Main: Seeds: No seeds specified!")
+	} else {
+		log.Printf("Main: Seeds(%d): %v", len(obj.seeds), obj.seeds)
+	}
+	EmbdEtcd := etcd.NewEmbdEtcd(
+		hostname,
+		obj.seeds,
+		obj.clientURLs,
+		obj.serverURLs,
+		obj.NoServer,
+		obj.idealClusterSize,
+		prefix,
+		converger,
+	)
+	if EmbdEtcd == nil {
+		// TODO: verify EmbdEtcd is not nil below...
+		obj.Exit(fmt.Errorf("Main: Etcd: Creation failed!"))
+	} else if err := EmbdEtcd.Startup(); err != nil { // startup (returns when etcd main loop is running)
+		obj.Exit(fmt.Errorf("Main: Etcd: Startup failed: %v", err))
+	}
+	convergerStateFn := func(b bool) error {
+		// exit if we are using the converged timeout and we are the
+		// root node. otherwise, if we are a child node in a remote
+		// execution hierarchy, we should only notify our converged
+		// state and wait for the parent to trigger the exit.
+		if t := obj.ConvergedTimeout; obj.Depth == 0 && t >= 0 {
+			if b {
+				log.Printf("Converged for %d seconds, exiting!", t)
+				obj.Exit(nil) // trigger an exit!
+			}
+			return nil
+		}
+		// send our individual state into etcd for others to see
+		return etcd.EtcdSetHostnameConverged(EmbdEtcd, hostname, b) // TODO: what should happen on error?
+	}
+	if EmbdEtcd != nil {
+		converger.SetStateFn(convergerStateFn)
+	}
+
+	var gapiChan chan error // stream events are nil errors
+	if obj.GAPI != nil {
+		data := gapi.Data{
+			Hostname: hostname,
+			EmbdEtcd: EmbdEtcd,
+			Noop:     obj.Noop,
+			NoWatch:  obj.NoWatch,
+		}
+		if err := obj.GAPI.Init(data); err != nil {
+			obj.Exit(fmt.Errorf("Main: GAPI: Init failed: %v", err))
+		} else if !obj.NoWatch {
+			gapiChan = obj.GAPI.SwitchStream() // stream of graph switch events!
+		}
+	}
+
+	exitchan := make(chan struct{}) // exit on close
+	go func() {
+		startchan := make(chan struct{}) // start signal
+		go func() { startchan <- struct{}{} }()
+
+		log.Println("Etcd: Starting...")
+		etcdchan := etcd.EtcdWatch(EmbdEtcd)
+		first := true // first loop or not
+		for {
+			log.Println("Main: Waiting...")
+			select {
+			case <-startchan: // kick the loop once at start
+				// pass
+
+			case b := <-etcdchan:
+				if !b { // ignore the message
+					continue
+				}
+				// everything else passes through to cause a compile!
+
+			case err, ok := <-gapiChan:
+				if !ok { // channel closed
+					continue
+				}
+				if err != nil {
+					obj.Exit(err) // trigger exit
+					continue
+					//return // TODO: return or wait for exitchan?
+				}
+				if obj.NoWatch { // extra safety for bad GAPI's
+					log.Printf("Main: GAPI stream should be quiet with NoWatch!") // fix the GAPI!
+					continue                                                      // no stream events should be sent
+				}
+
+			case <-exitchan:
+				return
+			}
+
+			if obj.GAPI == nil {
+				log.Printf("Config: GAPI is empty!")
+				continue
+			}
+
+			// we need the vertices to be paused to work on them, so
+			// run graph vertex LOCK...
+			if !first { // TODO: we can flatten this check out I think
+				converger.Pause() // FIXME: add sync wait?
+				G.Pause()         // sync
+
+				//G.UnGroup() // FIXME: implement me if needed!
+			}
+
+			// make the graph from yaml, lib, puppet->yaml, or dsl!
+			newGraph, err := obj.GAPI.Graph() // generate graph!
+			if err != nil {
+				log.Printf("Config: Error creating new graph: %v", err)
+				// unpause!
+				if !first {
+					G.Start(&wg, first) // sync
+					converger.Start()   // after G.Start()
+				}
+				continue
+			}
+
+			// apply the global noop parameter if requested
+			if obj.Noop {
+				for _, m := range newGraph.GraphMetas() {
+					m.Noop = obj.Noop
+				}
+			}
+
+			// FIXME: make sure we "UnGroup()" any semi-destructive
+			// changes to the resources so our efficient GraphSync
+			// will be able to re-use and cmp to the old graph.
+			newFullGraph, err := newGraph.GraphSync(oldGraph)
+			if err != nil {
+				log.Printf("Config: Error running graph sync: %v", err)
+				// unpause!
+				if !first {
+					G.Start(&wg, first) // sync
+					converger.Start()   // after G.Start()
+				}
+				continue
+			}
+			oldGraph = newFullGraph // save old graph
+			G = oldGraph.Copy()     // copy to active graph
+
+			G.AutoEdges() // add autoedges; modifies the graph
+			G.AutoGroup() // run autogroup; modifies the graph
+			// TODO: do we want to do a transitive reduction?
+
+			log.Printf("Graph: %v", G) // show graph
+			if obj.GraphvizFilter != "" {
+				if err := G.ExecGraphviz(obj.GraphvizFilter, obj.Graphviz); err != nil {
+					log.Printf("Graphviz: %v", err)
+				} else {
+					log.Printf("Graphviz: Successfully generated graph!")
+				}
+			}
+			G.AssociateData(converger)
+			// G.Start(...) needs to be synchronous or wait,
+			// because if half of the nodes are started and
+			// some are not ready yet and the EtcdWatch
+			// loops, we'll cause G.Pause(...) before we
+			// even got going, thus causing nil pointer errors
+			G.Start(&wg, first) // sync
+			converger.Start()   // after G.Start()
+			first = false
+		}
+	}()
+
+	configWatcher := recwatch.NewConfigWatcher()
+	events := configWatcher.Events()
+	if !obj.NoWatch {
+		configWatcher.Add(obj.Remotes...) // add all the files...
+	} else {
+		events = nil // signal that no-watch is true
+	}
+	go func() {
+		select {
+		case err := <-configWatcher.Error():
+			obj.Exit(err) // trigger an exit!
+
+		case <-exitchan:
+			return
+		}
+	}()
+
+	// initialize the add watcher, which calls the f callback on map changes
+	convergerCb := func(f func(map[string]bool) error) (func(), error) {
+		return etcd.EtcdAddHostnameConvergedWatcher(EmbdEtcd, f)
+	}
+
+	// build remotes struct for remote ssh
+	remotes := remote.NewRemotes(
+		EmbdEtcd.LocalhostClientURLs().StringSlice(),
+		[]string{etcd.DefaultClientURL},
+		obj.Noop,
+		obj.Remotes, // list of files
+		events,      // watch for file changes
+		obj.CConns,
+		obj.AllowInteractive,
+		obj.SSHPrivIDRsa,
+		!obj.NoCaching,
+		obj.Depth,
+		prefix,
+		converger,
+		convergerCb,
+		obj.Program,
+	)
+
+	// TODO: is there any benefit to running the remotes above in the loop?
+	// wait for etcd to be running before we remote in, which we do above!
+	go remotes.Run()
+
+	if obj.GAPI == nil {
+		converger.Start() // better start this for empty graphs
+	}
+	log.Println("Main: Running...")
+
+	reterr := <-obj.exit // wait for exit signal
+
+	log.Println("Destroy...")
+
+	if obj.GAPI != nil {
+		if err := obj.GAPI.Close(); err != nil {
+			err = errwrap.Wrapf(err, "GAPI closed poorly!")
+			reterr = multierr.Append(reterr, err) // list of errors
+		}
+	}
+
+	configWatcher.Close()                  // stop sending file changes to remotes
+	if err := remotes.Exit(); err != nil { // tell all the remote connections to shutdown; waits!
+		err = errwrap.Wrapf(err, "Remote exited poorly!")
+		reterr = multierr.Append(reterr, err) // list of errors
+	}
+
+	G.Exit() // tell all the children to exit
+
+	// tell inner main loop to exit
+	close(exitchan)
+
+	// cleanup etcd main loop last so it can process everything first
+	if err := EmbdEtcd.Destroy(); err != nil { // shutdown and cleanup etcd
+		err = errwrap.Wrapf(err, "Etcd exited poorly!")
+		reterr = multierr.Append(reterr, err) // list of errors
+	}
+
+	if obj.DEBUG {
+		log.Printf("Graph: %v", G)
+	}
+
+	wg.Wait() // wait for primary go routines to exit
+
+	// TODO: wait for each vertex to exit...
+	log.Println("Goodbye!")
+	return reterr
+}

misc/go

@@ -0,0 +1,15 @@
+#!/bin/bash
+# hack around stupid $GOPATH semantics, with ~/bin/go helper
+# thanks to Nilium in #go-nuts for 1/3 of the idea
+[ -z "$GOPATH" ] && echo '$GOPATH is not set!' && exit 1
+GO="$(which -a go | sed -e '2q;d')"	# TODO: pick /usr/bin/go in a better way
+if [ "$1" = "generate" ]; then
+	exec $GO "$@"	# go generate is stupid and gets confused by $GOPATH
+fi
+# the idea is to have $project/gopath/src/ be a symlink to ../vendor but you put
+# all of your vendored things in vendor/ but with this gopath can be per project
+if [ -d "$PWD/vendor/" ] && [ -d "$PWD/gopath/" ] && [ "`readlink $PWD/gopath/src`" = "../vendor" ] ; then
+	GOPATH="$PWD/gopath/:$GOPATH" $GO "$@"
+else
+	$GO "$@"
+fi

misc/make-deps.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # setup a simple go environment
 XPWD=`pwd`
 ROOT="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && cd .. && pwd )"	# dir!
@@ -9,43 +9,53 @@ if env | grep -q '^TRAVIS=true$'; then
 	travis=1
 fi
 
+sudo_command=$(which sudo)
+
+YUM=`which yum 2>/dev/null`
+APT=`which apt-get 2>/dev/null`
+if [ -z "$YUM" -a -z "$APT" ]; then
+	echo "The package managers can't be found."
+	exit 1
+fi
+
+if [ ! -z "$YUM" ]; then
+	$sudo_command $YUM install -y libvirt-devel
+
+fi
+if [ ! -z "$APT" ]; then
+	$sudo_command $APT install -y libvirt-dev || true
+	$sudo_command $APT install -y libpcap0.8-dev || true
+fi
+
 if [ $travis -eq 0 ]; then
-	YUM=`which yum 2>/dev/null`
-	APT=`which apt-get 2>/dev/null`
-	if [ -z "$YUM" -a -z "$APT" ]; then
-		echo "The package managers can't be found."
-		exit 1
-	fi
 	if [ ! -z "$YUM" ]; then
 		# some go dependencies are stored in mercurial
-		sudo $YUM install -y golang golang-googlecode-tools-stringer hg
+		$sudo_command $YUM install -y golang golang-googlecode-tools-stringer hg
 
 	fi
 	if [ ! -z "$APT" ]; then
-		sudo $APT update
+		$sudo_command $APT update
-		sudo $APT install -y golang make gcc packagekit mercurial
+		$sudo_command $APT install -y golang make gcc packagekit mercurial
 		# one of these two golang tools packages should work on debian
-		sudo $APT install -y golang-golang-x-tools || true
+		$sudo_command $APT install -y golang-golang-x-tools || true
-		sudo $APT install -y golang-go.tools || true
+		$sudo_command $APT install -y golang-go.tools || true
 	fi
 fi
 
-# build etcd
+# if golang is too old, we don't want to fail with an obscure error later
-git clone --recursive https://github.com/coreos/etcd/ && cd etcd
+if go version | grep 'go1\.[0123]\.'; then
-goversion=$(go version)
+	echo "mgmt requires go1.4 or higher."
-# if 'go version' contains string 'devel', then use git master of etcd...
+	exit 1
-if [ "${goversion#*devel}" == "$goversion" ]; then
-	git checkout v2.2.4	# TODO: update to newer versions as needed
 fi
-[ -x build ] && ./build
-mkdir -p ~/bin/
-cp bin/etcd ~/bin/
-cd - >/dev/null
-rm -rf etcd	# clean up to avoid failing on upstream gofmt errors
 
-go get ./...	# get all the go dependencies
+go get -d ./...	# get all the go dependencies
 [ -e "$GOBIN/mgmt" ] && rm -f "$GOBIN/mgmt"	# the `go get` version has no -X
-go get golang.org/x/tools/cmd/vet	# add in `go vet` for travis
+# vet is built-in in go 1.6 - we check for go vet command
+go vet 1> /dev/null 2>&1
+ret=$?
+if [[ $ret != 0 ]]; then
+	go get golang.org/x/tools/cmd/vet      # add in `go vet` for travis
+fi
 go get golang.org/x/tools/cmd/stringer	# for automatic stringer-ing
 go get github.com/golang/lint/golint	# for `golint`-ing
 cd "$XPWD" >/dev/null

misc/mgmt.service

@@ -0,0 +1,13 @@
+[Unit]
+Description=Run mgmt configuration management
+Documentation=https://github.com/purpleidea/mgmt/
+After=systemd-networkd.service
+Requires=systemd-networkd.service
+
+[Service]
+ExecStart=/usr/bin/mgmt run ${OPTS}
+RestartSec=5s
+Restart=always
+
+[Install]
+WantedBy=multi-user.target

pgraph/autoedge.go

@@ -0,0 +1,104 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package pgraph represents the internal "pointer graph" that we use.
+package pgraph
+
+import (
+	"fmt"
+	"log"
+
+	"github.com/purpleidea/mgmt/global"
+	"github.com/purpleidea/mgmt/resources"
+)
+
+// add edges to the vertex in a graph based on if it matches a uid list
+func (g *Graph) addEdgesByMatchingUIDS(v *Vertex, uids []resources.ResUID) []bool {
+	// search for edges and see what matches!
+	var result []bool
+
+	// loop through each uid, and see if it matches any vertex
+	for _, uid := range uids {
+		var found = false
+		// uid is a ResUID object
+		for _, vv := range g.GetVertices() { // search
+			if v == vv { // skip self
+				continue
+			}
+			if global.DEBUG {
+				log.Printf("Compile: AutoEdge: Match: %v[%v] with UID: %v[%v]", vv.Kind(), vv.GetName(), uid.Kind(), uid.GetName())
+			}
+			// we must match to an effective UID for the resource,
+			// that is to say, the name value of a res is a helpful
+			// handle, but it is not necessarily a unique identity!
+			// remember, resources can return multiple UID's each!
+			if resources.UIDExistsInUIDs(uid, vv.GetUIDs()) {
+				// add edge from: vv -> v
+				if uid.Reversed() {
+					txt := fmt.Sprintf("AutoEdge: %v[%v] -> %v[%v]", vv.Kind(), vv.GetName(), v.Kind(), v.GetName())
+					log.Printf("Compile: Adding %v", txt)
+					g.AddEdge(vv, v, NewEdge(txt))
+				} else { // edges go the "normal" way, eg: pkg resource
+					txt := fmt.Sprintf("AutoEdge: %v[%v] -> %v[%v]", v.Kind(), v.GetName(), vv.Kind(), vv.GetName())
+					log.Printf("Compile: Adding %v", txt)
+					g.AddEdge(v, vv, NewEdge(txt))
+				}
+				found = true
+				break
+			}
+		}
+		result = append(result, found)
+	}
+	return result
+}
+
+// AutoEdges adds the automatic edges to the graph.
+func (g *Graph) AutoEdges() {
+	log.Println("Compile: Adding AutoEdges...")
+	for _, v := range g.GetVertices() { // for each vertexes autoedges
+		if !v.Meta().AutoEdge { // is the metaparam true?
+			continue
+		}
+		autoEdgeObj := v.AutoEdges()
+		if autoEdgeObj == nil {
+			log.Printf("%v[%v]: Config: No auto edges were found!", v.Kind(), v.GetName())
+			continue // next vertex
+		}
+
+		for { // while the autoEdgeObj has more uids to add...
+			uids := autoEdgeObj.Next() // get some!
+			if uids == nil {
+				log.Printf("%v[%v]: Config: The auto edge list is empty!", v.Kind(), v.GetName())
+				break // inner loop
+			}
+			if global.DEBUG {
+				log.Println("Compile: AutoEdge: UIDS:")
+				for i, u := range uids {
+					log.Printf("Compile: AutoEdge: UID%d: %v", i, u)
+				}
+			}
+
+			// match and add edges
+			result := g.addEdgesByMatchingUIDS(v, uids)
+
+			// report back, and find out if we should continue
+			if !autoEdgeObj.Test(result) {
+				break
+			}
+		}
+	}
+}

pgraph/autogroup.go

@@ -0,0 +1,350 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package pgraph
+
+import (
+	"fmt"
+	"log"
+
+	"github.com/purpleidea/mgmt/global"
+
+	errwrap "github.com/pkg/errors"
+)
+
+// AutoGrouper is the required interface to implement for an autogroup algorithm
+type AutoGrouper interface {
+	// listed in the order these are typically called in...
+	name() string                                  // friendly identifier
+	init(*Graph) error                             // only call once
+	vertexNext() (*Vertex, *Vertex, error)         // mostly algorithmic
+	vertexCmp(*Vertex, *Vertex) error              // can we merge these ?
+	vertexMerge(*Vertex, *Vertex) (*Vertex, error) // vertex merge fn to use
+	edgeMerge(*Edge, *Edge) *Edge                  // edge merge fn to use
+	vertexTest(bool) (bool, error)                 // call until false
+}
+
+// baseGrouper is the base type for implementing the AutoGrouper interface
+type baseGrouper struct {
+	graph    *Graph    // store a pointer to the graph
+	vertices []*Vertex // cached list of vertices
+	i        int
+	j        int
+	done     bool
+}
+
+// name provides a friendly name for the logs to see
+func (ag *baseGrouper) name() string {
+	return "baseGrouper"
+}
+
+// init is called only once and before using other AutoGrouper interface methods
+// the name method is the only exception: call it any time without side effects!
+func (ag *baseGrouper) init(g *Graph) error {
+	if ag.graph != nil {
+		return fmt.Errorf("The init method has already been called!")
+	}
+	ag.graph = g                               // pointer
+	ag.vertices = ag.graph.GetVerticesSorted() // cache in deterministic order!
+	ag.i = 0
+	ag.j = 0
+	if len(ag.vertices) == 0 { // empty graph
+		ag.done = true
+		return nil
+	}
+	return nil
+}
+
+// vertexNext is a simple iterator that loops through vertex (pair) combinations
+// an intelligent algorithm would selectively offer only valid pairs of vertices
+// these should satisfy logical grouping requirements for the autogroup designs!
+// the desired algorithms can override, but keep this method as a base iterator!
+func (ag *baseGrouper) vertexNext() (v1, v2 *Vertex, err error) {
+	// this does a for v... { for w... { return v, w }} but stepwise!
+	l := len(ag.vertices)
+	if ag.i < l {
+		v1 = ag.vertices[ag.i]
+	}
+	if ag.j < l {
+		v2 = ag.vertices[ag.j]
+	}
+
+	// in case the vertex was deleted
+	if !ag.graph.HasVertex(v1) {
+		v1 = nil
+	}
+	if !ag.graph.HasVertex(v2) {
+		v2 = nil
+	}
+
+	// two nested loops...
+	if ag.j < l {
+		ag.j++
+	}
+	if ag.j == l {
+		ag.j = 0
+		if ag.i < l {
+			ag.i++
+		}
+		if ag.i == l {
+			ag.done = true
+		}
+	}
+
+	return
+}
+
+func (ag *baseGrouper) vertexCmp(v1, v2 *Vertex) error {
+	if v1 == nil || v2 == nil {
+		return fmt.Errorf("Vertex is nil!")
+	}
+	if v1 == v2 { // skip yourself
+		return fmt.Errorf("Vertices are the same!")
+	}
+	if v1.Kind() != v2.Kind() { // we must group similar kinds
+		// TODO: maybe future resources won't need this limitation?
+		return fmt.Errorf("The two resources aren't the same kind!")
+	}
+	// someone doesn't want to group!
+	if !v1.Meta().AutoGroup || !v2.Meta().AutoGroup {
+		return fmt.Errorf("One of the autogroup flags is false!")
+	}
+	if v1.Res.IsGrouped() { // already grouped!
+		return fmt.Errorf("Already grouped!")
+	}
+	if len(v2.Res.GetGroup()) > 0 { // already has children grouped!
+		return fmt.Errorf("Already has groups!")
+	}
+	if !v1.Res.GroupCmp(v2.Res) { // resource groupcmp failed!
+		return fmt.Errorf("The GroupCmp failed!")
+	}
+	return nil // success
+}
+
+func (ag *baseGrouper) vertexMerge(v1, v2 *Vertex) (v *Vertex, err error) {
+	// NOTE: it's important to use w.Res instead of w, b/c
+	// the w by itself is the *Vertex obj, not the *Res obj
+	// which is contained within it! They both satisfy the
+	// Res interface, which is why both will compile! :(
+	err = v1.Res.GroupRes(v2.Res) // GroupRes skips stupid groupings
+	return                        // success or fail, and no need to merge the actual vertices!
+}
+
+func (ag *baseGrouper) edgeMerge(e1, e2 *Edge) *Edge {
+	return e1 // noop
+}
+
+// vertexTest processes the results of the grouping for the algorithm to know
+// return an error if something went horribly wrong, and bool false to stop
+func (ag *baseGrouper) vertexTest(b bool) (bool, error) {
+	// NOTE: this particular baseGrouper version doesn't track what happens
+	// because since we iterate over every pair, we don't care which merge!
+	if ag.done {
+		return false, nil
+	}
+	return true, nil
+}
+
+// TODO: this algorithm may not be correct in all cases. replace if needed!
+type nonReachabilityGrouper struct {
+	baseGrouper // "inherit" what we want, and reimplement the rest
+}
+
+func (ag *nonReachabilityGrouper) name() string {
+	return "nonReachabilityGrouper"
+}
+
+// this algorithm relies on the observation that if there's a path from a to b,
+// then they *can't* be merged (b/c of the existing dependency) so therefore we
+// merge anything that *doesn't* satisfy this condition or that of the reverse!
+func (ag *nonReachabilityGrouper) vertexNext() (v1, v2 *Vertex, err error) {
+	for {
+		v1, v2, err = ag.baseGrouper.vertexNext() // get all iterable pairs
+		if err != nil {
+			log.Fatalf("Error running autoGroup(vertexNext): %v", err)
+		}
+
+		if v1 != v2 { // ignore self cmp early (perf optimization)
+			// if NOT reachable, they're viable...
+			out1 := ag.graph.Reachability(v1, v2)
+			out2 := ag.graph.Reachability(v2, v1)
+			if len(out1) == 0 && len(out2) == 0 {
+				return // return v1 and v2, they're viable
+			}
+		}
+
+		// if we got here, it means we're skipping over this candidate!
+		if ok, err := ag.baseGrouper.vertexTest(false); err != nil {
+			log.Fatalf("Error running autoGroup(vertexTest): %v", err)
+		} else if !ok {
+			return nil, nil, nil // done!
+		}
+
+		// the vertexTest passed, so loop and try with a new pair...
+	}
+}
+
+// VertexMerge merges v2 into v1 by reattaching the edges where appropriate,
+// and then by deleting v2 from the graph. Since more than one edge between two
+// vertices is not allowed, duplicate edges are merged as well. an edge merge
+// function can be provided if you'd like to control how you merge the edges!
+func (g *Graph) VertexMerge(v1, v2 *Vertex, vertexMergeFn func(*Vertex, *Vertex) (*Vertex, error), edgeMergeFn func(*Edge, *Edge) *Edge) error {
+	// methodology
+	// 1) edges between v1 and v2 are removed
+	//Loop:
+	for k1 := range g.Adjacency {
+		for k2 := range g.Adjacency[k1] {
+			// v1 -> v2 || v2 -> v1
+			if (k1 == v1 && k2 == v2) || (k1 == v2 && k2 == v1) {
+				delete(g.Adjacency[k1], k2) // delete map & edge
+				// NOTE: if we assume this is a DAG, then we can
+				// assume only v1 -> v2 OR v2 -> v1 exists, and
+				// we can break out of these loops immediately!
+				//break Loop
+				break
+			}
+		}
+	}
+
+	// 2) edges that point towards v2 from X now point to v1 from X (no dupes)
+	for _, x := range g.IncomingGraphEdges(v2) { // all to vertex v (??? -> v)
+		e := g.Adjacency[x][v2] // previous edge
+		r := g.Reachability(x, v1)
+		// merge e with ex := g.Adjacency[x][v1] if it exists!
+		if ex, exists := g.Adjacency[x][v1]; exists && edgeMergeFn != nil && len(r) == 0 {
+			e = edgeMergeFn(e, ex)
+		}
+		if len(r) == 0 { // if not reachable, add it
+			g.AddEdge(x, v1, e) // overwrite edge
+		} else if edgeMergeFn != nil { // reachable, merge e through...
+			prev := x // initial condition
+			for i, next := range r {
+				if i == 0 {
+					// next == prev, therefore skip
+					continue
+				}
+				// this edge is from: prev, to: next
+				ex, _ := g.Adjacency[prev][next] // get
+				ex = edgeMergeFn(ex, e)
+				g.Adjacency[prev][next] = ex // set
+				prev = next
+			}
+		}
+		delete(g.Adjacency[x], v2) // delete old edge
+	}
+
+	// 3) edges that point from v2 to X now point from v1 to X (no dupes)
+	for _, x := range g.OutgoingGraphEdges(v2) { // all from vertex v (v -> ???)
+		e := g.Adjacency[v2][x] // previous edge
+		r := g.Reachability(v1, x)
+		// merge e with ex := g.Adjacency[v1][x] if it exists!
+		if ex, exists := g.Adjacency[v1][x]; exists && edgeMergeFn != nil && len(r) == 0 {
+			e = edgeMergeFn(e, ex)
+		}
+		if len(r) == 0 {
+			g.AddEdge(v1, x, e) // overwrite edge
+		} else if edgeMergeFn != nil { // reachable, merge e through...
+			prev := v1 // initial condition
+			for i, next := range r {
+				if i == 0 {
+					// next == prev, therefore skip
+					continue
+				}
+				// this edge is from: prev, to: next
+				ex, _ := g.Adjacency[prev][next]
+				ex = edgeMergeFn(ex, e)
+				g.Adjacency[prev][next] = ex
+				prev = next
+			}
+		}
+		delete(g.Adjacency[v2], x)
+	}
+
+	// 4) merge and then remove the (now merged/grouped) vertex
+	if vertexMergeFn != nil { // run vertex merge function
+		if v, err := vertexMergeFn(v1, v2); err != nil {
+			return err
+		} else if v != nil { // replace v1 with the "merged" version...
+			v1 = v // XXX: will this replace v1 the way we want?
+		}
+	}
+	g.DeleteVertex(v2) // remove grouped vertex
+
+	// 5) creation of a cyclic graph should throw an error
+	if _, err := g.TopologicalSort(); err != nil { // am i a dag or not?
+		return errwrap.Wrapf(err, "TopologicalSort failed") // not a dag
+	}
+	return nil // success
+}
+
+// autoGroup is the mechanical auto group "runner" that runs the interface spec
+func (g *Graph) autoGroup(ag AutoGrouper) chan string {
+	strch := make(chan string) // output log messages here
+	go func(strch chan string) {
+		strch <- fmt.Sprintf("Compile: Grouping: Algorithm: %v...", ag.name())
+		if err := ag.init(g); err != nil {
+			log.Fatalf("Error running autoGroup(init): %v", err)
+		}
+
+		for {
+			var v, w *Vertex
+			v, w, err := ag.vertexNext() // get pair to compare
+			if err != nil {
+				log.Fatalf("Error running autoGroup(vertexNext): %v", err)
+			}
+			merged := false
+			// save names since they change during the runs
+			vStr := fmt.Sprintf("%s", v) // valid even if it is nil
+			wStr := fmt.Sprintf("%s", w)
+
+			if err := ag.vertexCmp(v, w); err != nil { // cmp ?
+				if global.DEBUG {
+					strch <- fmt.Sprintf("Compile: Grouping: !GroupCmp for: %s into %s", wStr, vStr)
+				}
+
+				// remove grouped vertex and merge edges (res is safe)
+			} else if err := g.VertexMerge(v, w, ag.vertexMerge, ag.edgeMerge); err != nil { // merge...
+				strch <- fmt.Sprintf("Compile: Grouping: !VertexMerge for: %s into %s", wStr, vStr)
+
+			} else { // success!
+				strch <- fmt.Sprintf("Compile: Grouping: Success for: %s into %s", wStr, vStr)
+				merged = true // woo
+			}
+
+			// did these get used?
+			if ok, err := ag.vertexTest(merged); err != nil {
+				log.Fatalf("Error running autoGroup(vertexTest): %v", err)
+			} else if !ok {
+				break // done!
+			}
+		}
+
+		close(strch)
+		return
+	}(strch) // call function
+	return strch
+}
+
+// AutoGroup runs the auto grouping on the graph and prints out log messages
+func (g *Graph) AutoGroup() {
+	// receive log messages from channel...
+	// this allows test cases to avoid printing them when they're unwanted!
+	// TODO: this algorithm may not be correct in all cases. replace if needed!
+	for str := range g.autoGroup(&nonReachabilityGrouper{}) {
+		log.Println(str)
+	}
+}

pgraph/pgraph.go

@@ -15,14 +15,14 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-// Pgraph (Pointer Graph)
+// Package pgraph represents the internal "pointer graph" that we use.
-package main
+package pgraph
 
 import (
-	"errors"
 	"fmt"
 	"io/ioutil"
 	"log"
+	"math"
 	"os"
 	"os/exec"
 	"sort"
@@ -30,6 +30,13 @@ import (
 	"sync"
 	"syscall"
 	"time"
+
+	"github.com/purpleidea/mgmt/converger"
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/global"
+	"github.com/purpleidea/mgmt/resources"
+
+	errwrap "github.com/pkg/errors"
 )
 
 //go:generate stringer -type=graphState -output=graphstate_stringer.go
@@ -43,6 +50,7 @@ const (
 	graphStatePaused
 )
 
+// Graph is the graph structure in this library.
 // The graph abstract data type (ADT) is defined as follows:
 // * the directed graph arrows point from left to right ( -> )
 // * the arrows point away from their dependencies (eg: arrows mean "before")
@@ -55,15 +63,18 @@ type Graph struct {
 	mutex     sync.Mutex // used when modifying graph State variable
 }
 
+// Vertex is the primary vertex struct in this library.
 type Vertex struct {
-	Res             // anonymous field
+	resources.Res       // anonymous field
-	timestamp int64 // last updated timestamp ?
+	timestamp     int64 // last updated timestamp ?
 }
 
+// Edge is the primary edge struct in this library.
 type Edge struct {
 	Name string
 }
 
+// NewGraph builds a new graph.
 func NewGraph(name string) *Graph {
 	return &Graph{
 		Name:      name,
@@ -72,12 +83,14 @@ func NewGraph(name string) *Graph {
 	}
 }
 
-func NewVertex(r Res) *Vertex {
+// NewVertex returns a new graph vertex struct with a contained resource.
+func NewVertex(r resources.Res) *Vertex {
 	return &Vertex{
 		Res: r,
 	}
 }
 
+// NewEdge returns a new graph edge struct.
 func NewEdge(name string) *Edge {
 	return &Edge{
 		Name: name,
@@ -97,38 +110,34 @@ func (g *Graph) Copy() *Graph {
 	return newGraph
 }
 
-// returns the name of the graph
+// GetName returns the name of the graph.
 func (g *Graph) GetName() string {
 	return g.Name
 }
 
-// set name of the graph
+// SetName sets the name of the graph.
 func (g *Graph) SetName(name string) {
 	g.Name = name
 }
 
-func (g *Graph) GetState() graphState {
+// getState returns the state of the graph. This state is used for optimizing
+// certain algorithms by knowing what part of processing the graph is currently
+// undergoing.
+func (g *Graph) getState() graphState {
 	//g.mutex.Lock()
 	//defer g.mutex.Unlock()
 	return g.state
 }
 
-// set graph state and return previous state
+// setState sets the graph state and returns the previous state.
-func (g *Graph) SetState(state graphState) graphState {
+func (g *Graph) setState(state graphState) graphState {
 	g.mutex.Lock()
 	defer g.mutex.Unlock()
-	prev := g.GetState()
+	prev := g.getState()
 	g.state = state
 	return prev
 }
 
-// store a pointer in the resource to it's parent vertex
-func (g *Graph) SetVertex() {
-	for v := range g.GetVerticesChan() {
-		v.Res.SetVertex(v)
-	}
-}
-
 // AddVertex uses variadic input to add all listed vertices to the graph
 func (g *Graph) AddVertex(xv ...*Vertex) {
 	for _, v := range xv {
@@ -138,6 +147,7 @@ func (g *Graph) AddVertex(xv ...*Vertex) {
 	}
 }
 
+// DeleteVertex deletes a particular vertex from the graph.
 func (g *Graph) DeleteVertex(v *Vertex) {
 	delete(g.Adjacency, v)
 	for k := range g.Adjacency {
@@ -145,7 +155,7 @@ func (g *Graph) DeleteVertex(v *Vertex) {
 	}
 }
 
-// adds a directed edge to the graph from v1 to v2
+// AddEdge adds a directed edge to the graph from v1 to v2.
 func (g *Graph) AddEdge(v1, v2 *Vertex, e *Edge) {
 	// NOTE: this doesn't allow more than one edge between two vertexes...
 	g.AddVertex(v1, v2) // supports adding N vertices now
@@ -154,7 +164,21 @@ func (g *Graph) AddEdge(v1, v2 *Vertex, e *Edge) {
 	g.Adjacency[v1][v2] = e
 }
 
-func (g *Graph) GetVertexMatch(obj Res) *Vertex {
+// DeleteEdge deletes a particular edge from the graph.
+// FIXME: add test cases
+func (g *Graph) DeleteEdge(e *Edge) {
+	for v1 := range g.Adjacency {
+		for v2, edge := range g.Adjacency[v1] {
+			if e == edge {
+				delete(g.Adjacency[v1], v2)
+			}
+		}
+	}
+}
+
+// GetVertexMatch searches for an equivalent resource in the graph and returns
+// the vertex it is found in, or nil if not found.
+func (g *Graph) GetVertexMatch(obj resources.Res) *Vertex {
 	for k := range g.Adjacency {
 		if k.Res.Compare(obj) {
 			return k
@@ -163,6 +187,7 @@ func (g *Graph) GetVertexMatch(obj Res) *Vertex {
 	return nil
 }
 
+// HasVertex returns if the input vertex exists in the graph.
 func (g *Graph) HasVertex(v *Vertex) bool {
 	if _, exists := g.Adjacency[v]; exists {
 		return true
@@ -170,12 +195,12 @@ func (g *Graph) HasVertex(v *Vertex) bool {
 	return false
 }
 
-// number of vertices in the graph
+// NumVertices returns the number of vertices in the graph.
 func (g *Graph) NumVertices() int {
 	return len(g.Adjacency)
 }
 
-// number of edges in the graph
+// NumEdges returns the number of edges in the graph.
 func (g *Graph) NumEdges() int {
 	count := 0
 	for k := range g.Adjacency {
@@ -194,7 +219,7 @@ func (g *Graph) GetVertices() []*Vertex {
 	return vertices
 }
 
-// returns a channel of all vertices in the graph
+// GetVerticesChan returns a channel of all vertices in the graph.
 func (g *Graph) GetVerticesChan() chan *Vertex {
 	ch := make(chan *Vertex)
 	go func(ch chan *Vertex) {
@@ -206,6 +231,7 @@ func (g *Graph) GetVerticesChan() chan *Vertex {
 	return ch
 }
 
+// VertexSlice is a linear list of vertices. It can be sorted.
 type VertexSlice []*Vertex
 
 func (vs VertexSlice) Len() int           { return len(vs) }
@@ -223,7 +249,7 @@ func (g *Graph) GetVerticesSorted() []*Vertex {
 	return vertices
 }
 
-// make the graph pretty print
+// String makes the graph pretty print.
 func (g *Graph) String() string {
 	return fmt.Sprintf("Vertices(%d), Edges(%d)", g.NumVertices(), g.NumEdges())
 }
@@ -233,7 +259,7 @@ func (v *Vertex) String() string {
 	return fmt.Sprintf("%s[%s]", v.Res.Kind(), v.Res.GetName())
 }
 
-// output the graph in graphviz format
+// Graphviz outputs the graph in graphviz format.
 // https://en.wikipedia.org/wiki/DOT_%28graph_description_language%29
 func (g *Graph) Graphviz() (out string) {
 	//digraph g {
@@ -265,17 +291,18 @@ func (g *Graph) Graphviz() (out string) {
 	return
 }
 
-// write out the graphviz data and run the correct graphviz filter command
+// ExecGraphviz writes out the graphviz data and runs the correct graphviz
+// filter command.
 func (g *Graph) ExecGraphviz(program, filename string) error {
 
 	switch program {
 	case "dot", "neato", "twopi", "circo", "fdp":
 	default:
-		return errors.New("Invalid graphviz program selected!")
+		return fmt.Errorf("Invalid graphviz program selected!")
 	}
 
 	if filename == "" {
-		return errors.New("No filename given!")
+		return fmt.Errorf("No filename given!")
 	}
 
 	// run as a normal user if possible when run with sudo
@@ -284,18 +311,18 @@ func (g *Graph) ExecGraphviz(program, filename string) error {
 
 	err := ioutil.WriteFile(filename, []byte(g.Graphviz()), 0644)
 	if err != nil {
-		return errors.New("Error writing to filename!")
+		return fmt.Errorf("Error writing to filename!")
 	}
 
 	if err1 == nil && err2 == nil {
 		if err := os.Chown(filename, uid, gid); err != nil {
-			return errors.New("Error changing file owner!")
+			return fmt.Errorf("Error changing file owner!")
 		}
 	}
 
 	path, err := exec.LookPath(program)
 	if err != nil {
-		return errors.New("Graphviz is missing!")
+		return fmt.Errorf("Graphviz is missing!")
 	}
 
 	out := fmt.Sprintf("%v.png", filename)
@@ -310,13 +337,13 @@ func (g *Graph) ExecGraphviz(program, filename string) error {
 	}
 	_, err = cmd.Output()
 	if err != nil {
-		return errors.New("Error writing to image!")
+		return fmt.Errorf("Error writing to image!")
 	}
 	return nil
 }
 
-// return an array (slice) of all directed vertices to vertex v (??? -> v)
+// IncomingGraphEdges returns an array (slice) of all directed vertices to
-// OKTimestamp should use this
+// vertex v (??? -> v). OKTimestamp should probably use this.
 func (g *Graph) IncomingGraphEdges(v *Vertex) []*Vertex {
 	// TODO: we might be able to implement this differently by reversing
 	// the Adjacency graph and then looping through it again...
@@ -331,8 +358,8 @@ func (g *Graph) IncomingGraphEdges(v *Vertex) []*Vertex {
 	return s
 }
 
-// return an array (slice) of all vertices that vertex v points to (v -> ???)
+// OutgoingGraphEdges returns an array (slice) of all vertices that vertex v
-// poke should use this
+// points to (v -> ???). Poke should probably use this.
 func (g *Graph) OutgoingGraphEdges(v *Vertex) []*Vertex {
 	var s []*Vertex
 	for k := range g.Adjacency[v] { // forward paths
@@ -341,7 +368,8 @@ func (g *Graph) OutgoingGraphEdges(v *Vertex) []*Vertex {
 	return s
 }
 
-// return an array (slice) of all vertices that connect to vertex v
+// GraphEdges returns an array (slice) of all vertices that connect to vertex v.
+// This is the union of IncomingGraphEdges and OutgoingGraphEdges.
 func (g *Graph) GraphEdges(v *Vertex) []*Vertex {
 	var s []*Vertex
 	s = append(s, g.IncomingGraphEdges(v)...)
@@ -349,6 +377,7 @@ func (g *Graph) GraphEdges(v *Vertex) []*Vertex {
 	return s
 }
 
+// DFS returns a depth first search for the graph, starting at the input vertex.
 func (g *Graph) DFS(start *Vertex) []*Vertex {
 	var d []*Vertex // discovered
 	var s []*Vertex // stack
@@ -371,7 +400,7 @@ func (g *Graph) DFS(start *Vertex) []*Vertex {
 	return d
 }
 
-// build a new graph containing only vertices from the list...
+// FilterGraph builds a new graph containing only vertices from the list.
 func (g *Graph) FilterGraph(name string, vertices []*Vertex) *Graph {
 	newgraph := NewGraph(name)
 	for k1, x := range g.Adjacency {
@@ -385,8 +414,8 @@ func (g *Graph) FilterGraph(name string, vertices []*Vertex) *Graph {
 	return newgraph
 }
 
-// return a channel containing the N disconnected graphs in our main graph
+// GetDisconnectedGraphs returns a channel containing the N disconnected graphs
-// we can then process each of these in parallel
+// in our main graph. We can then process each of these in parallel.
 func (g *Graph) GetDisconnectedGraphs() chan *Graph {
 	ch := make(chan *Graph)
 	go func() {
@@ -421,8 +450,7 @@ func (g *Graph) GetDisconnectedGraphs() chan *Graph {
 	return ch
 }
 
-// return the indegree for the graph, IOW the count of vertices that point to me
+// InDegree returns the count of vertices that point to me in one big lookup map.
-// NOTE: this returns the values for all vertices in one big lookup table
 func (g *Graph) InDegree() map[*Vertex]int {
 	result := make(map[*Vertex]int)
 	for k := range g.Adjacency {
@@ -437,24 +465,23 @@ func (g *Graph) InDegree() map[*Vertex]int {
 	return result
 }
 
-// return the outdegree for the graph, IOW the count of vertices that point away
+// OutDegree returns the count of vertices that point away in one big lookup map.
-// NOTE: this returns the values for all vertices in one big lookup table
 func (g *Graph) OutDegree() map[*Vertex]int {
 	result := make(map[*Vertex]int)
 
 	for k := range g.Adjacency {
 		result[k] = 0 // initialize
-		for _ = range g.Adjacency[k] {
+		for range g.Adjacency[k] {
 			result[k]++
 		}
 	}
 	return result
 }
 
-// returns a topological sort for the graph
+// TopologicalSort returns the sort of graph vertices in that order.
 // based on descriptions and code from wikipedia and rosetta code
 // TODO: add memoization, and cache invalidation to speed this up :)
-func (g *Graph) TopologicalSort() (result []*Vertex, ok bool) { // kahn's algorithm
+func (g *Graph) TopologicalSort() ([]*Vertex, error) { // kahn's algorithm
 	var L []*Vertex                    // empty list that will contain the sorted elements
 	var S []*Vertex                    // set of all nodes with no incoming edges
 	remaining := make(map[*Vertex]int) // amount of edges remaining
@@ -491,13 +518,13 @@ func (g *Graph) TopologicalSort() (result []*Vertex, ok bool) { // kahn's algori
 		if in > 0 {
 			for n := range g.Adjacency[c] {
 				if remaining[n] > 0 {
-					return nil, false // not a dag!
+					return nil, fmt.Errorf("Not a dag!")
 				}
 			}
 		}
 	}
 
-	return L, true
+	return L, nil
 }
 
 // Reachability finds the shortest path in a DAG from a to b, and returns the
@@ -540,108 +567,6 @@ func (g *Graph) Reachability(a, b *Vertex) []*Vertex {
 	return result
 }
 
-// VertexMerge merges v2 into v1 by reattaching the edges where appropriate,
-// and then by deleting v2 from the graph. Since more than one edge between two
-// vertices is not allowed, duplicate edges are merged as well. an edge merge
-// function can be provided if you'd like to control how you merge the edges!
-func (g *Graph) VertexMerge(v1, v2 *Vertex, vertexMergeFn func(*Vertex, *Vertex) (*Vertex, error), edgeMergeFn func(*Edge, *Edge) *Edge) error {
-	// methodology
-	// 1) edges between v1 and v2 are removed
-	//Loop:
-	for k1 := range g.Adjacency {
-		for k2 := range g.Adjacency[k1] {
-			// v1 -> v2 || v2 -> v1
-			if (k1 == v1 && k2 == v2) || (k1 == v2 && k2 == v1) {
-				delete(g.Adjacency[k1], k2) // delete map & edge
-				// NOTE: if we assume this is a DAG, then we can
-				// assume only v1 -> v2 OR v2 -> v1 exists, and
-				// we can break out of these loops immediately!
-				//break Loop
-				break
-			}
-		}
-	}
-
-	// 2) edges that point towards v2 from X now point to v1 from X (no dupes)
-	for _, x := range g.IncomingGraphEdges(v2) { // all to vertex v (??? -> v)
-		e := g.Adjacency[x][v2] // previous edge
-		r := g.Reachability(x, v1)
-		// merge e with ex := g.Adjacency[x][v1] if it exists!
-		if ex, exists := g.Adjacency[x][v1]; exists && edgeMergeFn != nil && len(r) == 0 {
-			e = edgeMergeFn(e, ex)
-		}
-		if len(r) == 0 { // if not reachable, add it
-			g.AddEdge(x, v1, e) // overwrite edge
-		} else if edgeMergeFn != nil { // reachable, merge e through...
-			prev := x // initial condition
-			for i, next := range r {
-				if i == 0 {
-					// next == prev, therefore skip
-					continue
-				}
-				// this edge is from: prev, to: next
-				ex, _ := g.Adjacency[prev][next] // get
-				ex = edgeMergeFn(ex, e)
-				g.Adjacency[prev][next] = ex // set
-				prev = next
-			}
-		}
-		delete(g.Adjacency[x], v2) // delete old edge
-	}
-
-	// 3) edges that point from v2 to X now point from v1 to X (no dupes)
-	for _, x := range g.OutgoingGraphEdges(v2) { // all from vertex v (v -> ???)
-		e := g.Adjacency[v2][x] // previous edge
-		r := g.Reachability(v1, x)
-		// merge e with ex := g.Adjacency[v1][x] if it exists!
-		if ex, exists := g.Adjacency[v1][x]; exists && edgeMergeFn != nil && len(r) == 0 {
-			e = edgeMergeFn(e, ex)
-		}
-		if len(r) == 0 {
-			g.AddEdge(v1, x, e) // overwrite edge
-		} else if edgeMergeFn != nil { // reachable, merge e through...
-			prev := v1 // initial condition
-			for i, next := range r {
-				if i == 0 {
-					// next == prev, therefore skip
-					continue
-				}
-				// this edge is from: prev, to: next
-				ex, _ := g.Adjacency[prev][next]
-				ex = edgeMergeFn(ex, e)
-				g.Adjacency[prev][next] = ex
-				prev = next
-			}
-		}
-		delete(g.Adjacency[v2], x)
-	}
-
-	// 4) merge and then remove the (now merged/grouped) vertex
-	if vertexMergeFn != nil { // run vertex merge function
-		if v, err := vertexMergeFn(v1, v2); err != nil {
-			return err
-		} else if v != nil { // replace v1 with the "merged" version...
-			v1 = v // XXX: will this replace v1 the way we want?
-		}
-	}
-	g.DeleteVertex(v2) // remove grouped vertex
-
-	// 5) creation of a cyclic graph should throw an error
-	if _, dag := g.TopologicalSort(); !dag { // am i a dag or not?
-		return fmt.Errorf("Graph is not a dag!")
-	}
-	return nil // success
-}
-
-func HeisenbergCount(ch chan *Vertex) int {
-	c := 0
-	for x := range ch {
-		_ = x
-		c++
-	}
-	return c
-}
-
 // GetTimestamp returns the timestamp of a vertex
 func (v *Vertex) GetTimestamp() int64 {
 	return v.timestamp
@@ -653,7 +578,7 @@ func (v *Vertex) UpdateTimestamp() int64 {
 	return v.timestamp
 }
 
-// can this element run right now?
+// OKTimestamp returns true if this element can run right now?
 func (g *Graph) OKTimestamp(v *Vertex) bool {
 	// these are all the vertices pointing TO v, eg: ??? -> v
 	for _, n := range g.IncomingGraphEdges(v) {
@@ -662,7 +587,7 @@ func (g *Graph) OKTimestamp(v *Vertex) bool {
 		// if they're equal (eg: on init of 0) then we also can't run
 		// b/c we should let our pre-req's go first...
 		x, y := v.GetTimestamp(), n.GetTimestamp()
-		if DEBUG {
+		if global.DEBUG {
 			log.Printf("%v[%v]: OKTimestamp: (%v) >= %v[%v](%v): !%v", v.Kind(), v.GetName(), x, n.Kind(), n.GetName(), y, x >= y)
 		}
 		if x >= y {
@@ -672,79 +597,79 @@ func (g *Graph) OKTimestamp(v *Vertex) bool {
 	return true
 }
 
-// notify nodes after me in the dependency graph that they need refreshing...
+// Poke notifies nodes after me in the dependency graph that they need refreshing...
 // NOTE: this assumes that this can never fail or need to be rescheduled
 func (g *Graph) Poke(v *Vertex, activity bool) {
 	// these are all the vertices pointing AWAY FROM v, eg: v -> ???
 	for _, n := range g.OutgoingGraphEdges(v) {
 		// XXX: if we're in state event and haven't been cancelled by
 		// apply, then we can cancel a poke to a child, right? XXX
-		// XXX: if n.Res.GetState() != resStateEvent { // is this correct?
+		// XXX: if n.Res.getState() != resources.ResStateEvent { // is this correct?
 		if true { // XXX
-			if DEBUG {
+			if global.DEBUG {
 				log.Printf("%v[%v]: Poke: %v[%v]", v.Kind(), v.GetName(), n.Kind(), n.GetName())
 			}
-			n.SendEvent(eventPoke, false, activity) // XXX: can this be switched to sync?
+			n.SendEvent(event.EventPoke, false, activity) // XXX: can this be switched to sync?
 		} else {
-			if DEBUG {
+			if global.DEBUG {
 				log.Printf("%v[%v]: Poke: %v[%v]: Skipped!", v.Kind(), v.GetName(), n.Kind(), n.GetName())
 			}
 		}
 	}
 }
 
-// poke the pre-requisites that are stale and need to run before I can run...
+// BackPoke pokes the pre-requisites that are stale and need to run before I can run.
 func (g *Graph) BackPoke(v *Vertex) {
 	// these are all the vertices pointing TO v, eg: ??? -> v
 	for _, n := range g.IncomingGraphEdges(v) {
 		x, y, s := v.GetTimestamp(), n.GetTimestamp(), n.Res.GetState()
 		// if the parent timestamp needs poking AND it's not in state
-		// resStateEvent, then poke it. If the parent is in resStateEvent it
+		// ResStateEvent, then poke it. If the parent is in ResStateEvent it
 		// means that an event is pending, so we'll be expecting a poke
 		// back soon, so we can safely discard the extra parent poke...
 		// TODO: implement a stateLT (less than) to tell if something
 		// happens earlier in the state cycle and that doesn't wrap nil
-		if x >= y && (s != resStateEvent && s != resStateCheckApply) {
+		if x >= y && (s != resources.ResStateEvent && s != resources.ResStateCheckApply) {
-			if DEBUG {
+			if global.DEBUG {
 				log.Printf("%v[%v]: BackPoke: %v[%v]", v.Kind(), v.GetName(), n.Kind(), n.GetName())
 			}
-			n.SendEvent(eventBackPoke, false, false) // XXX: can this be switched to sync?
+			n.SendEvent(event.EventBackPoke, false, false) // XXX: can this be switched to sync?
 		} else {
-			if DEBUG {
+			if global.DEBUG {
 				log.Printf("%v[%v]: BackPoke: %v[%v]: Skipped!", v.Kind(), v.GetName(), n.Kind(), n.GetName())
 			}
 		}
 	}
 }
 
-// XXX: rename this function
+// Process is the primary function to execute for a particular vertex in the graph.
-func (g *Graph) Process(v *Vertex) {
+func (g *Graph) Process(v *Vertex) error {
 	obj := v.Res
-	if DEBUG {
+	if global.DEBUG {
 		log.Printf("%v[%v]: Process()", obj.Kind(), obj.GetName())
 	}
-	obj.SetState(resStateEvent)
+	obj.SetState(resources.ResStateEvent)
 	var ok = true
 	var apply = false // did we run an apply?
 	// is it okay to run dependency wise right now?
 	// if not, that's okay because when the dependency runs, it will poke
 	// us back and we will run if needed then!
 	if g.OKTimestamp(v) {
-		if DEBUG {
+		if global.DEBUG {
 			log.Printf("%v[%v]: OKTimestamp(%v)", obj.Kind(), obj.GetName(), v.GetTimestamp())
 		}
 
-		obj.SetState(resStateCheckApply)
+		obj.SetState(resources.ResStateCheckApply)
 		// if this fails, don't UpdateTimestamp()
-		stateok, err := obj.CheckApply(true)
+		checkok, err := obj.CheckApply(!obj.Meta().Noop)
-		if stateok && err != nil { // should never return this way
+		if checkok && err != nil { // should never return this way
-			log.Fatalf("%v[%v]: CheckApply(): %t, %+v", obj.Kind(), obj.GetName(), stateok, err)
+			log.Fatalf("%v[%v]: CheckApply(): %t, %+v", obj.Kind(), obj.GetName(), checkok, err)
 		}
-		if DEBUG {
+		if global.DEBUG {
-			log.Printf("%v[%v]: CheckApply(): %t, %v", obj.Kind(), obj.GetName(), stateok, err)
+			log.Printf("%v[%v]: CheckApply(): %t, %v", obj.Kind(), obj.GetName(), checkok, err)
 		}
 
-		if !stateok { // if state *was* not ok, we had to have apply'ed
+		if !checkok { // if state *was* not ok, we had to have apply'ed
 			if err != nil { // error during check or apply
 				ok = false
 			} else {
@@ -752,24 +677,205 @@ func (g *Graph) Process(v *Vertex) {
 			}
 		}
 
+		// when noop is true we always want to update timestamp
+		if obj.Meta().Noop && err == nil {
+			ok = true
+		}
+
 		if ok {
 			// update this timestamp *before* we poke or the poked
 			// nodes might fail due to having a too old timestamp!
-			v.UpdateTimestamp()          // this was touched...
+			v.UpdateTimestamp()                    // this was touched...
-			obj.SetState(resStatePoking) // can't cancel parent poke
+			obj.SetState(resources.ResStatePoking) // can't cancel parent poke
 			g.Poke(v, apply)
 		}
 		// poke at our pre-req's instead since they need to refresh/run...
-	} else {
+		return err
-		// only poke at the pre-req's that need to run
-		go g.BackPoke(v)
 	}
+	// else... only poke at the pre-req's that need to run
+	go g.BackPoke(v)
+	return nil
 }
 
-// main kick to start the graph
+// SentinelErr is a sentinal as an error type that wraps an arbitrary error.
+type SentinelErr struct {
+	err error
+}
+
+// Error is the required method to fulfill the error type.
+func (obj *SentinelErr) Error() string {
+	return obj.err.Error()
+}
+
+// Worker is the common run frontend of the vertex. It handles all of the retry
+// and retry delay common code, and ultimately returns the final status of this
+// vertex execution.
+func (g *Graph) Worker(v *Vertex) error {
+	// listen for chan events from Watch() and run
+	// the Process() function when they're received
+	// this avoids us having to pass the data into
+	// the Watch() function about which graph it is
+	// running on, which isolates things nicely...
+	obj := v.Res
+	chanProcess := make(chan event.Event)
+	go func() {
+		running := false
+		var timer = time.NewTimer(time.Duration(math.MaxInt64)) // longest duration
+		if !timer.Stop() {
+			<-timer.C // unnecessary, shouldn't happen
+		}
+		var delay = time.Duration(v.Meta().Delay) * time.Millisecond
+		var retry = v.Meta().Retry // number of tries left, -1 for infinite
+		var saved event.Event
+	Loop:
+		for {
+			// this has to be synchronous, because otherwise the Res
+			// event loop will keep running and change state,
+			// causing the converged timeout to fire!
+			select {
+			case event, ok := <-chanProcess: // must use like this
+				if running && ok {
+					// we got an event that wasn't a close,
+					// while we were waiting for the timer!
+					// if this happens, it might be a bug:(
+					log.Fatalf("%v[%v]: Worker: Unexpected event: %+v", v.Kind(), v.GetName(), event)
+				}
+				if !ok { // chanProcess closed, let's exit
+					break Loop // no event, so no ack!
+				}
+
+				// the above mentioned synchronous part, is the
+				// running of this function, paired with an ack.
+				if e := g.Process(v); e != nil {
+					saved = event
+					log.Printf("%v[%v]: CheckApply errored: %v", v.Kind(), v.GetName(), e)
+					if retry == 0 {
+						// wrap the error in the sentinel
+						event.ACKNACK(&SentinelErr{e}) // fail the Watch()
+						break Loop
+					}
+					if retry > 0 { // don't decrement the -1
+						retry--
+					}
+					log.Printf("%v[%v]: CheckApply: Retrying after %.4f seconds (%d left)", v.Kind(), v.GetName(), delay.Seconds(), retry)
+					// start the timer...
+					timer.Reset(delay)
+					running = true
+					continue
+				}
+				retry = v.Meta().Retry // reset on success
+				event.ACK()            // sync
+
+			case <-timer.C:
+				if !timer.Stop() {
+					//<-timer.C // blocks, docs are wrong!
+				}
+				running = false
+				log.Printf("%s[%s]: CheckApply delay expired!", v.Kind(), v.GetName())
+				// re-send this failed event, to trigger a CheckApply()
+				go func() { chanProcess <- saved }()
+				// TODO: should we send a fake event instead?
+				//saved = nil
+			}
+		}
+	}()
+	var err error // propagate the error up (this is a permanent BAD error!)
+	// the watch delay runs inside of the Watch resource loop, so that it
+	// can still process signals and exit if needed. It shouldn't run any
+	// resource specific code since this is supposed to be a retry delay.
+	// NOTE: we're using the same retry and delay metaparams that CheckApply
+	// uses. This is for practicality. We can separate them later if needed!
+	var watchDelay time.Duration
+	var watchRetry = v.Meta().Retry // number of tries left, -1 for infinite
+	// watch blocks until it ends, & errors to retry
+	for {
+		// TODO: do we have to stop the converged-timeout when in this block (perhaps we're in the delay block!)
+		// TODO: should we setup/manage some of the converged timeout stuff in here anyways?
+
+		// if a retry-delay was requested, wait, but don't block our events!
+		if watchDelay > 0 {
+			//var pendingSendEvent bool
+			timer := time.NewTimer(watchDelay)
+		Loop:
+			for {
+				select {
+				case <-timer.C: // the wait is over
+					break Loop // critical
+
+				// TODO: resources could have a separate exit channel to avoid this complexity!?
+				case event := <-obj.Events():
+					// NOTE: this code should match the similar Res code!
+					//cuid.SetConverged(false) // TODO: ?
+					if exit, send := obj.ReadEvent(&event); exit {
+						return nil // exit
+					} else if send {
+						// if we dive down this rabbit hole, our
+						// timer.C won't get seen until we get out!
+						// in this situation, the Watch() is blocked
+						// from performing until CheckApply returns
+						// successfully, or errors out. This isn't
+						// so bad, but we should document it. Is it
+						// possible that some resource *needs* Watch
+						// to run to be able to execute a CheckApply?
+						// That situation shouldn't be common, and
+						// should probably not be allowed. Can we
+						// avoid it though?
+						//if exit, err := doSend(); exit || err != nil {
+						//	return err // we exit or bubble up a NACK...
+						//}
+						// Instead of doing the above, we can
+						// add events to a pending list, and
+						// when we finish the delay, we can run
+						// them.
+						//pendingSendEvent = true // all events are identical for now...
+					}
+				}
+			}
+			timer.Stop() // it's nice to cleanup
+			log.Printf("%s[%s]: Watch delay expired!", v.Kind(), v.GetName())
+			// NOTE: we can avoid the send if running Watch guarantees
+			// one CheckApply event on startup!
+			//if pendingSendEvent { // TODO: should this become a list in the future?
+			//	if exit, err := obj.DoSend(chanProcess, ""); exit || err != nil {
+			//		return err // we exit or bubble up a NACK...
+			//	}
+			//}
+		}
+
+		// TODO: reset the watch retry count after some amount of success
+		e := v.Res.Watch(chanProcess)
+		if e == nil { // exit signal
+			err = nil // clean exit
+			break
+		}
+		if sentinelErr, ok := e.(*SentinelErr); ok { // unwrap the sentinel
+			err = sentinelErr.err
+			break // sentinel means, perma-exit
+		}
+		log.Printf("%v[%v]: Watch errored: %v", v.Kind(), v.GetName(), e)
+		if watchRetry == 0 {
+			err = fmt.Errorf("Permanent watch error: %v", e)
+			break
+		}
+		if watchRetry > 0 { // don't decrement the -1
+			watchRetry--
+		}
+		watchDelay = time.Duration(v.Meta().Delay) * time.Millisecond
+		log.Printf("%v[%v]: Watch: Retrying after %.4f seconds (%d left)", v.Kind(), v.GetName(), watchDelay.Seconds(), watchRetry)
+		// We need to trigger a CheckApply after Watch restarts, so that
+		// we catch any lost events that happened while down. We do this
+		// by getting the Watch resource to send one event once it's up!
+		//v.SendEvent(eventPoke, false, false)
+	}
+	close(chanProcess)
+	return err
+}
+
+// Start is a main kick to start the graph. It goes through in reverse topological
+// sort order so that events can't hit un-started vertices.
 func (g *Graph) Start(wg *sync.WaitGroup, first bool) { // start or continue
-	log.Printf("State: %v -> %v", g.SetState(graphStateStarting), g.GetState())
+	log.Printf("State: %v -> %v", g.setState(graphStateStarting), g.getState())
-	defer log.Printf("State: %v -> %v", g.SetState(graphStateStarted), g.GetState())
+	defer log.Printf("State: %v -> %v", g.setState(graphStateStarted), g.getState())
 	t, _ := g.TopologicalSort()
 	// TODO: only calculate indegree if `first` is true to save resources
 	indegree := g.InDegree() // compute all of the indegree's
@@ -781,25 +887,13 @@ func (g *Graph) Start(wg *sync.WaitGroup, first bool) { // start or continue
 			// see: https://ttboj.wordpress.com/2015/07/27/golang-parallelism-issues-causing-too-many-open-files-error/
 			go func(vv *Vertex) {
 				defer wg.Done()
-				// listen for chan events from Watch() and run
+				// TODO: if a sufficient number of workers error,
-				// the Process() function when they're received
+				// should something be done? Will these restart
-				// this avoids us having to pass the data into
+				// after perma-failure if we have a graph change?
-				// the Watch() function about which graph it is
+				if err := g.Worker(vv); err != nil { // contains the Watch and CheckApply loops
-				// running on, which isolates things nicely...
+					log.Printf("%s[%s]: Exited with failure: %v", vv.Kind(), vv.GetName(), err)
-				chanProcess := make(chan Event)
+					return
-				go func() {
+				}
-					for event := range chanProcess {
-						// this has to be synchronous,
-						// because otherwise the Res
-						// event loop will keep running
-						// and change state, causing the
-						// converged timeout to fire!
-						g.Process(vv)
-						event.ACK() // sync
-					}
-				}()
-				vv.Res.Watch(chanProcess) // i block until i end
-				close(chanProcess)
 				log.Printf("%v[%v]: Exited", vv.Kind(), vv.GetName())
 			}(v)
 		}
@@ -818,8 +912,8 @@ func (g *Graph) Start(wg *sync.WaitGroup, first bool) { // start or continue
 		// and not just selectively the subset with no indegree.
 		if (!first) || indegree[v] == 0 {
 			// ensure state is started before continuing on to next vertex
-			for !v.SendEvent(eventStart, true, false) {
+			for !v.SendEvent(event.EventStart, true, false) {
-				if DEBUG {
+				if global.DEBUG {
 					// if SendEvent fails, we aren't up yet
 					log.Printf("%v[%v]: Retrying SendEvent(Start)", v.Kind(), v.GetName())
 					// sleep here briefly or otherwise cause
@@ -831,15 +925,17 @@ func (g *Graph) Start(wg *sync.WaitGroup, first bool) { // start or continue
 	}
 }
 
+// Pause sends pause events to the graph in a topological sort order.
 func (g *Graph) Pause() {
-	log.Printf("State: %v -> %v", g.SetState(graphStatePausing), g.GetState())
+	log.Printf("State: %v -> %v", g.setState(graphStatePausing), g.getState())
-	defer log.Printf("State: %v -> %v", g.SetState(graphStatePaused), g.GetState())
+	defer log.Printf("State: %v -> %v", g.setState(graphStatePaused), g.getState())
 	t, _ := g.TopologicalSort()
 	for _, v := range t { // squeeze out the events...
-		v.SendEvent(eventPause, true, false)
+		v.SendEvent(event.EventPause, true, false)
 	}
 }
 
+// Exit sends exit events to the graph in a topological sort order.
 func (g *Graph) Exit() {
 	if g == nil {
 		return
@@ -847,21 +943,111 @@ func (g *Graph) Exit() {
 	t, _ := g.TopologicalSort()
 	for _, v := range t { // squeeze out the events...
 		// turn off the taps...
+		// XXX: consider instead doing this by closing the Res.events channel instead?
 		// XXX: do this by sending an exit signal, and then returning
 		// when we hit the 'default' in the select statement!
 		// XXX: we can do this to quiesce, but it's not necessary now
 
-		v.SendEvent(eventExit, true, false)
+		v.SendEvent(event.EventExit, true, false)
 	}
 }
 
-func (g *Graph) SetConvergedCallback(ctimeout int, converged chan bool) {
+// GraphSync updates the oldGraph so that it matches the newGraph receiver. It
+// leaves identical elements alone so that they don't need to be refreshed.
+// FIXME: add test cases
+func (g *Graph) GraphSync(oldGraph *Graph) (*Graph, error) {
+
+	if oldGraph == nil {
+		oldGraph = NewGraph(g.GetName()) // copy over the name
+	}
+	oldGraph.SetName(g.GetName()) // overwrite the name
+
+	var lookup = make(map[*Vertex]*Vertex)
+	var vertexKeep []*Vertex // list of vertices which are the same in new graph
+	var edgeKeep []*Edge     // list of vertices which are the same in new graph
+
+	for v := range g.Adjacency { // loop through the vertices (resources)
+		res := v.Res // resource
+
+		vertex := oldGraph.GetVertexMatch(res)
+		if vertex == nil { // no match found
+			if err := res.Init(); err != nil {
+				return nil, errwrap.Wrapf(err, "could not Init() resource")
+			}
+			vertex = NewVertex(res)
+			oldGraph.AddVertex(vertex) // call standalone in case not part of an edge
+		}
+		lookup[v] = vertex                      // used for constructing edges
+		vertexKeep = append(vertexKeep, vertex) // append
+	}
+
+	// get rid of any vertices we shouldn't keep (that aren't in new graph)
+	for v := range oldGraph.Adjacency {
+		if !VertexContains(v, vertexKeep) {
+			// wait for exit before starting new graph!
+			v.SendEvent(event.EventExit, true, false)
+			oldGraph.DeleteVertex(v)
+		}
+	}
+
+	// compare edges
+	for v1 := range g.Adjacency { // loop through the vertices (resources)
+		for v2, e := range g.Adjacency[v1] {
+			// we have an edge!
+
+			// lookup vertices (these should exist now)
+			//res1 := v1.Res // resource
+			//res2 := v2.Res
+			//vertex1 := oldGraph.GetVertexMatch(res1)
+			//vertex2 := oldGraph.GetVertexMatch(res2)
+			vertex1, exists1 := lookup[v1]
+			vertex2, exists2 := lookup[v2]
+			if !exists1 || !exists2 { // no match found, bug?
+				//if vertex1 == nil || vertex2 == nil { // no match found
+				return nil, fmt.Errorf("New vertices weren't found!") // programming error
+			}
+
+			edge, exists := oldGraph.Adjacency[vertex1][vertex2]
+			if !exists || edge.Name != e.Name { // TODO: edgeCmp
+				edge = e // use or overwrite edge
+			}
+			oldGraph.Adjacency[vertex1][vertex2] = edge // store it (AddEdge)
+			edgeKeep = append(edgeKeep, edge)           // mark as saved
+		}
+	}
+
+	// delete unused edges
+	for v1 := range oldGraph.Adjacency {
+		for _, e := range oldGraph.Adjacency[v1] {
+			// we have an edge!
+			if !EdgeContains(e, edgeKeep) {
+				oldGraph.DeleteEdge(e)
+			}
+		}
+	}
+
+	return oldGraph, nil
+}
+
+// GraphMetas returns a list of pointers to each of the resource MetaParams.
+func (g *Graph) GraphMetas() []*resources.MetaParams {
+	metas := []*resources.MetaParams{}
+	for v := range g.Adjacency { // loop through the vertices (resources))
+		res := v.Res // resource
+		meta := res.Meta()
+		metas = append(metas, meta)
+	}
+	return metas
+}
+
+// AssociateData associates some data with the object in the graph in question
+func (g *Graph) AssociateData(converger converger.Converger) {
 	for v := range g.GetVerticesChan() {
-		v.Res.SetConvergedCallback(ctimeout, converged)
+		v.Res.AssociateData(converger)
 	}
 }
 
-// in array function to test *Vertex in a slice of *Vertices
+// VertexContains is an "in array" function to test for a vertex in a slice of vertices.
 func VertexContains(needle *Vertex, haystack []*Vertex) bool {
 	for _, v := range haystack {
 		if needle == v {
@@ -871,7 +1057,17 @@ func VertexContains(needle *Vertex, haystack []*Vertex) bool {
 	return false
 }
 
-// reverse a list of vertices
+// EdgeContains is an "in array" function to test for an edge in a slice of edges.
+func EdgeContains(needle *Edge, haystack []*Edge) bool {
+	for _, v := range haystack {
+		if needle == v {
+			return true
+		}
+	}
+	return false
+}
+
+// Reverse reverses a list of vertices.
 func Reverse(vs []*Vertex) []*Vertex {
 	//var out []*Vertex       // XXX: golint suggests, but it fails testing
 	out := make([]*Vertex, 0) // empty list

pgraph/pgraph_test.go

@@ -15,9 +15,7 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-// NOTE: this is pgraph, a pointer graph
+package pgraph
-
-package main
 
 import (
 	"fmt"
@@ -25,8 +23,18 @@ import (
 	"sort"
 	"strings"
 	"testing"
+	"time"
 )
 
+// NV is a helper function to make testing easier. It creates a new noop vertex.
+func NV(s string) *Vertex {
+	obj, err := NewNoopRes(s)
+	if err != nil {
+		panic(err) // unlikely test failure!
+	}
+	return NewVertex(obj)
+}
+
 func TestPgraphT1(t *testing.T) {
 
 	G := NewGraph("g1")
@@ -39,8 +47,8 @@ func TestPgraphT1(t *testing.T) {
 		t.Errorf("Should have 0 edges instead of: %d.", i)
 	}
 
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
 	e1 := NewEdge("e1")
 	G.AddEdge(v1, v2, e1)
 
@@ -56,12 +64,12 @@ func TestPgraphT1(t *testing.T) {
 func TestPgraphT2(t *testing.T) {
 
 	G := NewGraph("g2")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -83,12 +91,12 @@ func TestPgraphT2(t *testing.T) {
 func TestPgraphT3(t *testing.T) {
 
 	G := NewGraph("g3")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -124,9 +132,9 @@ func TestPgraphT3(t *testing.T) {
 func TestPgraphT4(t *testing.T) {
 
 	G := NewGraph("g4")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -146,12 +154,12 @@ func TestPgraphT4(t *testing.T) {
 
 func TestPgraphT5(t *testing.T) {
 	G := NewGraph("g5")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -175,12 +183,12 @@ func TestPgraphT5(t *testing.T) {
 
 func TestPgraphT6(t *testing.T) {
 	G := NewGraph("g6")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -213,9 +221,9 @@ func TestPgraphT6(t *testing.T) {
 func TestPgraphT7(t *testing.T) {
 
 	G := NewGraph("g7")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -254,28 +262,28 @@ func TestPgraphT7(t *testing.T) {
 
 func TestPgraphT8(t *testing.T) {
 
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
 	if VertexContains(v1, []*Vertex{v1, v2, v3}) != true {
 		t.Errorf("Should be true instead of false.")
 	}
 
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	if VertexContains(v4, []*Vertex{v5, v6}) != false {
 		t.Errorf("Should be false instead of true.")
 	}
 
-	v7 := NewVertex(NewNoopRes("v7"))
+	v7 := NV("v7")
-	v8 := NewVertex(NewNoopRes("v8"))
+	v8 := NV("v8")
-	v9 := NewVertex(NewNoopRes("v9"))
+	v9 := NV("v9")
 	if VertexContains(v8, []*Vertex{v7, v8, v9}) != true {
 		t.Errorf("Should be true instead of false.")
 	}
 
-	v1b := NewVertex(NewNoopRes("v1")) // same value, different objects
+	v1b := NV("v1") // same value, different objects
 	if VertexContains(v1b, []*Vertex{v1, v2, v3}) != false {
 		t.Errorf("Should be false instead of true.")
 	}
@@ -284,12 +292,12 @@ func TestPgraphT8(t *testing.T) {
 func TestPgraphT9(t *testing.T) {
 
 	G := NewGraph("g9")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -344,11 +352,11 @@ func TestPgraphT9(t *testing.T) {
 		t.Errorf("Outdegree of v6 should be 0 instead of: %d.", i)
 	}
 
-	s, ok := G.TopologicalSort()
+	s, err := G.TopologicalSort()
 	// either possibility is a valid toposort
 	match := reflect.DeepEqual(s, []*Vertex{v1, v2, v3, v4, v5, v6}) || reflect.DeepEqual(s, []*Vertex{v1, v3, v2, v4, v5, v6})
-	if !ok || !match {
+	if err != nil || !match {
-		t.Errorf("Topological sort failed, status: %v.", ok)
+		t.Errorf("Topological sort failed, error: %v.", err)
 		str := "Found:"
 		for _, v := range s {
 			str += " " + v.Res.GetName()
@@ -360,12 +368,12 @@ func TestPgraphT9(t *testing.T) {
 func TestPgraphT10(t *testing.T) {
 
 	G := NewGraph("g10")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -379,8 +387,8 @@ func TestPgraphT10(t *testing.T) {
 	G.AddEdge(v5, v6, e5)
 	G.AddEdge(v4, v2, e6) // cycle
 
-	if _, ok := G.TopologicalSort(); ok {
+	if _, err := G.TopologicalSort(); err == nil {
-		t.Errorf("Topological sort passed, but graph is cyclic.")
+		t.Errorf("Topological sort passed, but graph is cyclic!")
 	}
 }
 
@@ -400,8 +408,8 @@ func TestPgraphReachability0(t *testing.T) {
 	}
 	{
 		G := NewGraph("g")
-		v1 := NewVertex(NewNoopRes("v1"))
+		v1 := NV("v1")
-		v6 := NewVertex(NewNoopRes("v6"))
+		v6 := NV("v6")
 
 		result := G.Reachability(v1, v6)
 		expected := []*Vertex{}
@@ -417,12 +425,12 @@ func TestPgraphReachability0(t *testing.T) {
 	}
 	{
 		G := NewGraph("g")
-		v1 := NewVertex(NewNoopRes("v1"))
+		v1 := NV("v1")
-		v2 := NewVertex(NewNoopRes("v2"))
+		v2 := NV("v2")
-		v3 := NewVertex(NewNoopRes("v3"))
+		v3 := NV("v3")
-		v4 := NewVertex(NewNoopRes("v4"))
+		v4 := NV("v4")
-		v5 := NewVertex(NewNoopRes("v5"))
+		v5 := NV("v5")
-		v6 := NewVertex(NewNoopRes("v6"))
+		v6 := NV("v6")
 		e1 := NewEdge("e1")
 		e2 := NewEdge("e2")
 		e3 := NewEdge("e3")
@@ -451,12 +459,12 @@ func TestPgraphReachability0(t *testing.T) {
 // simple linear path
 func TestPgraphReachability1(t *testing.T) {
 	G := NewGraph("g")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -485,12 +493,12 @@ func TestPgraphReachability1(t *testing.T) {
 // pick one of two correct paths
 func TestPgraphReachability2(t *testing.T) {
 	G := NewGraph("g")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -522,12 +530,12 @@ func TestPgraphReachability2(t *testing.T) {
 // pick shortest path
 func TestPgraphReachability3(t *testing.T) {
 	G := NewGraph("g")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -557,12 +565,12 @@ func TestPgraphReachability3(t *testing.T) {
 // direct path
 func TestPgraphReachability4(t *testing.T) {
 	G := NewGraph("g")
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 	e1 := NewEdge("e1")
 	e2 := NewEdge("e2")
 	e3 := NewEdge("e3")
@@ -590,12 +598,12 @@ func TestPgraphReachability4(t *testing.T) {
 }
 
 func TestPgraphT11(t *testing.T) {
-	v1 := NewVertex(NewNoopRes("v1"))
+	v1 := NV("v1")
-	v2 := NewVertex(NewNoopRes("v2"))
+	v2 := NV("v2")
-	v3 := NewVertex(NewNoopRes("v3"))
+	v3 := NV("v3")
-	v4 := NewVertex(NewNoopRes("v4"))
+	v4 := NV("v4")
-	v5 := NewVertex(NewNoopRes("v5"))
+	v5 := NV("v5")
-	v6 := NewVertex(NewNoopRes("v6"))
+	v6 := NV("v6")
 
 	if rev := Reverse([]*Vertex{}); !reflect.DeepEqual(rev, []*Vertex{}) {
 		t.Errorf("Reverse of vertex slice failed.")
@@ -638,7 +646,7 @@ func NewNoopResTest(name string) *NoopResTest {
 		NoopRes: NoopRes{
 			BaseRes: BaseRes{
 				Name: name,
-				Meta: MetaParams{
+				MetaParams: MetaParams{
 					AutoGroup: true, // always autogroup
 				},
 			},
@@ -807,7 +815,7 @@ func (g *Graph) fullPrint() (str string) {
 // helper function
 func runGraphCmp(t *testing.T, g1, g2 *Graph) {
 	ch := g1.autoGroup(&testGrouper{}) // edits the graph
-	for _ = range ch {                 // bleed the channel or it won't run :(
+	for range ch {                     // bleed the channel or it won't run :(
 		// pass
 	}
 	err := GraphCmp(g1, g2)
@@ -819,7 +827,7 @@ func runGraphCmp(t *testing.T, g1, g2 *Graph) {
 	}
 }
 
-// all of the following test cases are layed out with the following semantics:
+// all of the following test cases are laid out with the following semantics:
 // * vertices which start with the same single letter are considered "like"
 // * "like" elements should be merged
 // * vertices can have any integer after their single letter "family" type
@@ -1282,3 +1290,13 @@ func TestPgraphGroupingConnected1(t *testing.T) {
 	}
 	runGraphCmp(t, g1, g2)
 }
+
+func TestDurationAssumptions(t *testing.T) {
+	var d time.Duration
+	if (d == 0) != true {
+		t.Errorf("Empty time.Duration is no longer equal to zero!")
+	}
+	if (d > 0) != false {
+		t.Errorf("Empty time.Duration is now greater than zero!")
+	}
+}

puppet/gapi.go

@@ -0,0 +1,121 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package puppet
+
+import (
+	"fmt"
+	"log"
+	"sync"
+	"time"
+
+	"github.com/purpleidea/mgmt/gapi"
+	"github.com/purpleidea/mgmt/pgraph"
+)
+
+// GAPI implements the main puppet GAPI interface.
+type GAPI struct {
+	PuppetParam *string // puppet mode to run; nil if undefined
+	PuppetConf  string  // the path to an alternate puppet.conf file
+
+	data        gapi.Data
+	initialized bool
+	closeChan   chan struct{}
+	wg          sync.WaitGroup // sync group for tunnel go routines
+}
+
+// NewGAPI creates a new puppet GAPI struct and calls Init().
+func NewGAPI(data gapi.Data, puppetParam *string, puppetConf string) (*GAPI, error) {
+	obj := &GAPI{
+		PuppetParam: puppetParam,
+		PuppetConf:  puppetConf,
+	}
+	return obj, obj.Init(data)
+}
+
+// Init initializes the puppet GAPI struct.
+func (obj *GAPI) Init(data gapi.Data) error {
+	if obj.initialized {
+		return fmt.Errorf("Already initialized!")
+	}
+	if obj.PuppetParam == nil {
+		return fmt.Errorf("The PuppetParam param must be specified!")
+	}
+	obj.data = data // store for later
+	obj.closeChan = make(chan struct{})
+	obj.initialized = true
+	return nil
+}
+
+// Graph returns a current Graph.
+func (obj *GAPI) Graph() (*pgraph.Graph, error) {
+	if !obj.initialized {
+		return nil, fmt.Errorf("Puppet: GAPI is not initialized!")
+	}
+	config := ParseConfigFromPuppet(*obj.PuppetParam, obj.PuppetConf)
+	if config == nil {
+		return nil, fmt.Errorf("Puppet: ParseConfigFromPuppet returned nil!")
+	}
+	g, err := config.NewGraphFromConfig(obj.data.Hostname, obj.data.EmbdEtcd, obj.data.Noop)
+	return g, err
+}
+
+// SwitchStream returns nil errors every time there could be a new graph.
+func (obj *GAPI) SwitchStream() chan error {
+	if obj.data.NoWatch {
+		return nil
+	}
+	puppetChan := func() <-chan time.Time { // helper function
+		return time.Tick(time.Duration(PuppetInterval(obj.PuppetConf)) * time.Second)
+	}
+	ch := make(chan error)
+	obj.wg.Add(1)
+	go func() {
+		defer obj.wg.Done()
+		defer close(ch) // this will run before the obj.wg.Done()
+		if !obj.initialized {
+			ch <- fmt.Errorf("Puppet: GAPI is not initialized!")
+			return
+		}
+		pChan := puppetChan()
+		for {
+			select {
+			case _, ok := <-pChan:
+				if !ok { // the channel closed!
+					return
+				}
+				log.Printf("Puppet: Generating new graph...")
+				pChan = puppetChan() // TODO: okay to update interval in case it changed?
+				ch <- nil            // trigger a run
+			case <-obj.closeChan:
+				return
+			}
+		}
+	}()
+	return ch
+}
+
+// Close shuts down the Puppet GAPI.
+func (obj *GAPI) Close() error {
+	if !obj.initialized {
+		return fmt.Errorf("Puppet: GAPI is not initialized!")
+	}
+	close(obj.closeChan)
+	obj.wg.Wait()
+	obj.initialized = false // closed = true
+	return nil
+}

puppet/puppet.go

@@ -0,0 +1,145 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package puppet provides the integration entrypoint for the puppet language.
+package puppet
+
+import (
+	"bufio"
+	"io"
+	"log"
+	"os/exec"
+	"strconv"
+	"strings"
+
+	"github.com/purpleidea/mgmt/global"
+	"github.com/purpleidea/mgmt/yamlgraph"
+)
+
+const (
+	// PuppetYAMLBufferSize is the maximum buffer size for the yaml input data
+	PuppetYAMLBufferSize = 65535
+)
+
+func runPuppetCommand(cmd *exec.Cmd) ([]byte, error) {
+	if global.DEBUG {
+		log.Printf("Puppet: running command: %v", cmd)
+	}
+
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		log.Printf("Puppet: Error opening pipe to puppet command: %v", err)
+		return nil, err
+	}
+	stderr, err := cmd.StderrPipe()
+	if err != nil {
+		log.Printf("Puppet: Error opening error pipe to puppet command: %v", err)
+		return nil, err
+	}
+
+	if err := cmd.Start(); err != nil {
+		log.Printf("Puppet: Error starting puppet command: %v", err)
+		return nil, err
+	}
+
+	// XXX: the current implementation is likely prone to fail
+	// as soon as the YAML data overflows the buffer.
+	data := make([]byte, PuppetYAMLBufferSize)
+	var result []byte
+	for err == nil {
+		var count int
+		count, err = stdout.Read(data)
+		if err != nil && err != io.EOF {
+			log.Printf("Puppet: Error reading YAML data from puppet: %v", err)
+			return nil, err
+		}
+		// Slicing down to the number of actual bytes is important, the YAML parser
+		// will choke on an oversized slice. http://stackoverflow.com/a/33726617/3356612
+		result = append(result, data[0:count]...)
+	}
+	if global.DEBUG {
+		log.Printf("Puppet: read %v bytes of data from puppet", len(result))
+	}
+	for scanner := bufio.NewScanner(stderr); scanner.Scan(); {
+		log.Printf("Puppet: (output) %v", scanner.Text())
+	}
+	if err := cmd.Wait(); err != nil {
+		log.Printf("Puppet: Error: puppet command did not complete: %v", err)
+		return nil, err
+	}
+
+	return result, nil
+}
+
+// ParseConfigFromPuppet takes a special puppet param string and config and
+// returns the graph configuration structure.
+func ParseConfigFromPuppet(puppetParam, puppetConf string) *yamlgraph.GraphConfig {
+	var puppetConfArg string
+	if puppetConf != "" {
+		puppetConfArg = "--config=" + puppetConf
+	}
+
+	var cmd *exec.Cmd
+	if puppetParam == "agent" {
+		cmd = exec.Command("puppet", "mgmtgraph", "print", puppetConfArg)
+	} else if strings.HasSuffix(puppetParam, ".pp") {
+		cmd = exec.Command("puppet", "mgmtgraph", "print", puppetConfArg, "--manifest", puppetParam)
+	} else {
+		cmd = exec.Command("puppet", "mgmtgraph", "print", puppetConfArg, "--code", puppetParam)
+	}
+
+	log.Println("Puppet: launching translator")
+
+	var config yamlgraph.GraphConfig
+	if data, err := runPuppetCommand(cmd); err != nil {
+		return nil
+	} else if err := config.Parse(data); err != nil {
+		log.Printf("Puppet: Error: Could not parse YAML output with Parse: %v", err)
+		return nil
+	}
+
+	return &config
+}
+
+// PuppetInterval returns the graph refresh interval from the puppet configuration.
+func PuppetInterval(puppetConf string) int {
+	if global.DEBUG {
+		log.Printf("Puppet: determining graph refresh interval")
+	}
+	var cmd *exec.Cmd
+	if puppetConf != "" {
+		cmd = exec.Command("puppet", "config", "print", "runinterval", "--config", puppetConf)
+	} else {
+		cmd = exec.Command("puppet", "config", "print", "runinterval")
+	}
+
+	log.Println("Puppet: inspecting runinterval configuration")
+
+	interval := 1800
+	data, err := runPuppetCommand(cmd)
+	if err != nil {
+		log.Printf("Puppet: could not determine configured run interval (%v), using default of %v", err, interval)
+		return interval
+	}
+	result, err := strconv.ParseInt(strings.TrimSpace(string(data)), 10, 0)
+	if err != nil {
+		log.Printf("Puppet: error reading numeric runinterval value (%v), using default of %v", err, interval)
+		return interval
+	}
+
+	return int(result)
+}

recwatch/configwatch.go

@@ -0,0 +1,134 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package recwatch
+
+import (
+	"log"
+	"sync"
+
+	"github.com/purpleidea/mgmt/global"
+)
+
+// ConfigWatcher returns events on a channel anytime one of its files events.
+type ConfigWatcher struct {
+	ch        chan string
+	wg        sync.WaitGroup
+	closechan chan struct{}
+	errorchan chan error
+}
+
+// NewConfigWatcher creates a new ConfigWatcher struct.
+func NewConfigWatcher() *ConfigWatcher {
+	return &ConfigWatcher{
+		ch:        make(chan string),
+		closechan: make(chan struct{}),
+		errorchan: make(chan error),
+	}
+}
+
+// Add new file paths to watch for events on.
+func (obj *ConfigWatcher) Add(file ...string) {
+	if len(file) == 0 {
+		return
+	}
+	if len(file) > 1 {
+		for _, f := range file { // add all the files...
+			obj.Add(f) // recurse
+		}
+		return
+	}
+	// otherwise, add the one file passed in...
+	obj.wg.Add(1)
+	go func() {
+		defer obj.wg.Done()
+		ch := ConfigWatch(file[0])
+		for {
+			select {
+			case e := <-ch:
+				if e != nil {
+					obj.errorchan <- e
+					return
+				}
+				obj.ch <- file[0]
+				continue
+			case <-obj.closechan:
+				return
+			}
+		}
+	}()
+}
+
+// Error returns a channel of errors that notifies us of permanent issues.
+func (obj *ConfigWatcher) Error() <-chan error {
+	return obj.errorchan
+}
+
+// Events returns a channel to listen on for file events. It closes when it is
+// emptied after the Close() method is called. You can test for closure with the
+// f, more := <-obj.Events() pattern.
+func (obj *ConfigWatcher) Events() chan string {
+	return obj.ch
+}
+
+// Close shuts down the ConfigWatcher object. It closes the Events channel after
+// all the currently pending events have been emptied.
+func (obj *ConfigWatcher) Close() {
+	if obj.ch == nil {
+		return
+	}
+	close(obj.closechan)
+	obj.wg.Wait() // wait until everyone is done sending on obj.ch
+	//obj.ch <- "" // send finished message
+	close(obj.ch)
+	obj.ch = nil
+	close(obj.errorchan)
+}
+
+// ConfigWatch writes on the channel every time an event is seen for the path.
+func ConfigWatch(file string) chan error {
+	ch := make(chan error)
+	go func() {
+		recWatcher, err := NewRecWatcher(file, false)
+		if err != nil {
+			ch <- err
+			close(ch)
+			return
+		}
+		defer recWatcher.Close()
+		for {
+			if global.DEBUG {
+				log.Printf("Watching: %v", file)
+			}
+			select {
+			case event, ok := <-recWatcher.Events():
+				if !ok { // channel is closed
+					close(ch)
+					return
+				}
+				if err := event.Error; err != nil {
+					ch <- err
+					close(ch)
+					return
+				}
+				ch <- nil // send event!
+			}
+		}
+		//close(ch)
+	}()
+	return ch
+}

recwatch/recwatch.go

@@ -0,0 +1,319 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package recwatch provides recursive file watching events via fsnotify.
+package recwatch
+
+import (
+	"fmt"
+	"log"
+	"math"
+	"os"
+	"path"
+	"path/filepath"
+	"strings"
+	"sync"
+	"syscall"
+
+	"github.com/purpleidea/mgmt/global" // XXX: package mgmtmain instead?
+	"github.com/purpleidea/mgmt/util"
+
+	"gopkg.in/fsnotify.v1"
+	//"github.com/go-fsnotify/fsnotify" // git master of "gopkg.in/fsnotify.v1"
+)
+
+// Event represents a watcher event. These can include errors.
+type Event struct {
+	Error error
+	Body  *fsnotify.Event
+}
+
+// RecWatcher is the struct for the recursive watcher. Run Init() on it.
+type RecWatcher struct {
+	Path     string // computed path
+	Recurse  bool   // should we watch recursively?
+	isDir    bool   // computed isDir
+	safename string // safe path
+	watcher  *fsnotify.Watcher
+	watches  map[string]struct{}
+	events   chan Event // one channel for events and err...
+	once     sync.Once
+	wg       sync.WaitGroup
+	exit     chan struct{}
+	closeErr error
+}
+
+// NewRecWatcher creates an initializes a new recursive watcher.
+func NewRecWatcher(path string, recurse bool) (*RecWatcher, error) {
+	obj := &RecWatcher{
+		Path:    path,
+		Recurse: recurse,
+	}
+	return obj, obj.Init()
+}
+
+// Init starts the recursive file watcher.
+func (obj *RecWatcher) Init() error {
+	obj.watcher = nil
+	obj.watches = make(map[string]struct{})
+	obj.events = make(chan Event)
+	obj.exit = make(chan struct{})
+	obj.isDir = strings.HasSuffix(obj.Path, "/") // dirs have trailing slashes
+	obj.safename = path.Clean(obj.Path)          // no trailing slash
+
+	var err error
+	obj.watcher, err = fsnotify.NewWatcher()
+	if err != nil {
+		return err
+	}
+
+	if obj.isDir {
+		if err := obj.addSubFolders(obj.safename); err != nil {
+			return err
+		}
+	}
+
+	go func() {
+		if err := obj.Watch(); err != nil {
+			obj.events <- Event{Error: err}
+		}
+		obj.Close()
+	}()
+	return nil
+}
+
+//func (obj *RecWatcher) Add(path string) error { // XXX: implement me or not?
+//
+//}
+//
+//func (obj *RecWatcher) Remove(path string) error { // XXX: implement me or not?
+//
+//}
+
+// Close shuts down the watcher.
+func (obj *RecWatcher) Close() error {
+	obj.once.Do(obj.close) // don't cause the channel to close twice
+	return obj.closeErr
+}
+
+// This close function is the function that actually does the close work. Don't
+// call it more than once!
+func (obj *RecWatcher) close() {
+	var err error
+	close(obj.exit) // send exit signal
+	obj.wg.Wait()
+	if obj.watcher != nil {
+		err = obj.watcher.Close()
+		obj.watcher = nil
+		// TODO: should we send the close error?
+		//if err != nil {
+		//	obj.events <- Event{Error: err}
+		//}
+	}
+	close(obj.events)
+	obj.closeErr = err // set the error
+}
+
+// Events returns a channel of events. These include events for errors.
+func (obj *RecWatcher) Events() chan Event { return obj.events }
+
+// Watch is the primary listener for this resource and it outputs events.
+func (obj *RecWatcher) Watch() error {
+	if obj.watcher == nil {
+		return fmt.Errorf("Watcher is not initialized!")
+	}
+	obj.wg.Add(1)
+	defer obj.wg.Done()
+
+	patharray := util.PathSplit(obj.safename) // tokenize the path
+	var index = len(patharray)                // starting index
+	var current string                        // current "watcher" location
+	var deltaDepth int                        // depth delta between watcher and event
+	var send = false                          // send event?
+
+	for {
+		current = strings.Join(patharray[0:index], "/")
+		if current == "" { // the empty string top is the root dir ("/")
+			current = "/"
+		}
+		if global.DEBUG {
+			log.Printf("Watching: %s", current) // attempting to watch...
+		}
+		// initialize in the loop so that we can reset on rm-ed handles
+		if err := obj.watcher.Add(current); err != nil {
+			if global.DEBUG {
+				log.Printf("watcher.Add(%s): Error: %v", current, err)
+			}
+
+			if err == syscall.ENOENT {
+				index-- // usually not found, move up one dir
+				index = int(math.Max(1, float64(index)))
+				continue
+			}
+
+			if err == syscall.ENOSPC {
+				// no space left on device, out of inotify watches
+				// TODO: consider letting the user fall back to
+				// polling if they hit this error very often...
+				return fmt.Errorf("Out of inotify watches: %v", err)
+			} else if os.IsPermission(err) {
+				return fmt.Errorf("Permission denied adding a watch: %v", err)
+			}
+			return fmt.Errorf("Unknown error: %v", err)
+		}
+
+		select {
+		case event := <-obj.watcher.Events:
+			if global.DEBUG {
+				log.Printf("Watch(%s), Event(%s): %v", current, event.Name, event.Op)
+			}
+			// the deeper you go, the bigger the deltaDepth is...
+			// this is the difference between what we're watching,
+			// and the event... doesn't mean we can't watch deeper
+			if current == event.Name {
+				deltaDepth = 0 // i was watching what i was looking for
+
+			} else if util.HasPathPrefix(event.Name, current) {
+				deltaDepth = len(util.PathSplit(current)) - len(util.PathSplit(event.Name)) // -1 or less
+
+			} else if util.HasPathPrefix(current, event.Name) {
+				deltaDepth = len(util.PathSplit(event.Name)) - len(util.PathSplit(current)) // +1 or more
+				// if below me...
+				if _, exists := obj.watches[event.Name]; exists {
+					send = true
+					if event.Op&fsnotify.Remove == fsnotify.Remove {
+						obj.watcher.Remove(event.Name)
+						delete(obj.watches, event.Name)
+					}
+					if (event.Op&fsnotify.Create == fsnotify.Create) && isDir(event.Name) {
+						obj.watcher.Add(event.Name)
+						obj.watches[event.Name] = struct{}{}
+						if err := obj.addSubFolders(event.Name); err != nil {
+							return err
+						}
+					}
+				}
+
+			} else {
+				// TODO: different watchers get each others events!
+				// https://github.com/go-fsnotify/fsnotify/issues/95
+				// this happened with two values such as:
+				// event.Name: /tmp/mgmt/f3 and current: /tmp/mgmt/f2
+				continue
+			}
+			//log.Printf("The delta depth is: %v", deltaDepth)
+
+			// if we have what we wanted, awesome, send an event...
+			if event.Name == obj.safename {
+				//log.Println("Event!")
+				// FIXME: should all these below cases trigger?
+				send = true
+
+				if obj.isDir {
+					if err := obj.addSubFolders(obj.safename); err != nil {
+						return err
+					}
+				}
+
+				// file removed, move the watch upwards
+				if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
+					//log.Println("Removal!")
+					obj.watcher.Remove(current)
+					index--
+				}
+
+				// we must be a parent watcher, so descend in
+				if deltaDepth < 0 {
+					// XXX: we can block here due to: https://github.com/fsnotify/fsnotify/issues/123
+					obj.watcher.Remove(current)
+					index++
+				}
+
+				// if safename starts with event.Name, we're above, and no event should be sent
+			} else if util.HasPathPrefix(obj.safename, event.Name) {
+				//log.Println("Above!")
+
+				if deltaDepth >= 0 && (event.Op&fsnotify.Remove == fsnotify.Remove) {
+					log.Println("Removal!")
+					obj.watcher.Remove(current)
+					index--
+				}
+
+				if deltaDepth < 0 {
+					log.Println("Parent!")
+					if util.PathPrefixDelta(obj.safename, event.Name) == 1 { // we're the parent dir
+						send = true
+					}
+					obj.watcher.Remove(current)
+					index++
+				}
+
+				// if event.Name startswith safename, send event, we're already deeper
+			} else if util.HasPathPrefix(event.Name, obj.safename) {
+				//log.Println("Event2!")
+				send = true
+			}
+
+			// do all our event sending all together to avoid duplicate msgs
+			if send {
+				send = false
+				// only invalid state on certain types of events
+				obj.events <- Event{Error: nil, Body: &event}
+			}
+
+		case err := <-obj.watcher.Errors:
+			return fmt.Errorf("Unknown watcher error: %v", err)
+
+		case <-obj.exit:
+			return nil
+		}
+	}
+}
+
+// addSubFolders is a helper that is used to add recursive dirs to the watches.
+func (obj *RecWatcher) addSubFolders(p string) error {
+	if !obj.Recurse {
+		return nil // if we're not watching recursively, just exit early
+	}
+	// look at all subfolders...
+	walkFn := func(path string, info os.FileInfo, err error) error {
+		if global.DEBUG {
+			log.Printf("Walk: %s (%v): %v", path, info, err)
+		}
+		if err != nil {
+			return nil
+		}
+		if info.IsDir() {
+			obj.watches[path] = struct{}{} // add key
+			err := obj.watcher.Add(path)
+			if err != nil {
+				return err // TODO: will this bubble up?
+			}
+		}
+		return nil
+	}
+	err := filepath.Walk(p, walkFn)
+	return err
+}
+
+func isDir(path string) bool {
+	finfo, err := os.Stat(path)
+	if err != nil {
+		return false
+	}
+	return finfo.IsDir()
+}

resources.go

@@ -1,353 +0,0 @@
-// Mgmt
-// Copyright (C) 2013-2016+ James Shubin and the project contributors
-// Written by James Shubin <james@shubin.ca> and the project contributors
-//
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as published by
-// the Free Software Foundation, either version 3 of the License, or
-// (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-package main
-
-import (
-	"bytes"
-	"encoding/base64"
-	"encoding/gob"
-	"fmt"
-	"log"
-)
-
-//go:generate stringer -type=resState -output=resstate_stringer.go
-type resState int
-
-const (
-	resStateNil resState = iota
-	resStateWatching
-	resStateEvent // an event has happened, but we haven't poked yet
-	resStateCheckApply
-	resStatePoking
-)
-
-//go:generate stringer -type=resConvergedState -output=resconvergedstate_stringer.go
-type resConvergedState int
-
-const (
-	resConvergedNil resConvergedState = iota
-	//resConverged
-	resConvergedTimeout
-)
-
-// a unique identifier for a resource, namely it's name, and the kind ("type")
-type ResUUID interface {
-	GetName() string
-	Kind() string
-	IFF(ResUUID) bool
-
-	Reversed() bool // true means this resource happens before the generator
-}
-
-type BaseUUID struct {
-	name string // name and kind are the values of where this is coming from
-	kind string
-
-	reversed *bool // piggyback edge information here
-}
-
-type AutoEdge interface {
-	Next() []ResUUID  // call to get list of edges to add
-	Test([]bool) bool // call until false
-}
-
-type MetaParams struct {
-	AutoEdge  bool `yaml:"autoedge"`  // metaparam, should we generate auto edges? // XXX should default to true
-	AutoGroup bool `yaml:"autogroup"` // metaparam, should we auto group? // XXX should default to true
-}
-
-// this interface is everything that is common to all resources
-// everything here only needs to be implemented once, in the BaseRes
-type Base interface {
-	GetName() string // can't be named "Name()" because of struct field
-	SetName(string)
-	Kind() string
-	GetMeta() MetaParams
-	SetVertex(*Vertex)
-	SetConvergedCallback(ctimeout int, converged chan bool)
-	IsWatching() bool
-	SetWatching(bool)
-	GetConvergedState() resConvergedState
-	SetConvergedState(resConvergedState)
-	GetState() resState
-	SetState(resState)
-	SendEvent(eventName, bool, bool) bool
-	ReadEvent(*Event) (bool, bool) // TODO: optional here?
-	GroupCmp(Res) bool             // TODO: is there a better name for this?
-	GroupRes(Res) error            // group resource (arg) into self
-	IsGrouped() bool               // am I grouped?
-	SetGrouped(bool)               // set grouped bool
-	GetGroup() []Res               // return everyone grouped inside me
-	SetGroup([]Res)
-}
-
-// this is the minimum interface you need to implement to make a new resource
-type Res interface {
-	Base // include everything from the Base interface
-	Init()
-	//Validate() bool    // TODO: this might one day be added
-	GetUUIDs() []ResUUID // most resources only return one
-	Watch(chan Event)    // send on channel to signal process() events
-	CheckApply(bool) (bool, error)
-	AutoEdges() AutoEdge
-	Compare(Res) bool
-	CollectPattern(string) // XXX: temporary until Res collection is more advanced
-}
-
-type BaseRes struct {
-	Name           string     `yaml:"name"`
-	Meta           MetaParams `yaml:"meta"` // struct of all the metaparams
-	kind           string
-	events         chan Event
-	vertex         *Vertex
-	state          resState
-	convergedState resConvergedState
-	watching       bool // is Watch() loop running ?
-	ctimeout       int  // converged timeout
-	converged      chan bool
-	isStateOK      bool  // whether the state is okay based on events or not
-	isGrouped      bool  // am i contained within a group?
-	grouped        []Res // list of any grouped resources
-}
-
-// wraps the IFF method when used with a list of UUID's
-func UUIDExistsInUUIDs(uuid ResUUID, uuids []ResUUID) bool {
-	for _, u := range uuids {
-		if uuid.IFF(u) {
-			return true
-		}
-	}
-	return false
-}
-
-func (obj *BaseUUID) GetName() string {
-	return obj.name
-}
-
-func (obj *BaseUUID) Kind() string {
-	return obj.kind
-}
-
-// if and only if they are equivalent, return true
-// if they are not equivalent, return false
-// most resource will want to override this method, since it does the important
-// work of actually discerning if two resources are identical in function
-func (obj *BaseUUID) IFF(uuid ResUUID) bool {
-	res, ok := uuid.(*BaseUUID)
-	if !ok {
-		return false
-	}
-	return obj.name == res.name
-}
-
-func (obj *BaseUUID) Reversed() bool {
-	if obj.reversed == nil {
-		log.Fatal("Programming error!")
-	}
-	return *obj.reversed
-}
-
-// initialize structures like channels if created without New constructor
-func (obj *BaseRes) Init() {
-	obj.events = make(chan Event) // unbuffered chan size to avoid stale events
-}
-
-// this method gets used by all the resources
-func (obj *BaseRes) GetName() string {
-	return obj.Name
-}
-
-func (obj *BaseRes) SetName(name string) {
-	obj.Name = name
-}
-
-// return the kind of resource this is
-func (obj *BaseRes) Kind() string {
-	return obj.kind
-}
-
-func (obj *BaseRes) GetMeta() MetaParams {
-	return obj.Meta
-}
-
-func (obj *BaseRes) GetVertex() *Vertex {
-	return obj.vertex
-}
-
-func (obj *BaseRes) SetVertex(v *Vertex) {
-	obj.vertex = v
-}
-
-func (obj *BaseRes) SetConvergedCallback(ctimeout int, converged chan bool) {
-	obj.ctimeout = ctimeout
-	obj.converged = converged
-}
-
-// is the Watch() function running?
-func (obj *BaseRes) IsWatching() bool {
-	return obj.watching
-}
-
-// store status of if the Watch() function is running
-func (obj *BaseRes) SetWatching(b bool) {
-	obj.watching = b
-}
-
-func (obj *BaseRes) GetConvergedState() resConvergedState {
-	return obj.convergedState
-}
-
-func (obj *BaseRes) SetConvergedState(state resConvergedState) {
-	obj.convergedState = state
-}
-
-func (obj *BaseRes) GetState() resState {
-	return obj.state
-}
-
-func (obj *BaseRes) SetState(state resState) {
-	if DEBUG {
-		log.Printf("%v[%v]: State: %v -> %v", obj.Kind(), obj.GetName(), obj.GetState(), state)
-	}
-	obj.state = state
-}
-
-// push an event into the message queue for a particular vertex
-func (obj *BaseRes) SendEvent(event eventName, sync bool, activity bool) bool {
-	// TODO: isn't this race-y ?
-	if !obj.IsWatching() { // element has already exited
-		return false // if we don't return, we'll block on the send
-	}
-	if !sync {
-		obj.events <- Event{event, nil, "", activity}
-		return true
-	}
-
-	resp := make(chan bool)
-	obj.events <- Event{event, resp, "", activity}
-	for {
-		value := <-resp
-		// wait until true value
-		if value {
-			return true
-		}
-	}
-}
-
-// process events when a select gets one, this handles the pause code too!
-// the return values specify if we should exit and poke respectively
-func (obj *BaseRes) ReadEvent(event *Event) (exit, poke bool) {
-	event.ACK()
-	switch event.Name {
-	case eventStart:
-		return false, true
-
-	case eventPoke:
-		return false, true
-
-	case eventBackPoke:
-		return false, true // forward poking in response to a back poke!
-
-	case eventExit:
-		return true, false
-
-	case eventPause:
-		// wait for next event to continue
-		select {
-		case e := <-obj.events:
-			e.ACK()
-			if e.Name == eventExit {
-				return true, false
-			} else if e.Name == eventStart { // eventContinue
-				return false, false // don't poke on unpause!
-			} else {
-				// if we get a poke event here, it's a bug!
-				log.Fatalf("%v[%v]: Unknown event: %v, while paused!", obj.Kind(), obj.GetName(), e)
-			}
-		}
-
-	default:
-		log.Fatal("Unknown event: ", event)
-	}
-	return true, false // required to keep the stupid go compiler happy
-}
-
-func (obj *BaseRes) GroupRes(res Res) error {
-	if l := len(res.GetGroup()); l > 0 {
-		return fmt.Errorf("Res: %v already contains %d grouped resources!", res, l)
-	}
-	if res.IsGrouped() {
-		return fmt.Errorf("Res: %v is already grouped!", res)
-	}
-
-	obj.grouped = append(obj.grouped, res)
-	res.SetGrouped(true) // i am contained _in_ a group
-	return nil
-}
-
-func (obj *BaseRes) IsGrouped() bool { // am I grouped?
-	return obj.isGrouped
-}
-
-func (obj *BaseRes) SetGrouped(b bool) {
-	obj.isGrouped = b
-}
-
-func (obj *BaseRes) GetGroup() []Res { // return everyone grouped inside me
-	return obj.grouped
-}
-
-func (obj *BaseRes) SetGroup(g []Res) {
-	obj.grouped = g
-}
-
-func (obj *BaseRes) CollectPattern(pattern string) {
-	// XXX: default method is empty
-}
-
-// ResToB64 encodes a resource to a base64 encoded string (after serialization)
-func ResToB64(res Res) (string, error) {
-	b := bytes.Buffer{}
-	e := gob.NewEncoder(&b)
-	err := e.Encode(&res) // pass with &
-	if err != nil {
-		return "", fmt.Errorf("Gob failed to encode: %v", err)
-	}
-	return base64.StdEncoding.EncodeToString(b.Bytes()), nil
-}
-
-// B64ToRes decodes a resource from a base64 encoded string (after deserialization)
-func B64ToRes(str string) (Res, error) {
-	var output interface{}
-	bb, err := base64.StdEncoding.DecodeString(str)
-	if err != nil {
-		return nil, fmt.Errorf("Base64 failed to decode: %v", err)
-	}
-	b := bytes.NewBuffer(bb)
-	d := gob.NewDecoder(b)
-	err = d.Decode(&output) // pass with &
-	if err != nil {
-		return nil, fmt.Errorf("Gob failed to decode: %v", err)
-	}
-	res, ok := output.(Res)
-	if !ok {
-		return nil, fmt.Errorf("Output %v is not a Res", res)
-
-	}
-	return res, nil
-}

resources/exec.go

@@ -15,22 +15,28 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-package main
+package resources
 
 import (
 	"bufio"
 	"bytes"
 	"encoding/gob"
 	"errors"
+	"fmt"
 	"log"
 	"os/exec"
 	"strings"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/util"
 )
 
 func init() {
 	gob.Register(&ExecRes{})
 }
 
+// ExecRes is an exec resource for running commands.
 type ExecRes struct {
 	BaseRes    `yaml:",inline"`
 	State      string `yaml:"state"`      // state: exists/present?, absent, (undefined?)
@@ -44,7 +50,8 @@ type ExecRes struct {
 	PollInt    int    `yaml:"pollint"`    // the poll interval for the ifcmd
 }
 
-func NewExecRes(name, cmd, shell string, timeout int, watchcmd, watchshell, ifcmd, ifshell string, pollint int, state string) *ExecRes {
+// NewExecRes is a constructor for this resource. It also calls Init() for you.
+func NewExecRes(name, cmd, shell string, timeout int, watchcmd, watchshell, ifcmd, ifshell string, pollint int, state string) (*ExecRes, error) {
 	obj := &ExecRes{
 		BaseRes: BaseRes{
 			Name: name,
@@ -59,31 +66,31 @@ func NewExecRes(name, cmd, shell string, timeout int, watchcmd, watchshell, ifcm
 		PollInt:    pollint,
 		State:      state,
 	}
-	obj.Init()
+	return obj, obj.Init()
-	return obj
 }
 
-func (obj *ExecRes) Init() {
+// Init runs some startup code for this resource.
+func (obj *ExecRes) Init() error {
 	obj.BaseRes.kind = "Exec"
-	obj.BaseRes.Init() // call base init, b/c we're overriding
+	return obj.BaseRes.Init() // call base init, b/c we're overriding
 }
 
-// validate if the params passed in are valid data
+// Validate if the params passed in are valid data.
 // FIXME: where should this get called ?
-func (obj *ExecRes) Validate() bool {
+func (obj *ExecRes) Validate() error {
 	if obj.Cmd == "" { // this is the only thing that is really required
-		return false
+		return fmt.Errorf("Command can't be empty!")
 	}
 
 	// if we have a watch command, then we don't poll with the if command!
 	if obj.WatchCmd != "" && obj.PollInt > 0 {
-		return false
+		return fmt.Errorf("Don't poll when we have a watch command.")
 	}
 
-	return true
+	return nil
 }
 
-// wraps the scanner output in a channel
+// BufioChanScanner wraps the scanner output in a channel.
 func (obj *ExecRes) BufioChanScanner(scanner *bufio.Scanner) (chan string, chan error) {
 	ch, errch := make(chan string), make(chan error)
 	go func() {
@@ -91,7 +98,7 @@ func (obj *ExecRes) BufioChanScanner(scanner *bufio.Scanner) (chan string, chan
 			ch <- scanner.Text() // blocks here ?
 			if e := scanner.Err(); e != nil {
 				errch <- e // send any misc errors we encounter
-				//break // TODO ?
+				//break // TODO: ?
 			}
 		}
 		close(ch)
@@ -101,18 +108,28 @@ func (obj *ExecRes) BufioChanScanner(scanner *bufio.Scanner) (chan string, chan
 	return ch, errch
 }
 
-// Exec watcher
+// Watch is the primary listener for this resource and it outputs events.
-func (obj *ExecRes) Watch(processChan chan Event) {
+func (obj *ExecRes) Watch(processChan chan event.Event) error {
 	if obj.IsWatching() {
-		return
+		return nil
 	}
 	obj.SetWatching(true)
 	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
 
 	var send = false // send event?
 	var exit = false
 	bufioch, errch := make(chan string), make(chan error)
-	//vertex := obj.GetVertex()         // stored with SetVertex
 
 	if obj.WatchCmd != "" {
 		var cmdName string
@@ -124,7 +141,7 @@ func (obj *ExecRes) Watch(processChan chan Event) {
 			cmdName = split[0]
 			//d, _ := os.Getwd() // TODO: how does this ever error ?
 			//cmdName = path.Join(d, cmdName)
-			cmdArgs = split[1:len(split)]
+			cmdArgs = split[1:]
 		} else {
 			cmdName = obj.Shell // usually bash, or sh
 			cmdArgs = []string{"-c", obj.WatchCmd}
@@ -134,8 +151,7 @@ func (obj *ExecRes) Watch(processChan chan Event) {
 
 		cmdReader, err := cmd.StdoutPipe()
 		if err != nil {
-			log.Printf("%v[%v]: Error creating StdoutPipe for Cmd: %v", obj.Kind(), obj.GetName(), err)
+			return fmt.Errorf("%s[%s]: Error creating StdoutPipe for Cmd: %v", obj.Kind(), obj.GetName(), err)
-			log.Fatal(err) // XXX: how should we handle errors?
 		}
 		scanner := bufio.NewScanner(cmdReader)
 
@@ -146,18 +162,17 @@ func (obj *ExecRes) Watch(processChan chan Event) {
 			cmd.Process.Kill() // TODO: is this necessary?
 		}()
 		if err := cmd.Start(); err != nil {
-			log.Printf("%v[%v]: Error starting Cmd: %v", obj.Kind(), obj.GetName(), err)
+			return fmt.Errorf("%s[%s]: Error starting Cmd: %v", obj.Kind(), obj.GetName(), err)
-			log.Fatal(err) // XXX: how should we handle errors?
 		}
 
 		bufioch, errch = obj.BufioChanScanner(scanner)
 	}
 
 	for {
-		obj.SetState(resStateWatching) // reset
+		obj.SetState(ResStateWatching) // reset
 		select {
 		case text := <-bufioch:
-			obj.SetConvergedState(resConvergedNil)
+			cuid.SetConverged(false)
 			// each time we get a line of output, we loop!
 			log.Printf("%v[%v]: Watch output: %s", obj.Kind(), obj.GetName(), text)
 			if text != "" {
@@ -165,43 +180,47 @@ func (obj *ExecRes) Watch(processChan chan Event) {
 			}
 
 		case err := <-errch:
-			obj.SetConvergedState(resConvergedNil) // XXX ?
+			cuid.SetConverged(false)
-			if err == nil {                        // EOF
+			if err == nil { // EOF
 				// FIXME: add an "if watch command ends/crashes"
 				// restart or generate error option
-				log.Printf("%v[%v]: Reached EOF", obj.Kind(), obj.GetName())
+				return fmt.Errorf("%s[%s]: Reached EOF", obj.Kind(), obj.GetName())
-				return
 			}
-			log.Printf("%v[%v]: Error reading input?: %v", obj.Kind(), obj.GetName(), err)
+			// error reading input?
-			log.Fatal(err)
+			return fmt.Errorf("Unknown %s[%s] error: %v", obj.Kind(), obj.GetName(), err)
-			// XXX: how should we handle errors?
 
-		case event := <-obj.events:
+		case event := <-obj.Events():
-			obj.SetConvergedState(resConvergedNil)
+			cuid.SetConverged(false)
 			if exit, send = obj.ReadEvent(&event); exit {
-				return // exit
+				return nil // exit
 			}
 
-		case _ = <-TimeAfterOrBlock(obj.ctimeout):
+		case <-cuid.ConvergedTimer():
-			obj.SetConvergedState(resConvergedTimeout)
+			cuid.SetConverged(true) // converged!
-			obj.converged <- true
 			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
 		}
 
 		// do all our event sending all together to avoid duplicate msgs
 		if send {
+			startup = true // startup finished
 			send = false
 			// it is okay to invalidate the clean state on poke too
 			obj.isStateOK = false // something made state dirty
-			resp := NewResp()
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
-			processChan <- Event{eventNil, resp, "", true} // trigger process
+				return err // we exit or bubble up a NACK...
-			resp.ACKWait()                                 // wait for the ACK()
+			}
 		}
 	}
 }
 
+// CheckApply checks the resource state and applies the resource if the bool
+// input is true. It returns error info and if the state check passed or not.
 // TODO: expand the IfCmd to be a list of commands
-func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
+func (obj *ExecRes) CheckApply(apply bool) (checkok bool, err error) {
 	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
 
 	// if there is a watch command, but no if command, run based on state
@@ -216,8 +235,8 @@ func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
 		//} else if obj.IfCmd != "" && obj.WatchCmd != "" {
 
 		if obj.PollInt > 0 { // && obj.WatchCmd == ""
-			// XXX have the Watch() command output onlyif poll events...
+			// XXX: have the Watch() command output onlyif poll events...
-			// XXX we can optimize by saving those results for returning here
+			// XXX: we can optimize by saving those results for returning here
 			// return XXX
 		}
 
@@ -230,7 +249,7 @@ func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
 			cmdName = split[0]
 			//d, _ := os.Getwd() // TODO: how does this ever error ?
 			//cmdName = path.Join(d, cmdName)
-			cmdArgs = split[1:len(split)]
+			cmdArgs = split[1:]
 		} else {
 			cmdName = obj.IfShell // usually bash, or sh
 			cmdArgs = []string{"-c", obj.IfCmd}
@@ -266,7 +285,7 @@ func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
 		cmdName = split[0]
 		//d, _ := os.Getwd() // TODO: how does this ever error ?
 		//cmdName = path.Join(d, cmdName)
-		cmdArgs = split[1:len(split)]
+		cmdArgs = split[1:]
 	} else {
 		cmdName = obj.Shell // usually bash, or sh
 		cmdArgs = []string{"-c", obj.Cmd}
@@ -295,7 +314,7 @@ func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
 			return false, err
 		}
 
-	case <-TimeAfterOrBlock(timeout):
+	case <-util.TimeAfterOrBlock(timeout):
 		log.Printf("%v[%v]: Timeout waiting for Cmd", obj.Kind(), obj.GetName())
 		//cmd.Process.Kill() // TODO: is this necessary?
 		return false, errors.New("Timeout waiting for Cmd!")
@@ -320,17 +339,17 @@ func (obj *ExecRes) CheckApply(apply bool) (stateok bool, err error) {
 	return false, nil    // success
 }
 
-type ExecUUID struct {
+// ExecUID is the UID struct for ExecRes.
-	BaseUUID
+type ExecUID struct {
+	BaseUID
 	Cmd   string
 	IfCmd string
 	// TODO: add more elements here
 }
 
-// if and only if they are equivalent, return true
+// IFF aka if and only if they are equivalent, return true. If not, false.
-// if they are not equivalent, return false
+func (obj *ExecUID) IFF(uid ResUID) bool {
-func (obj *ExecUUID) IFF(uuid ResUUID) bool {
+	res, ok := uid.(*ExecUID)
-	res, ok := uuid.(*ExecUUID)
 	if !ok {
 		return false
 	}
@@ -362,35 +381,43 @@ func (obj *ExecUUID) IFF(uuid ResUUID) bool {
 	return true
 }
 
+// AutoEdges returns the AutoEdge interface. In this case no autoedges are used.
 func (obj *ExecRes) AutoEdges() AutoEdge {
 	// TODO: parse as many exec params to look for auto edges, for example
 	// the path of the binary in the Cmd variable might be from in a pkg
 	return nil
 }
 
-// include all params to make a unique identification of this object
+// GetUIDs includes all params to make a unique identification of this object.
-func (obj *ExecRes) GetUUIDs() []ResUUID {
+// Most resources only return one, although some resources can return multiple.
-	x := &ExecUUID{
+func (obj *ExecRes) GetUIDs() []ResUID {
-		BaseUUID: BaseUUID{name: obj.GetName(), kind: obj.Kind()},
+	x := &ExecUID{
-		Cmd:      obj.Cmd,
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
-		IfCmd:    obj.IfCmd,
+		Cmd:     obj.Cmd,
+		IfCmd:   obj.IfCmd,
 		// TODO: add more params here
 	}
-	return []ResUUID{x}
+	return []ResUID{x}
 }
 
+// GroupCmp returns whether two resources can be grouped together or not.
 func (obj *ExecRes) GroupCmp(r Res) bool {
-	_, ok := r.(*SvcRes)
+	_, ok := r.(*ExecRes)
 	if !ok {
 		return false
 	}
 	return false // not possible atm
 }
 
+// Compare two resources and return if they are equivalent.
 func (obj *ExecRes) Compare(res Res) bool {
 	switch res.(type) {
 	case *ExecRes:
 		res := res.(*ExecRes)
+		if !obj.BaseRes.Compare(res) { // call base Compare
+			return false
+		}
+
 		if obj.Name != res.Name {
 			return false
 		}

resources/file.go

@@ -0,0 +1,815 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package resources
+
+import (
+	"bytes"
+	"crypto/sha256"
+	"encoding/gob"
+	"encoding/hex"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"log"
+	"os"
+	"path"
+	"path/filepath"
+	"strings"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/global" // XXX: package mgmtmain instead?
+	"github.com/purpleidea/mgmt/recwatch"
+	"github.com/purpleidea/mgmt/util"
+)
+
+func init() {
+	gob.Register(&FileRes{})
+}
+
+// FileRes is a file and directory resource.
+type FileRes struct {
+	BaseRes    `yaml:",inline"`
+	Path       string `yaml:"path"` // path variable (should default to name)
+	Dirname    string `yaml:"dirname"`
+	Basename   string `yaml:"basename"`
+	Content    string `yaml:"content"` // FIXME: how do you describe: "leave content alone" - state = "create" ?
+	Source     string `yaml:"source"`  // file path for source content
+	State      string `yaml:"state"`   // state: exists/present?, absent, (undefined?)
+	Recurse    bool   `yaml:"recurse"`
+	Force      bool   `yaml:"force"`
+	path       string // computed path
+	isDir      bool   // computed isDir
+	sha256sum  string
+	recWatcher *recwatch.RecWatcher
+}
+
+// NewFileRes is a constructor for this resource. It also calls Init() for you.
+func NewFileRes(name, path, dirname, basename, content, source, state string, recurse, force bool) (*FileRes, error) {
+	obj := &FileRes{
+		BaseRes: BaseRes{
+			Name: name,
+		},
+		Path:     path,
+		Dirname:  dirname,
+		Basename: basename,
+		Content:  content,
+		Source:   source,
+		State:    state,
+		Recurse:  recurse,
+		Force:    force,
+	}
+	return obj, obj.Init()
+}
+
+// Init runs some startup code for this resource.
+func (obj *FileRes) Init() error {
+	obj.sha256sum = ""
+	if obj.Path == "" { // use the name as the path default if missing
+		obj.Path = obj.BaseRes.Name
+	}
+	obj.path = obj.GetPath()                     // compute once
+	obj.isDir = strings.HasSuffix(obj.path, "/") // dirs have trailing slashes
+
+	obj.BaseRes.kind = "File"
+	return obj.BaseRes.Init() // call base init, b/c we're overriding
+}
+
+// GetPath returns the actual path to use for this resource. It computes this
+// after analysis of the Path, Dirname and Basename values. Dirs end with slash.
+func (obj *FileRes) GetPath() string {
+	d := util.Dirname(obj.Path)
+	b := util.Basename(obj.Path)
+	if obj.Dirname == "" && obj.Basename == "" {
+		return obj.Path
+	}
+	if obj.Dirname == "" {
+		return d + obj.Basename
+	}
+	if obj.Basename == "" {
+		return obj.Dirname + b
+	}
+	// if obj.dirname != "" && obj.basename != ""
+	return obj.Dirname + obj.Basename
+}
+
+// Validate reports any problems with the struct definition.
+func (obj *FileRes) Validate() error {
+	if obj.Dirname != "" && !strings.HasSuffix(obj.Dirname, "/") {
+		return fmt.Errorf("Dirname must end with a slash.")
+	}
+
+	if strings.HasPrefix(obj.Basename, "/") {
+		return fmt.Errorf("Basename must not start with a slash.")
+	}
+
+	if obj.Content != "" && obj.Source != "" {
+		return fmt.Errorf("Can't specify both Content and Source.")
+	}
+
+	if obj.isDir && obj.Content != "" { // makes no sense
+		return fmt.Errorf("Can't specify Content when creating a Dir.")
+	}
+
+	// XXX: should this specify that we create an empty directory instead?
+	//if obj.Source == "" && obj.isDir {
+	//	return fmt.Errorf("Can't specify an empty source when creating a Dir.")
+	//}
+
+	return nil
+}
+
+// Watch is the primary listener for this resource and it outputs events.
+// This one is a file watcher for files and directories.
+// Modify with caution, it is probably important to write some test cases first!
+// If the Watch returns an error, it means that something has gone wrong, and it
+// must be restarted. On a clean exit it returns nil.
+// FIXME: Also watch the source directory when using obj.Source !!!
+func (obj *FileRes) Watch(processChan chan event.Event) error {
+	if obj.IsWatching() {
+		return nil // TODO: should this be an error?
+	}
+	obj.SetWatching(true)
+	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
+
+	var err error
+	obj.recWatcher, err = recwatch.NewRecWatcher(obj.Path, obj.Recurse)
+	if err != nil {
+		return err
+	}
+	defer obj.recWatcher.Close()
+
+	var send = false // send event?
+	var exit = false
+	var dirty = false
+
+	for {
+		if global.DEBUG {
+			log.Printf("%s[%s]: Watching: %s", obj.Kind(), obj.GetName(), obj.Path) // attempting to watch...
+		}
+
+		obj.SetState(ResStateWatching) // reset
+		select {
+		case event, ok := <-obj.recWatcher.Events():
+			if !ok { // channel shutdown
+				return nil
+			}
+			cuid.SetConverged(false)
+			if err := event.Error; err != nil {
+				return fmt.Errorf("Unknown %s[%s] watcher error: %v", obj.Kind(), obj.GetName(), err)
+			}
+			if global.DEBUG { // don't access event.Body if event.Error isn't nil
+				log.Printf("%s[%s]: Event(%s): %v", obj.Kind(), obj.GetName(), event.Body.Name, event.Body.Op)
+			}
+			send = true
+			dirty = true
+
+		case event := <-obj.Events():
+			cuid.SetConverged(false)
+			if exit, send = obj.ReadEvent(&event); exit {
+				return nil // exit
+			}
+			//dirty = false // these events don't invalidate state
+
+		case <-cuid.ConvergedTimer():
+			cuid.SetConverged(true) // converged!
+			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
+			dirty = true
+		}
+
+		// do all our event sending all together to avoid duplicate msgs
+		if send {
+			startup = true // startup finished
+			send = false
+			// only invalid state on certain types of events
+			if dirty {
+				dirty = false
+				obj.isStateOK = false // something made state dirty
+			}
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
+				return err // we exit or bubble up a NACK...
+			}
+		}
+	}
+}
+
+// smartPath adds a trailing slash to the path if it is a directory.
+func smartPath(fileInfo os.FileInfo) string {
+	smartPath := fileInfo.Name() // absolute path
+	if fileInfo.IsDir() {
+		smartPath += "/" // add a trailing slash for dirs
+	}
+	return smartPath
+}
+
+// FileInfo is an enhanced variant of the traditional os.FileInfo struct. It can
+// store both the absolute and the relative paths (when built from our ReadDir),
+// and those two paths contain a trailing slash when they refer to a directory.
+type FileInfo struct {
+	os.FileInfo        // embed
+	AbsPath     string // smart variant
+	RelPath     string // smart variant
+}
+
+// ReadDir reads a directory path, and returns a list of enhanced FileInfo's.
+func ReadDir(path string) ([]FileInfo, error) {
+	if !strings.HasSuffix(path, "/") { // dirs have trailing slashes
+		return nil, fmt.Errorf("Path must be a directory.")
+	}
+	output := []FileInfo{} // my file info
+	fileInfos, err := ioutil.ReadDir(path)
+	if os.IsNotExist(err) {
+		return output, err // return empty list
+	}
+	if err != nil {
+		return nil, err
+	}
+	for _, fi := range fileInfos {
+		abs := path + smartPath(fi)
+		rel, err := filepath.Rel(path, abs) // NOTE: calls Clean()
+		if err != nil {                     // shouldn't happen
+			return nil, fmt.Errorf("ReadDir: Unhandled error: %v", err)
+		}
+		if fi.IsDir() {
+			rel += "/" // add a trailing slash for dirs
+		}
+		x := FileInfo{
+			FileInfo: fi,
+			AbsPath:  abs,
+			RelPath:  rel,
+		}
+		output = append(output, x)
+	}
+	return output, nil
+}
+
+// smartMapPaths adds a trailing slash to every path that is a directory. It
+// returns the data as a map where the keys are the smart paths and where the
+// values are the original os.FileInfo entries.
+func mapPaths(fileInfos []FileInfo) map[string]FileInfo {
+	paths := make(map[string]FileInfo)
+	for _, fileInfo := range fileInfos {
+		paths[fileInfo.RelPath] = fileInfo
+	}
+	return paths
+}
+
+// fileCheckApply is the CheckApply operation for a source and destination file.
+// It can accept an io.Reader as the source, which can be a regular file, or it
+// can be a bytes Buffer struct. It can take an input sha256 hash to use instead
+// of computing the source data hash, and it returns the computed value if this
+// function reaches that stage. As usual, it respects the apply action variable,
+// and it symmetry with the main CheckApply function returns checkOK and error.
+func (obj *FileRes) fileCheckApply(apply bool, src io.ReadSeeker, dst string, sha256sum string) (string, bool, error) {
+	// TODO: does it make sense to switch dst to an io.Writer ?
+	// TODO: use obj.Force when dealing with symlinks and other file types!
+	if global.DEBUG {
+		log.Printf("fileCheckApply: %s -> %s", src, dst)
+	}
+
+	srcFile, isFile := src.(*os.File)
+	_, isBytes := src.(*bytes.Reader) // supports seeking!
+	if !isFile && !isBytes {
+		return "", false, fmt.Errorf("Can't open src as either file or buffer!")
+	}
+
+	var srcStat os.FileInfo
+	if isFile {
+		var err error
+		srcStat, err = srcFile.Stat()
+		if err != nil {
+			return "", false, err
+		}
+		// TODO: deal with symlinks
+		if !srcStat.Mode().IsRegular() { // can't copy non-regular files or dirs
+			return "", false, fmt.Errorf("Non-regular src file: %s (%q)", srcStat.Name(), srcStat.Mode())
+		}
+	}
+
+	dstFile, err := os.Open(dst)
+	if err != nil && !os.IsNotExist(err) { // ignore ErrNotExist errors
+		return "", false, err
+	}
+	dstClose := func() error {
+		return dstFile.Close() // calling this twice is safe :)
+	}
+	defer dstClose()
+	dstExists := !os.IsNotExist(err)
+
+	dstStat, err := dstFile.Stat()
+	if err != nil && dstExists {
+		return "", false, err
+	}
+
+	if dstExists && dstStat.IsDir() { // oops, dst is a dir, and we want a file...
+		if !apply {
+			return "", false, nil
+		}
+		if !obj.Force {
+			return "", false, fmt.Errorf("Can't force dir into file: %s", dst)
+		}
+
+		cleanDst := path.Clean(dst)
+		if cleanDst == "" || cleanDst == "/" {
+			return "", false, fmt.Errorf("Don't want to remove root!") // safety
+		}
+		// FIXME: respect obj.Recurse here...
+		// there is a dir here, where we want a file...
+		log.Printf("fileCheckApply: Removing (force): %s", cleanDst)
+		if err := os.RemoveAll(cleanDst); err != nil { // dangerous ;)
+			return "", false, err
+		}
+		dstExists = false // now it's gone!
+
+	} else if err == nil {
+		if !dstStat.Mode().IsRegular() {
+			return "", false, fmt.Errorf("Non-regular dst file: %s (%q)", dstStat.Name(), dstStat.Mode())
+		}
+		if isFile && os.SameFile(srcStat, dstStat) { // same inode, we're done!
+			return "", true, nil
+		}
+	}
+
+	if dstExists { // if dst doesn't exist, no need to compare hashes
+		// hash comparison (efficient because we can cache hash of content str)
+		if sha256sum == "" { // cache is invalid
+			hash := sha256.New()
+			// TODO: file existence test?
+			if _, err := io.Copy(hash, src); err != nil {
+				return "", false, err
+			}
+			sha256sum = hex.EncodeToString(hash.Sum(nil))
+			// since we re-use this src handler below, it is
+			// *critical* to seek to 0, or we'll copy nothing!
+			if n, err := src.Seek(0, 0); err != nil || n != 0 {
+				return sha256sum, false, err
+			}
+		}
+
+		// dst hash
+		hash := sha256.New()
+		if _, err := io.Copy(hash, dstFile); err != nil {
+			return "", false, err
+		}
+		if h := hex.EncodeToString(hash.Sum(nil)); h == sha256sum {
+			return sha256sum, true, nil // same!
+		}
+	}
+
+	// state is not okay, no work done, exit, but without error
+	if !apply {
+		return sha256sum, false, nil
+	}
+	if global.DEBUG {
+		log.Printf("fileCheckApply: Apply: %s -> %s", src, dst)
+	}
+
+	dstClose() // unlock file usage so we can write to it
+	dstFile, err = os.Create(dst)
+	if err != nil {
+		return sha256sum, false, err
+	}
+	defer dstFile.Close() // TODO: is this redundant because of the earlier defered Close() ?
+
+	if isFile { // set mode because it's a new file
+		if err := dstFile.Chmod(srcStat.Mode()); err != nil {
+			return sha256sum, false, err
+		}
+	}
+
+	// TODO: attempt to reflink with Splice() and int(file.Fd()) as input...
+	// syscall.Splice(rfd int, roff *int64, wfd int, woff *int64, len int, flags int) (n int64, err error)
+
+	// TODO: should we offer a way to cancel the copy on ^C ?
+	if global.DEBUG {
+		log.Printf("fileCheckApply: Copy: %s -> %s", src, dst)
+	}
+	if n, err := io.Copy(dstFile, src); err != nil {
+		return sha256sum, false, err
+	} else if global.DEBUG {
+		log.Printf("fileCheckApply: Copied: %v", n)
+	}
+	return sha256sum, false, dstFile.Sync()
+}
+
+// syncCheckApply is the CheckApply operation for a source and destination dir.
+// It is recursive and can create directories directly, and files via the usual
+// fileCheckApply method. It returns checkOK and error as is normally expected.
+func (obj *FileRes) syncCheckApply(apply bool, src, dst string) (bool, error) {
+	if global.DEBUG {
+		log.Printf("syncCheckApply: %s -> %s", src, dst)
+	}
+	if src == "" || dst == "" {
+		return false, fmt.Errorf("The src and dst must not be empty!")
+	}
+
+	var checkOK = true
+	// TODO: handle ./ cases or ../ cases that need cleaning ?
+
+	srcIsDir := strings.HasSuffix(src, "/")
+	dstIsDir := strings.HasSuffix(dst, "/")
+
+	if srcIsDir != dstIsDir {
+		return false, fmt.Errorf("The src and dst must be both either files or directories.")
+	}
+
+	if !srcIsDir && !dstIsDir {
+		if global.DEBUG {
+			log.Printf("syncCheckApply: %s -> %s", src, dst)
+		}
+		fin, err := os.Open(src)
+		if err != nil {
+			if global.DEBUG && os.IsNotExist(err) { // if we get passed an empty src
+				log.Printf("syncCheckApply: Missing src: %s", src)
+			}
+			return false, err
+		}
+
+		_, checkOK, err := obj.fileCheckApply(apply, fin, dst, "")
+		if err != nil {
+			fin.Close()
+			return false, err
+		}
+		return checkOK, fin.Close()
+	}
+
+	// else: if srcIsDir && dstIsDir
+	srcFiles, err := ReadDir(src)          // if src does not exist...
+	if err != nil && !os.IsNotExist(err) { // an empty map comes out below!
+		return false, err
+	}
+	dstFiles, err := ReadDir(dst)
+	if err != nil && !os.IsNotExist(err) {
+		return false, err
+	}
+	//log.Printf("syncCheckApply: srcFiles: %v", srcFiles)
+	//log.Printf("syncCheckApply: dstFiles: %v", dstFiles)
+	smartSrc := mapPaths(srcFiles)
+	smartDst := mapPaths(dstFiles)
+
+	for relPath, fileInfo := range smartSrc {
+		absSrc := fileInfo.AbsPath // absolute path
+		absDst := dst + relPath    // absolute dest
+
+		if _, exists := smartDst[relPath]; !exists {
+			if fileInfo.IsDir() {
+				if !apply { // only checking and not identical!
+					return false, nil
+				}
+
+				// file exists, but we want a dir: we need force
+				// we check for the file w/o the smart dir slash
+				relPathFile := strings.TrimSuffix(relPath, "/")
+				if _, ok := smartDst[relPathFile]; ok {
+					absCleanDst := path.Clean(absDst)
+					if !obj.Force {
+						return false, fmt.Errorf("Can't force file into dir: %s", absCleanDst)
+					}
+					if absCleanDst == "" || absCleanDst == "/" {
+						return false, fmt.Errorf("Don't want to remove root!") // safety
+					}
+					log.Printf("syncCheckApply: Removing (force): %s", absCleanDst)
+					if err := os.Remove(absCleanDst); err != nil {
+						return false, err
+					}
+					delete(smartDst, relPathFile) // rm from purge list
+				}
+
+				if global.DEBUG {
+					log.Printf("syncCheckApply: mkdir -m %s '%s'", fileInfo.Mode(), absDst)
+				}
+				if err := os.Mkdir(absDst, fileInfo.Mode()); err != nil {
+					return false, err
+				}
+				checkOK = false // we did some work
+			}
+			// if we're a regular file, the recurse will create it
+		}
+
+		if global.DEBUG {
+			log.Printf("syncCheckApply: Recurse: %s -> %s", absSrc, absDst)
+		}
+		if obj.Recurse {
+			if c, err := obj.syncCheckApply(apply, absSrc, absDst); err != nil { // recurse
+				return false, fmt.Errorf("syncCheckApply: Recurse failed: %v", err)
+			} else if !c { // don't let subsequent passes make this true
+				checkOK = false
+			}
+		}
+		if !apply && !checkOK { // check failed, and no apply to do, so exit!
+			return false, nil
+		}
+		delete(smartDst, relPath) // rm from purge list
+	}
+
+	if !apply && len(smartDst) > 0 { // we know there are files to remove!
+		return false, nil // so just exit now
+	}
+	// any files that now remain in smartDst need to be removed...
+	for relPath, fileInfo := range smartDst {
+		absSrc := src + relPath    // absolute dest (should not exist!)
+		absDst := fileInfo.AbsPath // absolute path (should get removed)
+		absCleanDst := path.Clean(absDst)
+		if absCleanDst == "" || absCleanDst == "/" {
+			return false, fmt.Errorf("Don't want to remove root!") // safety
+		}
+
+		// FIXME: respect obj.Recurse here...
+
+		// NOTE: we could use os.RemoveAll instead of recursing, but I
+		// think the symmetry is more elegant and correct here for now
+		// Avoiding this is also useful if we had a recurse limit arg!
+		if true { // switch
+			log.Printf("syncCheckApply: Removing: %s", absCleanDst)
+			if apply {
+				if err := os.RemoveAll(absCleanDst); err != nil { // dangerous ;)
+					return false, err
+				}
+				checkOK = false
+			}
+			continue
+		}
+		_ = absSrc
+		//log.Printf("syncCheckApply: Recurse rm: %s -> %s", absSrc, absDst)
+		//if c, err := obj.syncCheckApply(apply, absSrc, absDst); err != nil {
+		//	return false, fmt.Errorf("syncCheckApply: Recurse rm failed: %v", err)
+		//} else if !c { // don't let subsequent passes make this true
+		//	checkOK = false
+		//}
+		//log.Printf("syncCheckApply: Removing: %s", absCleanDst)
+		//if apply { // safety
+		//	if err := os.Remove(absCleanDst); err != nil {
+		//		return false, err
+		//	}
+		//	checkOK = false
+		//}
+	}
+
+	return checkOK, nil
+}
+
+// contentCheckApply performs a CheckApply for the file existence and content.
+func (obj *FileRes) contentCheckApply(apply bool) (checkOK bool, _ error) {
+	log.Printf("%v[%v]: contentCheckApply(%t)", obj.Kind(), obj.GetName(), apply)
+
+	if obj.State == "absent" {
+		if _, err := os.Stat(obj.path); os.IsNotExist(err) {
+			// no such file or directory, but
+			// file should be missing, phew :)
+			return true, nil
+
+		} else if err != nil { // what could this error be?
+			return false, err
+		}
+
+		// state is not okay, no work done, exit, but without error
+		if !apply {
+			return false, nil
+		}
+
+		// apply portion
+		if obj.path == "" || obj.path == "/" {
+			return false, fmt.Errorf("Don't want to remove root!") // safety
+		}
+		log.Printf("contentCheckApply: Removing: %s", obj.path)
+		// FIXME: respect obj.Recurse here...
+		// TODO: add recurse limit here
+		err := os.RemoveAll(obj.path) // dangerous ;)
+		return false, err             // either nil or not
+	}
+
+	if obj.Source == "" { // do the obj.Content checks first...
+		if obj.isDir { // TODO: should we create an empty dir this way?
+			log.Fatal("XXX: Not implemented!") // XXX
+		}
+
+		bufferSrc := bytes.NewReader([]byte(obj.Content))
+		sha256sum, checkOK, err := obj.fileCheckApply(apply, bufferSrc, obj.path, obj.sha256sum)
+		if sha256sum != "" { // empty values mean errored or didn't hash
+			// this can be valid even when the whole function errors
+			obj.sha256sum = sha256sum // cache value
+		}
+		if err != nil {
+			return false, err
+		}
+		// if no err, but !ok, then...
+		return checkOK, nil // success
+	}
+
+	checkOK, err := obj.syncCheckApply(apply, obj.Source, obj.path)
+	if err != nil {
+		log.Printf("syncCheckApply: Error: %v", err)
+		return false, err
+	}
+
+	return checkOK, nil
+}
+
+// CheckApply checks the resource state and applies the resource if the bool
+// input is true. It returns error info and if the state check passed or not.
+func (obj *FileRes) CheckApply(apply bool) (checkOK bool, _ error) {
+	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
+
+	if obj.isStateOK { // cache the state
+		return true, nil
+	}
+
+	checkOK = true
+
+	if c, err := obj.contentCheckApply(apply); err != nil {
+		return false, err
+	} else if !c {
+		checkOK = false
+	}
+
+	// TODO
+	//if c, err := obj.chmodCheckApply(apply); err != nil {
+	//	return false, err
+	//} else if !c {
+	//	checkOK = false
+	//}
+
+	// TODO
+	//if c, err := obj.chownCheckApply(apply); err != nil {
+	//	return false, err
+	//} else if !c {
+	//	checkOK = false
+	//}
+
+	// if we did work successfully, or are in a good state, then state is ok
+	if apply || checkOK {
+		obj.isStateOK = true
+	}
+	return checkOK, nil // w00t
+}
+
+// FileUID is the UID struct for FileRes.
+type FileUID struct {
+	BaseUID
+	path string
+}
+
+// IFF aka if and only if they are equivalent, return true. If not, false.
+func (obj *FileUID) IFF(uid ResUID) bool {
+	res, ok := uid.(*FileUID)
+	if !ok {
+		return false
+	}
+	return obj.path == res.path
+}
+
+// FileResAutoEdges holds the state of the auto edge generator.
+type FileResAutoEdges struct {
+	data    []ResUID
+	pointer int
+	found   bool
+}
+
+// Next returns the next automatic edge.
+func (obj *FileResAutoEdges) Next() []ResUID {
+	if obj.found {
+		log.Fatal("Shouldn't be called anymore!")
+	}
+	if len(obj.data) == 0 { // check length for rare scenarios
+		return nil
+	}
+	value := obj.data[obj.pointer]
+	obj.pointer++
+	return []ResUID{value} // we return one, even though api supports N
+}
+
+// Test gets results of the earlier Next() call, & returns if we should continue!
+func (obj *FileResAutoEdges) Test(input []bool) bool {
+	// if there aren't any more remaining
+	if len(obj.data) <= obj.pointer {
+		return false
+	}
+	if obj.found { // already found, done!
+		return false
+	}
+	if len(input) != 1 { // in case we get given bad data
+		log.Fatal("Expecting a single value!")
+	}
+	if input[0] { // if a match is found, we're done!
+		obj.found = true // no more to find!
+		return false
+	}
+	return true // keep going
+}
+
+// AutoEdges generates a simple linear sequence of each parent directory from
+// the bottom up!
+func (obj *FileRes) AutoEdges() AutoEdge {
+	var data []ResUID                              // store linear result chain here...
+	values := util.PathSplitFullReversed(obj.path) // build it
+	_, values = values[0], values[1:]              // get rid of first value which is me!
+	for _, x := range values {
+		var reversed = true // cheat by passing a pointer
+		data = append(data, &FileUID{
+			BaseUID: BaseUID{
+				name:     obj.GetName(),
+				kind:     obj.Kind(),
+				reversed: &reversed,
+			},
+			path: x, // what matters
+		}) // build list
+	}
+	return &FileResAutoEdges{
+		data:    data,
+		pointer: 0,
+		found:   false,
+	}
+}
+
+// GetUIDs includes all params to make a unique identification of this object.
+// Most resources only return one, although some resources can return multiple.
+func (obj *FileRes) GetUIDs() []ResUID {
+	x := &FileUID{
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
+		path:    obj.path,
+	}
+	return []ResUID{x}
+}
+
+// GroupCmp returns whether two resources can be grouped together or not.
+func (obj *FileRes) GroupCmp(r Res) bool {
+	_, ok := r.(*FileRes)
+	if !ok {
+		return false
+	}
+	// TODO: we might be able to group directory children into a single
+	// recursive watcher in the future, thus saving fanotify watches
+	return false // not possible atm
+}
+
+// Compare two resources and return if they are equivalent.
+func (obj *FileRes) Compare(res Res) bool {
+	switch res.(type) {
+	case *FileRes:
+		res := res.(*FileRes)
+		if !obj.BaseRes.Compare(res) { // call base Compare
+			return false
+		}
+
+		if obj.Name != res.Name {
+			return false
+		}
+		if obj.path != res.Path {
+			return false
+		}
+		if obj.Content != res.Content {
+			return false
+		}
+		if obj.Source != res.Source {
+			return false
+		}
+		if obj.State != res.State {
+			return false
+		}
+		if obj.Recurse != res.Recurse {
+			return false
+		}
+		if obj.Force != res.Force {
+			return false
+		}
+	default:
+		return false
+	}
+	return true
+}
+
+// CollectPattern applies the pattern for collection resources.
+func (obj *FileRes) CollectPattern(pattern string) {
+	// XXX: currently the pattern for files can only override the Dirname variable :P
+	obj.Dirname = pattern // XXX: simplistic for now
+}

resources/msg.go

@@ -0,0 +1,271 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package resources
+
+import (
+	"encoding/gob"
+	"fmt"
+	"log"
+	"regexp"
+	"strings"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+
+	"github.com/coreos/go-systemd/journal"
+)
+
+func init() {
+	gob.Register(&MsgRes{})
+}
+
+// MsgRes is a resource that writes messages to logs.
+type MsgRes struct {
+	BaseRes        `yaml:",inline"`
+	Body           string            `yaml:"body"`
+	Priority       string            `yaml:"priority"`
+	Fields         map[string]string `yaml:"fields"`
+	Journal        bool              `yaml:"journal"` // enable systemd journal output
+	Syslog         bool              `yaml:"syslog"`  // enable syslog output
+	logStateOK     bool
+	journalStateOK bool
+	syslogStateOK  bool
+}
+
+// MsgUID is a unique representation for a MsgRes object.
+type MsgUID struct {
+	BaseUID
+	body string
+}
+
+// NewMsgRes is a constructor for this resource.
+func NewMsgRes(name, body, priority string, journal, syslog bool, fields map[string]string) (*MsgRes, error) {
+	message := name
+	if body != "" {
+		message = body
+	}
+
+	obj := &MsgRes{
+		BaseRes: BaseRes{
+			Name: name,
+		},
+		Body:     message,
+		Priority: priority,
+		Fields:   fields,
+		Journal:  journal,
+		Syslog:   syslog,
+	}
+
+	return obj, obj.Init()
+}
+
+// Init runs some startup code for this resource.
+func (obj *MsgRes) Init() error {
+	obj.BaseRes.kind = "Msg"
+	return obj.BaseRes.Init() // call base init, b/c we're overrriding
+}
+
+// Validate the params that are passed to MsgRes
+func (obj *MsgRes) Validate() error {
+	invalidCharacters := regexp.MustCompile("[^a-zA-Z0-9_]")
+	for field := range obj.Fields {
+		if invalidCharacters.FindString(field) != "" {
+			return fmt.Errorf("Invalid character in field %s.", field)
+		}
+		if strings.HasPrefix(field, "_") {
+			return fmt.Errorf("Fields cannot begin with _.")
+		}
+	}
+	return nil
+}
+
+// Watch is the primary listener for this resource and it outputs events.
+func (obj *MsgRes) Watch(processChan chan event.Event) error {
+	if obj.IsWatching() {
+		return nil
+	}
+	obj.SetWatching(true)
+	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
+
+	var send = false // send event?
+	var exit = false
+	for {
+		obj.SetState(ResStateWatching) // reset
+		select {
+		case event := <-obj.Events():
+			cuid.SetConverged(false)
+			// we avoid sending events on unpause
+			if exit, send = obj.ReadEvent(&event); exit {
+				return nil // exit
+			}
+
+			/*
+				// TODO: invalidate cached state on poke events
+				obj.logStateOK = false
+				if obj.Journal {
+					obj.journalStateOK = false
+				}
+				if obj.Syslog {
+					obj.syslogStateOK = false
+				}
+			*/
+			send = true
+
+		case <-cuid.ConvergedTimer():
+			cuid.SetConverged(true) // converged!
+			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
+		}
+
+		// do all our event sending all together to avoid duplicate msgs
+		if send {
+			startup = true // startup finished
+			send = false
+			// only do this on certain types of events
+			//obj.isStateOK = false // something made state dirty
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
+				return err // we exit or bubble up a NACK...
+			}
+		}
+	}
+}
+
+// GetUIDs includes all params to make a unique identification of this object.
+// Most resources only return one, although some resources can return multiple.
+func (obj *MsgRes) GetUIDs() []ResUID {
+	x := &MsgUID{
+		BaseUID: BaseUID{
+			name: obj.GetName(),
+			kind: obj.Kind(),
+		},
+		body: obj.Body,
+	}
+	return []ResUID{x}
+}
+
+// AutoEdges returns the AutoEdges. In this case none are used.
+func (obj *MsgRes) AutoEdges() AutoEdge {
+	return nil
+}
+
+// Compare two resources and return if they are equivalent.
+func (obj *MsgRes) Compare(res Res) bool {
+	switch res.(type) {
+	case *MsgRes:
+		res := res.(*MsgRes)
+		if !obj.BaseRes.Compare(res) {
+			return false
+		}
+		if obj.Body != res.Body {
+			return false
+		}
+		if obj.Priority != res.Priority {
+			return false
+		}
+		if len(obj.Fields) != len(res.Fields) {
+			return false
+		}
+		for field, value := range obj.Fields {
+			if res.Fields[field] != value {
+				return false
+			}
+		}
+	default:
+		return false
+	}
+	return true
+}
+
+// IsAllStateOK derives a compound state from all internal cache flags that apply to this resource.
+func (obj *MsgRes) isAllStateOK() bool {
+	if obj.Journal && !obj.journalStateOK {
+		return false
+	}
+	if obj.Syslog && !obj.syslogStateOK {
+		return false
+	}
+	return obj.logStateOK
+}
+
+// JournalPriority converts a string description to a numeric priority.
+// XXX: Have Validate() make sure it actually is one of these.
+func (obj *MsgRes) journalPriority() journal.Priority {
+	switch obj.Priority {
+	case "Emerg":
+		return journal.PriEmerg
+	case "Alert":
+		return journal.PriAlert
+	case "Crit":
+		return journal.PriCrit
+	case "Err":
+		return journal.PriErr
+	case "Warning":
+		return journal.PriWarning
+	case "Notice":
+		return journal.PriNotice
+	case "Info":
+		return journal.PriInfo
+	case "Debug":
+		return journal.PriDebug
+	}
+	return journal.PriNotice
+}
+
+// CheckApply method for Msg resource.
+// Every check leads to an apply, meaning that the message is flushed to the journal.
+func (obj *MsgRes) CheckApply(apply bool) (bool, error) {
+	log.Printf("%s[%s]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
+
+	if obj.isAllStateOK() {
+		return true, nil
+	}
+
+	if !obj.logStateOK {
+		log.Printf("%s[%s]: Body: %s", obj.Kind(), obj.GetName(), obj.Body)
+		obj.logStateOK = true
+	}
+
+	if !apply {
+		return false, nil
+	}
+	if obj.Journal && !obj.journalStateOK {
+		if err := journal.Send(obj.Body, obj.journalPriority(), obj.Fields); err != nil {
+			return false, err
+		}
+		obj.journalStateOK = true
+	}
+	if obj.Syslog && !obj.syslogStateOK {
+		// TODO: implement syslog client
+		obj.syslogStateOK = true
+	}
+	return false, nil
+}

resources/noop.go

@@ -15,107 +15,130 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-package main
+package resources
 
 import (
 	"encoding/gob"
 	"log"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
 )
 
 func init() {
 	gob.Register(&NoopRes{})
 }
 
+// NoopRes is a no-op resource that does nothing.
 type NoopRes struct {
 	BaseRes `yaml:",inline"`
 	Comment string `yaml:"comment"` // extra field for example purposes
 }
 
-func NewNoopRes(name string) *NoopRes {
+// NewNoopRes is a constructor for this resource. It also calls Init() for you.
+func NewNoopRes(name string) (*NoopRes, error) {
 	obj := &NoopRes{
 		BaseRes: BaseRes{
 			Name: name,
 		},
 		Comment: "",
 	}
-	obj.Init()
+	return obj, obj.Init()
-	return obj
 }
 
-func (obj *NoopRes) Init() {
+// Init runs some startup code for this resource.
+func (obj *NoopRes) Init() error {
 	obj.BaseRes.kind = "Noop"
-	obj.BaseRes.Init() // call base init, b/c we're overriding
+	return obj.BaseRes.Init() // call base init, b/c we're overriding
 }
 
-// validate if the params passed in are valid data
+// Validate if the params passed in are valid data.
 // FIXME: where should this get called ?
-func (obj *NoopRes) Validate() bool {
+func (obj *NoopRes) Validate() error {
-	return true
+	return nil
 }
 
-func (obj *NoopRes) Watch(processChan chan Event) {
+// Watch is the primary listener for this resource and it outputs events.
+func (obj *NoopRes) Watch(processChan chan event.Event) error {
 	if obj.IsWatching() {
-		return
+		return nil // TODO: should this be an error?
 	}
 	obj.SetWatching(true)
 	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
 
-	//vertex := obj.vertex // stored with SetVertex
 	var send = false // send event?
 	var exit = false
 	for {
-		obj.SetState(resStateWatching) // reset
+		obj.SetState(ResStateWatching) // reset
 		select {
-		case event := <-obj.events:
+		case event := <-obj.Events():
-			obj.SetConvergedState(resConvergedNil)
+			cuid.SetConverged(false)
 			// we avoid sending events on unpause
 			if exit, send = obj.ReadEvent(&event); exit {
-				return // exit
+				return nil // exit
 			}
 
-		case _ = <-TimeAfterOrBlock(obj.ctimeout):
+		case <-cuid.ConvergedTimer():
-			obj.SetConvergedState(resConvergedTimeout)
+			cuid.SetConverged(true) // converged!
-			obj.converged <- true
 			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
 		}
 
 		// do all our event sending all together to avoid duplicate msgs
 		if send {
+			startup = true // startup finished
 			send = false
 			// only do this on certain types of events
 			//obj.isStateOK = false // something made state dirty
-			resp := NewResp()
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
-			processChan <- Event{eventNil, resp, "", true} // trigger process
+				return err // we exit or bubble up a NACK...
-			resp.ACKWait()                                 // wait for the ACK()
+			}
 		}
 	}
 }
 
 // CheckApply method for Noop resource. Does nothing, returns happy!
-func (obj *NoopRes) CheckApply(apply bool) (stateok bool, err error) {
+func (obj *NoopRes) CheckApply(apply bool) (checkok bool, err error) {
 	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
 	return true, nil // state is always okay
 }
 
-type NoopUUID struct {
+// NoopUID is the UID struct for NoopRes.
-	BaseUUID
+type NoopUID struct {
+	BaseUID
 	name string
 }
 
+// AutoEdges returns the AutoEdge interface. In this case no autoedges are used.
 func (obj *NoopRes) AutoEdges() AutoEdge {
 	return nil
 }
 
-// include all params to make a unique identification of this object
+// GetUIDs includes all params to make a unique identification of this object.
-// most resources only return one, although some resources return multiple
+// Most resources only return one, although some resources can return multiple.
-func (obj *NoopRes) GetUUIDs() []ResUUID {
+func (obj *NoopRes) GetUIDs() []ResUID {
-	x := &NoopUUID{
+	x := &NoopUID{
-		BaseUUID: BaseUUID{name: obj.GetName(), kind: obj.Kind()},
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
-		name:     obj.Name,
+		name:    obj.Name,
 	}
-	return []ResUUID{x}
+	return []ResUID{x}
 }
 
+// GroupCmp returns whether two resources can be grouped together or not.
 func (obj *NoopRes) GroupCmp(r Res) bool {
 	_, ok := r.(*NoopRes)
 	if !ok {
@@ -128,11 +151,16 @@ func (obj *NoopRes) GroupCmp(r Res) bool {
 	return true // noop resources can always be grouped together!
 }
 
+// Compare two resources and return if they are equivalent.
 func (obj *NoopRes) Compare(res Res) bool {
 	switch res.(type) {
 	// we can only compare NoopRes to others of the same resource
 	case *NoopRes:
 		res := res.(*NoopRes)
+		// calling base Compare is unneeded for the noop res
+		//if !obj.BaseRes.Compare(res) { // call base Compare
+		//	return false
+		//}
 		if obj.Name != res.Name {
 			return false
 		}

resources/packagekit/packagekit.go

@@ -15,24 +15,29 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-// DOCS: https://www.freedesktop.org/software/PackageKit/gtk-doc/index.html
+// Package packagekit provides an interface to interact with packagekit.
-
+// See: https://www.freedesktop.org/software/PackageKit/gtk-doc/index.html for
-//package packagekit // TODO
+// more information.
-package main
+package packagekit
 
 import (
 	"fmt"
-	"github.com/godbus/dbus"
 	"log"
 	"runtime"
 	"strings"
+
+	"github.com/purpleidea/mgmt/util"
+
+	"github.com/godbus/dbus"
 )
 
+// global tweaks of verbosity and code path
 const (
 	PK_DEBUG = false
 	PARANOID = false // enable if you see any ghosts
 )
 
+// constants which might need to be tweaked or which contain special dbus strings.
 const (
 	// FIXME: if PkBufferSize is too low, install seems to drop signals
 	PkBufferSize = 1000
@@ -46,11 +51,13 @@ const (
 )
 
 var (
+	// PkArchMap contains the mapping from PackageKit arch to GOARCH.
 	// GOARCH's: 386, amd64, arm, arm64, mips64, mips64le, ppc64, ppc64le
 	PkArchMap = map[string]string{ // map of PackageKit arch to GOARCH
 		// TODO: add more values
 		// noarch
-		"noarch": "ANY", // special value "ANY"
+		"noarch": "ANY", // special value "ANY" (noarch as seen in Fedora)
+		"all":    "ANY", // special value "ANY" ('all' as seen in Debian)
 		// fedora
 		"x86_64":  "amd64",
 		"aarch64": "arm64",
@@ -97,6 +104,7 @@ const ( //static const PkEnumMatch enum_filter[]
 	PK_FILTER_ENUM_NOT_DOWNLOADED                     // "~downloaded"
 )
 
+// constants from packagekit c library.
 const ( //static const PkEnumMatch enum_transaction_flag[]
 	PK_TRANSACTION_FLAG_ENUM_NONE            uint64 = 1 << iota // "none"
 	PK_TRANSACTION_FLAG_ENUM_ONLY_TRUSTED                       // "only-trusted"
@@ -107,6 +115,7 @@ const ( //static const PkEnumMatch enum_transaction_flag[]
 	PK_TRANSACTION_FLAG_ENUM_ALLOW_DOWNGRADE                    // "allow-downgrade"
 )
 
+// constants from packagekit c library.
 const ( //typedef enum
 	PK_INFO_ENUM_UNKNOWN uint64 = 1 << iota
 	PK_INFO_ENUM_INSTALLED
@@ -137,12 +146,12 @@ const ( //typedef enum
 	PK_INFO_ENUM_LAST
 )
 
-// wrapper struct so we can pass bus connection around in the struct
+// Conn is a wrapper struct so we can pass bus connection around in the struct.
 type Conn struct {
 	conn *dbus.Conn
 }
 
-// struct that is returned by PackagesToPackageIDs in the map values
+// PkPackageIDActionData is a struct that is returned by PackagesToPackageIDs in the map values.
 type PkPackageIDActionData struct {
 	Found     bool
 	Installed bool
@@ -151,10 +160,10 @@ type PkPackageIDActionData struct {
 	Newest    bool
 }
 
-// get a new bus connection
+// NewBus returns a new bus connection.
 func NewBus() *Conn {
 	// if we share the bus with others, we will get each others messages!!
-	bus, err := SystemBusPrivateUsable() // don't share the bus connection!
+	bus, err := util.SystemBusPrivateUsable() // don't share the bus connection!
 	if err != nil {
 		return nil
 	}
@@ -163,12 +172,12 @@ func NewBus() *Conn {
 	}
 }
 
-// get the dbus connection object
+// GetBus gets the dbus connection object.
 func (bus *Conn) GetBus() *dbus.Conn {
 	return bus.conn
 }
 
-// close the dbus connection object
+// Close closes the dbus connection object.
 func (bus *Conn) Close() error {
 	return bus.conn.Close()
 }
@@ -204,7 +213,7 @@ func (bus *Conn) matchSignal(ch chan *dbus.Signal, path dbus.ObjectPath, iface s
 	return nil
 }
 
-// get a signal anytime an event happens
+// WatchChanges gets a signal anytime an event happens.
 func (bus *Conn) WatchChanges() (chan *dbus.Signal, error) {
 	ch := make(chan *dbus.Signal, PkBufferSize)
 	// NOTE: the TransactionListChanged signal fires much more frequently,
@@ -246,7 +255,7 @@ func (bus *Conn) WatchChanges() (chan *dbus.Signal, error) {
 	return ch, nil
 }
 
-// create and return a transaction path
+// CreateTransaction creates and returns a transaction path.
 func (bus *Conn) CreateTransaction() (dbus.ObjectPath, error) {
 	if PK_DEBUG {
 		log.Println("PackageKit: CreateTransaction()")
@@ -263,6 +272,7 @@ func (bus *Conn) CreateTransaction() (dbus.ObjectPath, error) {
 	return interfacePath, nil
 }
 
+// ResolvePackages runs the PackageKit Resolve method and returns the result.
 func (bus *Conn) ResolvePackages(packages []string, filter uint64) ([]string, error) {
 	packageIDs := []string{}
 	ch := make(chan *dbus.Signal, PkBufferSize)   // we need to buffer :(
@@ -326,6 +336,7 @@ loop:
 	return packageIDs, nil
 }
 
+// IsInstalledList queries a list of packages to see if they are installed.
 func (bus *Conn) IsInstalledList(packages []string) ([]bool, error) {
 	var filter uint64             // initializes at the "zero" value of 0
 	filter += PK_FILTER_ENUM_ARCH // always search in our arch
@@ -362,7 +373,7 @@ func (bus *Conn) IsInstalledList(packages []string) ([]bool, error) {
 	return r, nil
 }
 
-// is package installed ?
+// IsInstalled returns if a package is installed.
 // TODO: this could be optimized by making the resolve call directly
 func (bus *Conn) IsInstalled(pkg string) (bool, error) {
 	p, e := bus.IsInstalledList([]string{pkg})
@@ -372,7 +383,7 @@ func (bus *Conn) IsInstalled(pkg string) (bool, error) {
 	return p[0], nil
 }
 
-// install list of packages by packageID
+// InstallPackages installs a list of packages by packageID.
 func (bus *Conn) InstallPackages(packageIDs []string, transactionFlags uint64) error {
 
 	ch := make(chan *dbus.Signal, PkBufferSize)   // we need to buffer :(
@@ -414,7 +425,7 @@ loop:
 			} else {
 				return fmt.Errorf("PackageKit: Error: %v", signal.Body)
 			}
-		case _ = <-TimeAfterOrBlock(timeout):
+		case <-util.TimeAfterOrBlock(timeout):
 			if finished {
 				log.Println("PackageKit: Timeout: InstallPackages: Waiting for 'Destroy'")
 				return nil // got tired of waiting for Destroy
@@ -424,7 +435,7 @@ loop:
 	}
 }
 
-// remove list of packages
+// RemovePackages removes a list of packages by packageID.
 func (bus *Conn) RemovePackages(packageIDs []string, transactionFlags uint64) error {
 
 	var allowDeps = true                          // TODO: configurable
@@ -472,7 +483,7 @@ loop:
 	return nil
 }
 
-// update list of packages to versions that are specified
+// UpdatePackages updates a list of packages to versions that are specified.
 func (bus *Conn) UpdatePackages(packageIDs []string, transactionFlags uint64) error {
 	ch := make(chan *dbus.Signal, PkBufferSize) // we need to buffer :(
 	interfacePath, err := bus.CreateTransaction()
@@ -515,7 +526,7 @@ loop:
 	return nil
 }
 
-// get the list of files that are contained inside a list of packageids
+// GetFilesByPackageID gets the list of files that are contained inside a list of packageIDs.
 func (bus *Conn) GetFilesByPackageID(packageIDs []string) (files map[string][]string, err error) {
 	// NOTE: the maximum number of files in an RPM is 52116 in Fedora 23
 	// https://gist.github.com/purpleidea/b98e60dcd449e1ac3b8a
@@ -580,7 +591,7 @@ loop:
 	return
 }
 
-// get list of packages that are installed and which can be updated, mod filter
+// GetUpdates gets a list of packages that are installed and which can be updated, mod filter.
 func (bus *Conn) GetUpdates(filter uint64) ([]string, error) {
 	if PK_DEBUG {
 		log.Println("PackageKit: GetUpdates()")
@@ -641,9 +652,10 @@ loop:
 	return packageIDs, nil
 }
 
-// this is a helper function that *might* be generally useful outside mgmtconfig
+// PackagesToPackageIDs is a helper function that *might* be generally useful
-// packageMap input has the package names as keys and requested states as values
+// outside mgmt. The packageMap input has the package names as keys and
-// these states can be installed, uninstalled, newest or a requested version str
+// requested states as values. These states can be: installed, uninstalled,
+// newest or a requested version str.
 func (bus *Conn) PackagesToPackageIDs(packageMap map[string]string, filter uint64) (map[string]*PkPackageIDActionData, error) {
 	count := 0
 	packages := make([]string, len(packageMap))
@@ -814,7 +826,7 @@ func (bus *Conn) PackagesToPackageIDs(packageMap map[string]string, filter uint6
 	return result, nil
 }
 
-// returns a list of packageIDs which match the set of package names in packages
+// FilterPackageIDs returns a list of packageIDs which match the set of package names in packages.
 func FilterPackageIDs(m map[string]*PkPackageIDActionData, packages []string) ([]string, error) {
 	result := []string{}
 	for _, k := range packages {
@@ -828,6 +840,7 @@ func FilterPackageIDs(m map[string]*PkPackageIDActionData, packages []string) ([
 	return result, nil
 }
 
+// FilterState returns a map of whether each package queried matches the particular state.
 func FilterState(m map[string]*PkPackageIDActionData, packages []string, state string) (result map[string]bool, err error) {
 	result = make(map[string]bool)
 	pkgs := []string{} // bad pkgs that don't have a bool state
@@ -857,7 +870,7 @@ func FilterState(m map[string]*PkPackageIDActionData, packages []string, state s
 	return result, err
 }
 
-// return all packages that are in package and match the specific state
+// FilterPackageState returns all packages that are in package and match the specific state.
 func FilterPackageState(m map[string]*PkPackageIDActionData, packages []string, state string) (result []string, err error) {
 	result = []string{}
 	for _, k := range packages {
@@ -883,7 +896,7 @@ func FilterPackageState(m map[string]*PkPackageIDActionData, packages []string,
 	return result, err
 }
 
-// does flag exist inside data portion of packageID field?
+// FlagInData asks whether a flag exists inside the data portion of a packageID field?
 func FlagInData(flag, data string) bool {
 	flags := strings.Split(data, ":")
 	for _, f := range flags {
@@ -894,11 +907,12 @@ func FlagInData(flag, data string) bool {
 	return false
 }
 
-// builds the transaction method string
+// FmtTransactionMethod builds the transaction method string properly.
 func FmtTransactionMethod(method string) string {
 	return fmt.Sprintf("%s.%s", PkIfaceTransaction, method)
 }
 
+// IsMyArch determines if a PackageKit architecture matches the current os arch.
 func IsMyArch(arch string) bool {
 	goarch, ok := PkArchMap[arch]
 	if !ok {

resources/pkg.go

@@ -15,106 +15,119 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-package main
+package resources
 
 import (
-	//"packagekit" // TODO
 	"encoding/gob"
 	"errors"
 	"fmt"
 	"log"
 	"path"
 	"strings"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/global" // XXX: package mgmtmain instead?
+	"github.com/purpleidea/mgmt/resources/packagekit"
+	"github.com/purpleidea/mgmt/util"
 )
 
 func init() {
 	gob.Register(&PkgRes{})
 }
 
+// PkgRes is a package resource for packagekit.
 type PkgRes struct {
 	BaseRes          `yaml:",inline"`
 	State            string `yaml:"state"`            // state: installed, uninstalled, newest, <version>
 	AllowUntrusted   bool   `yaml:"allowuntrusted"`   // allow untrusted packages to be installed?
 	AllowNonFree     bool   `yaml:"allownonfree"`     // allow nonfree packages to be found?
 	AllowUnsupported bool   `yaml:"allowunsupported"` // allow unsupported packages to be found?
-	//bus              *Conn    // pk bus connection
+	//bus              *packagekit.Conn    // pk bus connection
 	fileList []string // FIXME: update if pkg changes
 }
 
-// helper function for creating new pkg resources that calls Init()
+// NewPkgRes is a constructor for this resource. It also calls Init() for you.
-func NewPkgRes(name, state string, allowuntrusted, allownonfree, allowunsupported bool) *PkgRes {
+func NewPkgRes(name, state string, allowuntrusted, allownonfree, allowunsupported bool) (*PkgRes, error) {
 	obj := &PkgRes{
 		BaseRes: BaseRes{
-			Name:   name,
+			Name: name,
-			events: make(chan Event),
-			vertex: nil,
 		},
 		State:            state,
 		AllowUntrusted:   allowuntrusted,
 		AllowNonFree:     allownonfree,
 		AllowUnsupported: allowunsupported,
 	}
-	obj.Init()
+	return obj, obj.Init()
-	return obj
 }
 
-func (obj *PkgRes) Init() {
+// Init runs some startup code for this resource.
+func (obj *PkgRes) Init() error {
 	obj.BaseRes.kind = "Pkg"
-	obj.BaseRes.Init() // call base init, b/c we're overriding
+	if err := obj.BaseRes.Init(); err != nil { // call base init, b/c we're overriding
+		return err
+	}
 
-	bus := NewBus()
+	bus := packagekit.NewBus()
 	if bus == nil {
-		log.Fatal("Can't connect to PackageKit bus.")
+		return fmt.Errorf("Can't connect to PackageKit bus.")
 	}
 	defer bus.Close()
 
 	result, err := obj.pkgMappingHelper(bus)
 	if err != nil {
-		// FIXME: return error?
+		return fmt.Errorf("The pkgMappingHelper failed with: %v.", err)
-		log.Fatalf("The pkgMappingHelper failed with: %v.", err)
-		return
 	}
 
 	data, ok := result[obj.Name] // lookup single package (init does just one)
 	// package doesn't exist, this is an error!
 	if !ok || !data.Found {
-		// FIXME: return error?
+		return fmt.Errorf("Can't find package named '%s'.", obj.Name)
-		log.Fatalf("Can't find package named '%s'.", obj.Name)
-		return
 	}
 
 	packageIDs := []string{data.PackageID} // just one for now
 	filesMap, err := bus.GetFilesByPackageID(packageIDs)
 	if err != nil {
-		// FIXME: return error?
+		return fmt.Errorf("Can't run GetFilesByPackageID: %v", err)
-		log.Fatalf("Can't run GetFilesByPackageID: %v", err)
-		return
 	}
 	if files, ok := filesMap[data.PackageID]; ok {
-		obj.fileList = DirifyFileList(files, false)
+		obj.fileList = util.DirifyFileList(files, false)
 	}
+	return nil
 }
 
-func (obj *PkgRes) Validate() bool {
+// Validate checks if the resource data structure was populated correctly.
-
+func (obj *PkgRes) Validate() error {
 	if obj.State == "" {
-		return false
+		return fmt.Errorf("State cannot be empty!")
 	}
 
-	return true
+	return nil
 }
 
-// use UpdatesChanged signal to watch for changes
+// Watch is the primary listener for this resource and it outputs events.
+// It uses the PackageKit UpdatesChanged signal to watch for changes.
 // TODO: https://github.com/hughsie/PackageKit/issues/109
 // TODO: https://github.com/hughsie/PackageKit/issues/110
-func (obj *PkgRes) Watch(processChan chan Event) {
+func (obj *PkgRes) Watch(processChan chan event.Event) error {
 	if obj.IsWatching() {
-		return
+		return nil
 	}
 	obj.SetWatching(true)
 	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
 
-	bus := NewBus()
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
+
+	bus := packagekit.NewBus()
 	if bus == nil {
 		log.Fatal("Can't connect to PackageKit bus.")
 	}
@@ -130,15 +143,17 @@ func (obj *PkgRes) Watch(processChan chan Event) {
 	var dirty = false
 
 	for {
-		if DEBUG {
+		if global.DEBUG {
 			log.Printf("%v: Watching...", obj.fmtNames(obj.getNames()))
 		}
 
-		obj.SetState(resStateWatching) // reset
+		obj.SetState(ResStateWatching) // reset
 		select {
 		case event := <-ch:
+			cuid.SetConverged(false)
+
 			// FIXME: ask packagekit for info on what packages changed
-			if DEBUG {
+			if global.DEBUG {
 				log.Printf("%v: Event: %v", obj.fmtNames(obj.getNames()), event.Name)
 			}
 
@@ -148,34 +163,38 @@ func (obj *PkgRes) Watch(processChan chan Event) {
 				<-ch // discard
 			}
 
-			obj.SetConvergedState(resConvergedNil)
 			send = true
 			dirty = true
 
-		case event := <-obj.events:
+		case event := <-obj.Events():
-			obj.SetConvergedState(resConvergedNil)
+			cuid.SetConverged(false)
 			if exit, send = obj.ReadEvent(&event); exit {
-				return // exit
+				return nil // exit
 			}
-			//dirty = false // these events don't invalidate state
+			dirty = false // these events don't invalidate state
 
-		case _ = <-TimeAfterOrBlock(obj.ctimeout):
+		case <-cuid.ConvergedTimer():
-			obj.SetConvergedState(resConvergedTimeout)
+			cuid.SetConverged(true) // converged!
-			obj.converged <- true
 			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
+			dirty = true
 		}
 
 		// do all our event sending all together to avoid duplicate msgs
 		if send {
+			startup = true // startup finished
 			send = false
 			// only invalid state on certain types of events
 			if dirty {
 				dirty = false
 				obj.isStateOK = false // something made state dirty
 			}
-			resp := NewResp()
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
-			processChan <- Event{eventNil, resp, "", true} // trigger process
+				return err // we exit or bubble up a NACK...
-			resp.ACKWait()                                 // wait for the ACK()
+			}
 		}
 	}
 }
@@ -217,23 +236,23 @@ func (obj *PkgRes) groupMappingHelper() map[string]string {
 	return result
 }
 
-func (obj *PkgRes) pkgMappingHelper(bus *Conn) (map[string]*PkPackageIDActionData, error) {
+func (obj *PkgRes) pkgMappingHelper(bus *packagekit.Conn) (map[string]*packagekit.PkPackageIDActionData, error) {
-	packageMap := obj.groupMappingHelper() // get the grouped values
+	packageMap := obj.groupMappingHelper()   // get the grouped values
-	packageMap[obj.Name] = obj.State       // key is pkg name, value is pkg state
+	packageMap[obj.Name] = obj.State         // key is pkg name, value is pkg state
-	var filter uint64                      // initializes at the "zero" value of 0
+	var filter uint64                        // initializes at the "zero" value of 0
-	filter += PK_FILTER_ENUM_ARCH          // always search in our arch (optional!)
+	filter += packagekit.PK_FILTER_ENUM_ARCH // always search in our arch (optional!)
 	// we're requesting latest version, or to narrow down install choices!
 	if obj.State == "newest" || obj.State == "installed" {
 		// if we add this, we'll still see older packages if installed
 		// this is an optimization, and is *optional*, this logic is
 		// handled inside of PackagesToPackageIDs now automatically!
-		filter += PK_FILTER_ENUM_NEWEST // only search for newest packages
+		filter += packagekit.PK_FILTER_ENUM_NEWEST // only search for newest packages
 	}
 	if !obj.AllowNonFree {
-		filter += PK_FILTER_ENUM_FREE
+		filter += packagekit.PK_FILTER_ENUM_FREE
 	}
 	if !obj.AllowUnsupported {
-		filter += PK_FILTER_ENUM_SUPPORTED
+		filter += packagekit.PK_FILTER_ENUM_SUPPORTED
 	}
 	result, e := bus.PackagesToPackageIDs(packageMap, filter)
 	if e != nil {
@@ -242,7 +261,9 @@ func (obj *PkgRes) pkgMappingHelper(bus *Conn) (map[string]*PkPackageIDActionDat
 	return result, nil
 }
 
-func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
+// CheckApply checks the resource state and applies the resource if the bool
+// input is true. It returns error info and if the state check passed or not.
+func (obj *PkgRes) CheckApply(apply bool) (checkok bool, err error) {
 	log.Printf("%v: CheckApply(%t)", obj.fmtNames(obj.getNames()), apply)
 
 	if obj.State == "" { // TODO: Validate() should replace this check!
@@ -253,7 +274,7 @@ func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
 		return true, nil
 	}
 
-	bus := NewBus()
+	bus := packagekit.NewBus()
 	if bus == nil {
 		return false, errors.New("Can't connect to PackageKit bus.")
 	}
@@ -266,18 +287,18 @@ func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
 
 	packageMap := obj.groupMappingHelper() // map[string]string
 	packageList := []string{obj.Name}
-	packageList = append(packageList, StrMapKeys(packageMap)...)
+	packageList = append(packageList, util.StrMapKeys(packageMap)...)
 	//stateList := []string{obj.State}
-	//stateList = append(stateList, StrMapValues(packageMap)...)
+	//stateList = append(stateList, util.StrMapValues(packageMap)...)
 
 	// TODO: at the moment, all the states are the same, but
 	// eventually we might be able to drop this constraint!
-	states, err := FilterState(result, packageList, obj.State)
+	states, err := packagekit.FilterState(result, packageList, obj.State)
 	if err != nil {
 		return false, fmt.Errorf("The FilterState method failed with: %v.", err)
 	}
 	data, _ := result[obj.Name] // if above didn't error, we won't either!
-	validState := BoolMapTrue(BoolMapValues(states))
+	validState := util.BoolMapTrue(util.BoolMapValues(states))
 
 	// obj.State == "installed" || "uninstalled" || "newest" || "4.2-1.fc23"
 	switch obj.State {
@@ -287,10 +308,12 @@ func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
 		fallthrough
 	case "newest":
 		if validState {
-			return true, nil // state is correct, exit!
+			obj.isStateOK = true // reset
+			return true, nil     // state is correct, exit!
 		}
 	default: // version string
 		if obj.State == data.Version && data.Version != "" {
+			obj.isStateOK = true // reset
 			return true, nil
 		}
 	}
@@ -302,20 +325,20 @@ func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
 
 	// apply portion
 	log.Printf("%v: Apply", obj.fmtNames(obj.getNames()))
-	readyPackages, err := FilterPackageState(result, packageList, obj.State)
+	readyPackages, err := packagekit.FilterPackageState(result, packageList, obj.State)
 	if err != nil {
 		return false, err // fail
 	}
 	// these are the packages that actually need their states applied!
-	applyPackages := StrFilterElementsInList(readyPackages, packageList)
+	applyPackages := util.StrFilterElementsInList(readyPackages, packageList)
-	packageIDs, _ := FilterPackageIDs(result, applyPackages) // would be same err as above
+	packageIDs, _ := packagekit.FilterPackageIDs(result, applyPackages) // would be same err as above
 
 	var transactionFlags uint64 // initializes at the "zero" value of 0
 	if !obj.AllowUntrusted {    // allow
-		transactionFlags += PK_TRANSACTION_FLAG_ENUM_ONLY_TRUSTED
+		transactionFlags += packagekit.PK_TRANSACTION_FLAG_ENUM_ONLY_TRUSTED
 	}
 	// apply correct state!
-	log.Printf("%v: Set: %v...", obj.fmtNames(StrListIntersection(applyPackages, obj.getNames())), obj.State)
+	log.Printf("%v: Set: %v...", obj.fmtNames(util.StrListIntersection(applyPackages, obj.getNames())), obj.State)
 	switch obj.State {
 	case "uninstalled": // run remove
 		// NOTE: packageID is different than when installed, because now
@@ -333,20 +356,21 @@ func (obj *PkgRes) CheckApply(apply bool) (stateok bool, err error) {
 	if err != nil {
 		return false, err // fail
 	}
-	log.Printf("%v: Set: %v success!", obj.fmtNames(StrListIntersection(applyPackages, obj.getNames())), obj.State)
+	log.Printf("%v: Set: %v success!", obj.fmtNames(util.StrListIntersection(applyPackages, obj.getNames())), obj.State)
-	return false, nil // success
+	obj.isStateOK = true // reset
+	return false, nil    // success
 }
 
-type PkgUUID struct {
+// PkgUID is the UID struct for PkgRes.
-	BaseUUID
+type PkgUID struct {
+	BaseUID
 	name  string // pkg name
 	state string // pkg state or "version"
 }
 
-// if and only if they are equivalent, return true
+// IFF aka if and only if they are equivalent, return true. If not, false.
-// if they are not equivalent, return false
+func (obj *PkgUID) IFF(uid ResUID) bool {
-func (obj *PkgUUID) IFF(uuid ResUUID) bool {
+	res, ok := uid.(*PkgUID)
-	res, ok := uuid.(*PkgUUID)
 	if !ok {
 		return false
 	}
@@ -354,31 +378,33 @@ func (obj *PkgUUID) IFF(uuid ResUUID) bool {
 	return obj.name == res.name
 }
 
+// PkgResAutoEdges holds the state of the auto edge generator.
 type PkgResAutoEdges struct {
 	fileList   []string
-	svcUUIDs   []ResUUID
+	svcUIDs    []ResUID
 	testIsNext bool   // safety
 	name       string // saved data from PkgRes obj
 	kind       string
 }
 
-func (obj *PkgResAutoEdges) Next() []ResUUID {
+// Next returns the next automatic edge.
+func (obj *PkgResAutoEdges) Next() []ResUID {
 	if obj.testIsNext {
 		log.Fatal("Expecting a call to Test()")
 	}
 	obj.testIsNext = true // set after all the errors paths are past
 
-	// first return any matching svcUUIDs
+	// first return any matching svcUIDs
-	if x := obj.svcUUIDs; len(x) > 0 {
+	if x := obj.svcUIDs; len(x) > 0 {
 		return x
 	}
 
-	var result []ResUUID
+	var result []ResUID
-	// return UUID's for whatever is in obj.fileList
+	// return UID's for whatever is in obj.fileList
 	for _, x := range obj.fileList {
 		var reversed = false // cheat by passing a pointer
-		result = append(result, &FileUUID{
+		result = append(result, &FileUID{
-			BaseUUID: BaseUUID{
+			BaseUID: BaseUID{
 				name:     obj.name,
 				kind:     obj.kind,
 				reversed: &reversed,
@@ -389,17 +415,18 @@ func (obj *PkgResAutoEdges) Next() []ResUUID {
 	return result
 }
 
+// Test gets results of the earlier Next() call, & returns if we should continue!
 func (obj *PkgResAutoEdges) Test(input []bool) bool {
 	if !obj.testIsNext {
 		log.Fatal("Expecting a call to Next()")
 	}
 
-	// ack the svcUUID's...
+	// ack the svcUID's...
-	if x := obj.svcUUIDs; len(x) > 0 {
+	if x := obj.svcUIDs; len(x) > 0 {
 		if y := len(x); y != len(input) {
 			log.Fatalf("Expecting %d value(s)!", y)
 		}
-		obj.svcUUIDs = []ResUUID{} // empty
+		obj.svcUIDs = []ResUID{} // empty
 		obj.testIsNext = false
 		return true
 	}
@@ -422,16 +449,16 @@ func (obj *PkgResAutoEdges) Test(input []bool) bool {
 	var dirs = make([]string, count)
 	done := []string{}
 	for i := 0; i < count; i++ {
-		dir := Dirname(obj.fileList[i]) // dirname of /foo/ should be /
+		dir := util.Dirname(obj.fileList[i]) // dirname of /foo/ should be /
 		dirs[i] = dir
 		if input[i] {
 			done = append(done, dir)
 		}
 	}
-	nodupes := StrRemoveDuplicatesInList(dirs)                // remove duplicates
+	nodupes := util.StrRemoveDuplicatesInList(dirs)                // remove duplicates
-	nodones := StrFilterElementsInList(done, nodupes)         // filter out done
+	nodones := util.StrFilterElementsInList(done, nodupes)         // filter out done
-	noempty := StrFilterElementsInList([]string{""}, nodones) // remove the "" from /
+	noempty := util.StrFilterElementsInList([]string{""}, nodones) // remove the "" from /
-	obj.fileList = RemoveCommonFilePrefixes(noempty)          // magic
+	obj.fileList = util.RemoveCommonFilePrefixes(noempty)          // magic
 
 	if len(obj.fileList) == 0 { // nothing more, don't continue
 		return false
@@ -439,46 +466,49 @@ func (obj *PkgResAutoEdges) Test(input []bool) bool {
 	return true // continue, there are more files!
 }
 
-// produce an object which generates a minimal pkg file optimization sequence
+// AutoEdges produces an object which generates a minimal pkg file optimization
+// sequence of edges.
 func (obj *PkgRes) AutoEdges() AutoEdge {
 	// in contrast with the FileRes AutoEdges() function which contains
 	// more of the mechanics, most of the AutoEdge mechanics for the PkgRes
 	// is contained in the Test() method! This design is completely okay!
 
 	// add matches for any svc resources found in pkg definition!
-	var svcUUIDs []ResUUID
+	var svcUIDs []ResUID
 	for _, x := range ReturnSvcInFileList(obj.fileList) {
 		var reversed = false
-		svcUUIDs = append(svcUUIDs, &SvcUUID{
+		svcUIDs = append(svcUIDs, &SvcUID{
-			BaseUUID: BaseUUID{
+			BaseUID: BaseUID{
 				name:     obj.GetName(),
 				kind:     obj.Kind(),
 				reversed: &reversed,
 			},
-			name: x, // the svc name itself in the SvcUUID object!
+			name: x, // the svc name itself in the SvcUID object!
 		}) // build list
 	}
 
 	return &PkgResAutoEdges{
-		fileList:   RemoveCommonFilePrefixes(obj.fileList), // clean start!
+		fileList:   util.RemoveCommonFilePrefixes(obj.fileList), // clean start!
-		svcUUIDs:   svcUUIDs,
+		svcUIDs:    svcUIDs,
 		testIsNext: false,         // start with Next() call
 		name:       obj.GetName(), // save data for PkgResAutoEdges obj
 		kind:       obj.Kind(),
 	}
 }
 
-// include all params to make a unique identification of this object
+// GetUIDs includes all params to make a unique identification of this object.
-func (obj *PkgRes) GetUUIDs() []ResUUID {
+// Most resources only return one, although some resources can return multiple.
-	x := &PkgUUID{
+func (obj *PkgRes) GetUIDs() []ResUID {
-		BaseUUID: BaseUUID{name: obj.GetName(), kind: obj.Kind()},
+	x := &PkgUID{
-		name:     obj.Name,
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
-		state:    obj.State,
+		name:    obj.Name,
+		state:   obj.State,
 	}
-	result := []ResUUID{x}
+	result := []ResUID{x}
 	return result
 }
 
+// GroupCmp returns whether two resources can be grouped together or not.
 // can these two resources be merged ?
 // (aka does this resource support doing so?)
 // will resource allow itself to be grouped _into_ this obj?
@@ -500,10 +530,15 @@ func (obj *PkgRes) GroupCmp(r Res) bool {
 	return true
 }
 
+// Compare two resources and return if they are equivalent.
 func (obj *PkgRes) Compare(res Res) bool {
 	switch res.(type) {
 	case *PkgRes:
 		res := res.(*PkgRes)
+		if !obj.BaseRes.Compare(res) { // call base Compare
+			return false
+		}
+
 		if obj.Name != res.Name {
 			return false
 		}
@@ -525,7 +560,7 @@ func (obj *PkgRes) Compare(res Res) bool {
 	return true
 }
 
-// return a list of svc names for matches like /usr/lib/systemd/system/*.service
+// ReturnSvcInFileList returns a list of svc names for matches like: `/usr/lib/systemd/system/*.service`.
 func ReturnSvcInFileList(fileList []string) []string {
 	result := []string{}
 	for _, x := range fileList {
@@ -537,7 +572,7 @@ func ReturnSvcInFileList(fileList []string) []string {
 		if !strings.HasSuffix(basename, ".service") {
 			continue
 		}
-		if s := strings.TrimSuffix(basename, ".service"); !StrInList(s, result) {
+		if s := strings.TrimSuffix(basename, ".service"); !util.StrInList(s, result) {
 			result = append(result, s)
 		}
 	}

resources/resources.go

@@ -0,0 +1,448 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+// Package resources provides the resource framework and idempotent primitives.
+package resources
+
+import (
+	"bytes"
+	"encoding/base64"
+	"encoding/gob"
+	"fmt"
+	"log"
+
+	// TODO: should each resource be a sub-package?
+	"github.com/purpleidea/mgmt/converger"
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/global"
+)
+
+//go:generate stringer -type=ResState -output=resstate_stringer.go
+
+// The ResState type represents the current activity state of each resource.
+type ResState int
+
+// Each ResState should be set properly in the relevant part of the resource.
+const (
+	ResStateNil ResState = iota
+	ResStateWatching
+	ResStateEvent // an event has happened, but we haven't poked yet
+	ResStateCheckApply
+	ResStatePoking
+)
+
+// ResUID is a unique identifier for a resource, namely it's name, and the kind ("type").
+type ResUID interface {
+	GetName() string
+	Kind() string
+	IFF(ResUID) bool
+
+	Reversed() bool // true means this resource happens before the generator
+}
+
+// The BaseUID struct is used to provide a unique resource identifier.
+type BaseUID struct {
+	name string // name and kind are the values of where this is coming from
+	kind string
+
+	reversed *bool // piggyback edge information here
+}
+
+// The AutoEdge interface is used to implement the autoedges feature.
+type AutoEdge interface {
+	Next() []ResUID   // call to get list of edges to add
+	Test([]bool) bool // call until false
+}
+
+// MetaParams is a struct will all params that apply to every resource.
+type MetaParams struct {
+	AutoEdge  bool `yaml:"autoedge"`  // metaparam, should we generate auto edges?
+	AutoGroup bool `yaml:"autogroup"` // metaparam, should we auto group?
+	Noop      bool `yaml:"noop"`
+	// NOTE: there are separate Watch and CheckApply retry and delay values,
+	// but I've decided to use the same ones for both until there's a proper
+	// reason to want to do something differently for the Watch errors.
+	Retry int16  `yaml:"retry"` // metaparam, number of times to retry on error. -1 for infinite
+	Delay uint64 `yaml:"delay"` // metaparam, number of milliseconds to wait between retries
+}
+
+// UnmarshalYAML is the custom unmarshal handler for the MetaParams struct. It
+// is primarily useful for setting the defaults.
+func (obj *MetaParams) UnmarshalYAML(unmarshal func(interface{}) error) error {
+	type rawMetaParams MetaParams           // indirection to avoid infinite recursion
+	raw := rawMetaParams(DefaultMetaParams) // convert; the defaults go here
+
+	if err := unmarshal(&raw); err != nil {
+		return err
+	}
+
+	*obj = MetaParams(raw) // restore from indirection with type conversion!
+	return nil
+}
+
+// DefaultMetaParams are the defaults to be used for undefined metaparams.
+var DefaultMetaParams = MetaParams{
+	AutoEdge:  true,
+	AutoGroup: true,
+	Noop:      false,
+	Retry:     0, // TODO: is this a good default?
+	Delay:     0, // TODO: is this a good default?
+}
+
+// The Base interface is everything that is common to all resources.
+// Everything here only needs to be implemented once, in the BaseRes.
+type Base interface {
+	GetName() string // can't be named "Name()" because of struct field
+	SetName(string)
+	SetKind(string)
+	Kind() string
+	Meta() *MetaParams
+	Events() chan event.Event
+	AssociateData(converger.Converger)
+	IsWatching() bool
+	SetWatching(bool)
+	GetState() ResState
+	SetState(ResState)
+	DoSend(chan event.Event, string) (bool, error)
+	SendEvent(event.EventName, bool, bool) bool
+	ReadEvent(*event.Event) (bool, bool) // TODO: optional here?
+	GroupCmp(Res) bool                   // TODO: is there a better name for this?
+	GroupRes(Res) error                  // group resource (arg) into self
+	IsGrouped() bool                     // am I grouped?
+	SetGrouped(bool)                     // set grouped bool
+	GetGroup() []Res                     // return everyone grouped inside me
+	SetGroup([]Res)
+}
+
+// Res is the minimum interface you need to implement to define a new resource.
+type Res interface {
+	Base // include everything from the Base interface
+	Init() error
+	//Validate() error    // TODO: this might one day be added
+	GetUIDs() []ResUID            // most resources only return one
+	Watch(chan event.Event) error // send on channel to signal process() events
+	CheckApply(apply bool) (checkOK bool, err error)
+	AutoEdges() AutoEdge
+	Compare(Res) bool
+	CollectPattern(string) // XXX: temporary until Res collection is more advanced
+}
+
+// BaseRes is the base struct that gets used in every resource.
+type BaseRes struct {
+	Name       string     `yaml:"name"`
+	MetaParams MetaParams `yaml:"meta"` // struct of all the metaparams
+	kind       string
+	events     chan event.Event
+	converger  converger.Converger // converged tracking
+	state      ResState
+	watching   bool  // is Watch() loop running ?
+	isStateOK  bool  // whether the state is okay based on events or not
+	isGrouped  bool  // am i contained within a group?
+	grouped    []Res // list of any grouped resources
+}
+
+// UIDExistsInUIDs wraps the IFF method when used with a list of UID's.
+func UIDExistsInUIDs(uid ResUID, uids []ResUID) bool {
+	for _, u := range uids {
+		if uid.IFF(u) {
+			return true
+		}
+	}
+	return false
+}
+
+// GetName returns the name of the resource.
+func (obj *BaseUID) GetName() string {
+	return obj.name
+}
+
+// Kind returns the kind of resource.
+func (obj *BaseUID) Kind() string {
+	return obj.kind
+}
+
+// IFF looks at two UID's and if and only if they are equivalent, returns true.
+// If they are not equivalent, it returns false.
+// Most resources will want to override this method, since it does the important
+// work of actually discerning if two resources are identical in function.
+func (obj *BaseUID) IFF(uid ResUID) bool {
+	res, ok := uid.(*BaseUID)
+	if !ok {
+		return false
+	}
+	return obj.name == res.name
+}
+
+// Reversed is part of the ResUID interface, and true means this resource
+// happens before the generator.
+func (obj *BaseUID) Reversed() bool {
+	if obj.reversed == nil {
+		log.Fatal("Programming error!")
+	}
+	return *obj.reversed
+}
+
+// Init initializes structures like channels if created without New constructor.
+func (obj *BaseRes) Init() error {
+	if obj.kind == "" {
+		return fmt.Errorf("Resource did not set kind!")
+	}
+	obj.events = make(chan event.Event) // unbuffered chan size to avoid stale events
+	return nil
+}
+
+// GetName is used by all the resources to Get the name.
+func (obj *BaseRes) GetName() string {
+	return obj.Name
+}
+
+// SetName is used to set the name of the resource.
+func (obj *BaseRes) SetName(name string) {
+	obj.Name = name
+}
+
+// SetKind sets the kind. This is used internally for exported resources.
+func (obj *BaseRes) SetKind(kind string) {
+	obj.kind = kind
+}
+
+// Kind returns the kind of resource this is.
+func (obj *BaseRes) Kind() string {
+	return obj.kind
+}
+
+// Meta returns the MetaParams as a reference, which we can then get/set on.
+func (obj *BaseRes) Meta() *MetaParams {
+	return &obj.MetaParams
+}
+
+// Events returns the channel of events to listen on.
+func (obj *BaseRes) Events() chan event.Event {
+	return obj.events
+}
+
+// AssociateData associates some data with the object in question.
+func (obj *BaseRes) AssociateData(converger converger.Converger) {
+	obj.converger = converger
+}
+
+// IsWatching tells us if the Watch() function is running.
+func (obj *BaseRes) IsWatching() bool {
+	return obj.watching
+}
+
+// SetWatching stores the status of if the Watch() function is running.
+func (obj *BaseRes) SetWatching(b bool) {
+	obj.watching = b
+}
+
+// GetState returns the state of the resource.
+func (obj *BaseRes) GetState() ResState {
+	return obj.state
+}
+
+// SetState sets the state of the resource.
+func (obj *BaseRes) SetState(state ResState) {
+	if global.DEBUG {
+		log.Printf("%v[%v]: State: %v -> %v", obj.Kind(), obj.GetName(), obj.GetState(), state)
+	}
+	obj.state = state
+}
+
+// DoSend sends off an event, but doesn't block the incoming event queue. It can
+// also recursively call itself when events need processing during the wait.
+// I'm not completely comfortable with this fn, but it will have to do for now.
+func (obj *BaseRes) DoSend(processChan chan event.Event, comment string) (bool, error) {
+	resp := event.NewResp()
+	processChan <- event.Event{Name: event.EventNil, Resp: resp, Msg: comment, Activity: true} // trigger process
+	e := resp.Wait()
+	return false, e // XXX: at the moment, we don't use the exit bool.
+	// XXX: this can cause a deadlock. do we need to recursively send? fix event stuff!
+	//select {
+	//case e := <-resp: // wait for the ACK()
+	//	if e != nil { // we got a NACK
+	//		return true, e // exit with error
+	//	}
+	//case event := <-obj.events:
+	//	// NOTE: this code should match the similar code below!
+	//	//cuid.SetConverged(false) // TODO: ?
+	//	if exit, send := obj.ReadEvent(&event); exit {
+	//		return true, nil // exit, without error
+	//	} else if send {
+	//		return obj.DoSend(processChan, comment) // recurse
+	//	}
+	//}
+	//return false, nil // return, no error or exit signal
+}
+
+// SendEvent pushes an event into the message queue for a particular vertex
+func (obj *BaseRes) SendEvent(ev event.EventName, sync bool, activity bool) bool {
+	// TODO: isn't this race-y ?
+	if !obj.IsWatching() { // element has already exited
+		return false // if we don't return, we'll block on the send
+	}
+	if !sync {
+		obj.events <- event.Event{Name: ev, Resp: nil, Msg: "", Activity: activity}
+		return true
+	}
+
+	resp := event.NewResp()
+	obj.events <- event.Event{Name: ev, Resp: resp, Msg: "", Activity: activity}
+	resp.ACKWait() // waits until true (nil) value
+	return true
+}
+
+// ReadEvent processes events when a select gets one, and handles the pause
+// code too! The return values specify if we should exit and poke respectively.
+func (obj *BaseRes) ReadEvent(ev *event.Event) (exit, poke bool) {
+	ev.ACK()
+	switch ev.Name {
+	case event.EventStart:
+		return false, true
+
+	case event.EventPoke:
+		return false, true
+
+	case event.EventBackPoke:
+		return false, true // forward poking in response to a back poke!
+
+	case event.EventExit:
+		return true, false
+
+	case event.EventPause:
+		// wait for next event to continue
+		select {
+		case e := <-obj.Events():
+			e.ACK()
+			if e.Name == event.EventExit {
+				return true, false
+			} else if e.Name == event.EventStart { // eventContinue
+				return false, false // don't poke on unpause!
+			} else {
+				// if we get a poke event here, it's a bug!
+				log.Fatalf("%v[%v]: Unknown event: %v, while paused!", obj.Kind(), obj.GetName(), e)
+			}
+		}
+
+	default:
+		log.Fatal("Unknown event: ", ev)
+	}
+	return true, false // required to keep the stupid go compiler happy
+}
+
+// GroupCmp compares two resources and decides if they're suitable for grouping
+// You'll probably want to override this method when implementing a resource...
+func (obj *BaseRes) GroupCmp(res Res) bool {
+	return false // base implementation assumes false, override me!
+}
+
+// GroupRes groups resource (arg) into self.
+func (obj *BaseRes) GroupRes(res Res) error {
+	if l := len(res.GetGroup()); l > 0 {
+		return fmt.Errorf("Res: %v already contains %d grouped resources!", res, l)
+	}
+	if res.IsGrouped() {
+		return fmt.Errorf("Res: %v is already grouped!", res)
+	}
+
+	obj.grouped = append(obj.grouped, res)
+	res.SetGrouped(true) // i am contained _in_ a group
+	return nil
+}
+
+// IsGrouped determines if we are grouped.
+func (obj *BaseRes) IsGrouped() bool { // am I grouped?
+	return obj.isGrouped
+}
+
+// SetGrouped sets a flag to tell if we are grouped.
+func (obj *BaseRes) SetGrouped(b bool) {
+	obj.isGrouped = b
+}
+
+// GetGroup returns everyone grouped inside me.
+func (obj *BaseRes) GetGroup() []Res { // return everyone grouped inside me
+	return obj.grouped
+}
+
+// SetGroup sets the grouped resources into me.
+func (obj *BaseRes) SetGroup(g []Res) {
+	obj.grouped = g
+}
+
+// Compare is the base compare method, which also handles the metaparams cmp
+func (obj *BaseRes) Compare(res Res) bool {
+	// TODO: should the AutoEdge values be compared?
+	if obj.Meta().AutoEdge != res.Meta().AutoEdge {
+		return false
+	}
+	if obj.Meta().AutoGroup != res.Meta().AutoGroup {
+		return false
+	}
+	if obj.Meta().Noop != res.Meta().Noop {
+		// obj is the existing res, res is the *new* resource
+		// if we go from no-noop -> noop, we can re-use the obj
+		// if we go from noop -> no-noop, we need to regenerate
+		if obj.Meta().Noop { // asymmetrical
+			return false // going from noop to no-noop!
+		}
+	}
+	if obj.Meta().Retry != res.Meta().Retry {
+		return false
+	}
+	if obj.Meta().Delay != res.Meta().Delay {
+		return false
+	}
+	return true
+}
+
+// CollectPattern is used for resource collection.
+func (obj *BaseRes) CollectPattern(pattern string) {
+	// XXX: default method is empty
+}
+
+// ResToB64 encodes a resource to a base64 encoded string (after serialization)
+func ResToB64(res Res) (string, error) {
+	b := bytes.Buffer{}
+	e := gob.NewEncoder(&b)
+	err := e.Encode(&res) // pass with &
+	if err != nil {
+		return "", fmt.Errorf("Gob failed to encode: %v", err)
+	}
+	return base64.StdEncoding.EncodeToString(b.Bytes()), nil
+}
+
+// B64ToRes decodes a resource from a base64 encoded string (after deserialization)
+func B64ToRes(str string) (Res, error) {
+	var output interface{}
+	bb, err := base64.StdEncoding.DecodeString(str)
+	if err != nil {
+		return nil, fmt.Errorf("Base64 failed to decode: %v", err)
+	}
+	b := bytes.NewBuffer(bb)
+	d := gob.NewDecoder(b)
+	err = d.Decode(&output) // pass with &
+	if err != nil {
+		return nil, fmt.Errorf("Gob failed to decode: %v", err)
+	}
+	res, ok := output.(Res)
+	if !ok {
+		return nil, fmt.Errorf("Output %v is not a Res", res)
+
+	}
+	return res, nil
+}

resources/resources_test.go

@@ -15,7 +15,7 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-package main
+package resources
 
 import (
 	"bytes"
@@ -103,3 +103,71 @@ func TestMiscEncodeDecode2(t *testing.T) {
 		t.Error("The input and output Res values do not match!")
 	}
 }
+
+func TestIFF(t *testing.T) {
+	uid := &BaseUID{name: "/tmp/unit-test"}
+	same := &BaseUID{name: "/tmp/unit-test"}
+	diff := &BaseUID{name: "/tmp/other-file"}
+
+	if !uid.IFF(same) {
+		t.Error("basic resource UIDs with the same name should satisfy each other's IFF condition.")
+	}
+
+	if uid.IFF(diff) {
+		t.Error("basic resource UIDs with different names should NOT satisfy each other's IFF condition.")
+	}
+}
+
+func TestReadEvent(t *testing.T) {
+	res := FileRes{}
+
+	shouldExit := map[eventName]bool{
+		eventStart:    false,
+		eventPoke:     false,
+		eventBackPoke: false,
+		eventExit:     true,
+	}
+	shouldPoke := map[eventName]bool{
+		eventStart:    true,
+		eventPoke:     true,
+		eventBackPoke: true,
+		eventExit:     false,
+	}
+
+	for event := range shouldExit {
+		exit, poke := res.ReadEvent(&Event{Name: event})
+		if exit != shouldExit[event] {
+			t.Errorf("resource.ReadEvent returned wrong exit flag for a %v event (%v, should be %v)",
+				event, exit, shouldExit[event])
+		}
+		if poke != shouldPoke[event] {
+			t.Errorf("resource.ReadEvent returned wrong poke flag for a %v event (%v, should be %v)",
+				event, poke, shouldPoke[event])
+		}
+	}
+
+	res.Init()
+	res.SetWatching(true)
+
+	// test result when a pause event is followed by start
+	go res.SendEvent(eventStart, false, false)
+	exit, poke := res.ReadEvent(&Event{Name: eventPause})
+	if exit {
+		t.Error("resource.ReadEvent returned wrong exit flag for a pause+start event (true, should be false)")
+	}
+	if poke {
+		t.Error("resource.ReadEvent returned wrong poke flag for a pause+start event (true, should be false)")
+	}
+
+	// test result when a pause event is followed by exit
+	go res.SendEvent(eventExit, false, false)
+	exit, poke = res.ReadEvent(&Event{Name: eventPause})
+	if !exit {
+		t.Error("resource.ReadEvent returned wrong exit flag for a pause+start event (false, should be true)")
+	}
+	if poke {
+		t.Error("resource.ReadEvent returned wrong poke flag for a pause+start event (true, should be false)")
+	}
+
+	// TODO: create a wrapper API around log, so that Fatals can be mocked and tested
+}

resources/svc.go

@@ -17,29 +17,36 @@
 
 // DOCS: https://godoc.org/github.com/coreos/go-systemd/dbus
 
-package main
+package resources
 
 import (
 	"encoding/gob"
 	"errors"
 	"fmt"
+	"log"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/util"
+
 	systemd "github.com/coreos/go-systemd/dbus" // change namespace
 	systemdUtil "github.com/coreos/go-systemd/util"
 	"github.com/godbus/dbus" // namespace collides with systemd wrapper
-	"log"
 )
 
 func init() {
 	gob.Register(&SvcRes{})
 }
 
+// SvcRes is a service resource for systemd units.
 type SvcRes struct {
 	BaseRes `yaml:",inline"`
 	State   string `yaml:"state"`   // state: running, stopped, undefined
 	Startup string `yaml:"startup"` // enabled, disabled, undefined
 }
 
-func NewSvcRes(name, state, startup string) *SvcRes {
+// NewSvcRes is a constructor for this resource. It also calls Init() for you.
+func NewSvcRes(name, state, startup string) (*SvcRes, error) {
 	obj := &SvcRes{
 		BaseRes: BaseRes{
 			Name: name,
@@ -47,49 +54,60 @@ func NewSvcRes(name, state, startup string) *SvcRes {
 		State:   state,
 		Startup: startup,
 	}
-	obj.Init()
+	return obj, obj.Init()
-	return obj
 }
 
-func (obj *SvcRes) Init() {
+// Init runs some startup code for this resource.
+func (obj *SvcRes) Init() error {
 	obj.BaseRes.kind = "Svc"
-	obj.BaseRes.Init() // call base init, b/c we're overriding
+	return obj.BaseRes.Init() // call base init, b/c we're overriding
 }
 
-func (obj *SvcRes) Validate() bool {
+// Validate checks if the resource data structure was populated correctly.
+func (obj *SvcRes) Validate() error {
 	if obj.State != "running" && obj.State != "stopped" && obj.State != "" {
-		return false
+		return fmt.Errorf("State must be either `running` or `stopped` or undefined.")
 	}
 	if obj.Startup != "enabled" && obj.Startup != "disabled" && obj.Startup != "" {
-		return false
+		return fmt.Errorf("Startup must be either `enabled` or `disabled` or undefined.")
 	}
-	return true
+	return nil
 }
 
-// Service watcher
+// Watch is the primary listener for this resource and it outputs events.
-func (obj *SvcRes) Watch(processChan chan Event) {
+func (obj *SvcRes) Watch(processChan chan event.Event) error {
 	if obj.IsWatching() {
-		return
+		return nil
 	}
 	obj.SetWatching(true)
 	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
 
 	// obj.Name: svc name
-	//vertex := obj.GetVertex()         // stored with SetVertex
 	if !systemdUtil.IsRunningSystemd() {
-		log.Fatal("Systemd is not running.")
+		return fmt.Errorf("Systemd is not running.")
 	}
 
 	conn, err := systemd.NewSystemdConnection() // needs root access
 	if err != nil {
-		log.Fatal("Failed to connect to systemd: ", err)
+		return fmt.Errorf("Failed to connect to systemd: %s", err)
 	}
 	defer conn.Close()
 
 	// if we share the bus with others, we will get each others messages!!
-	bus, err := SystemBusPrivateUsable() // don't share the bus connection!
+	bus, err := util.SystemBusPrivateUsable() // don't share the bus connection!
 	if err != nil {
-		log.Fatal("Failed to connect to bus: ", err)
+		return fmt.Errorf("Failed to connect to bus: %s", err)
 	}
 
 	// XXX: will this detect new units?
@@ -126,7 +144,7 @@ func (obj *SvcRes) Watch(processChan chan Event) {
 			var notFound = (loadstate.Value == dbus.MakeVariant("not-found"))
 			if notFound { // XXX: in the loop we'll handle changes better...
 				log.Printf("Failed to find svc: %v", svc)
-				invalid = true // XXX ?
+				invalid = true // XXX: ?
 			}
 		}
 
@@ -142,26 +160,30 @@ func (obj *SvcRes) Watch(processChan chan Event) {
 				set.Remove(svc) // no return value should ever occur
 			}
 
-			obj.SetState(resStateWatching) // reset
+			obj.SetState(ResStateWatching) // reset
 			select {
-			case _ = <-buschan: // XXX wait for new units event to unstick
+			case <-buschan: // XXX: wait for new units event to unstick
-				obj.SetConvergedState(resConvergedNil)
+				cuid.SetConverged(false)
 				// loop so that we can see the changed invalid signal
 				log.Printf("Svc[%v]->DaemonReload()", svc)
 
-			case event := <-obj.events:
+			case event := <-obj.Events():
-				obj.SetConvergedState(resConvergedNil)
+				cuid.SetConverged(false)
 				if exit, send = obj.ReadEvent(&event); exit {
-					return // exit
+					return nil // exit
 				}
 				if event.GetActivity() {
 					dirty = true
 				}
 
-			case _ = <-TimeAfterOrBlock(obj.ctimeout):
+			case <-cuid.ConvergedTimer():
-				obj.SetConvergedState(resConvergedTimeout)
+				cuid.SetConverged(true) // converged!
-				obj.converged <- true
 				continue
+
+			case <-Startup(startup):
+				cuid.SetConverged(false)
+				send = true
+				dirty = true
 			}
 		} else {
 			if !activeSet {
@@ -170,7 +192,7 @@ func (obj *SvcRes) Watch(processChan chan Event) {
 			}
 
 			log.Printf("Watching: %v", svc) // attempting to watch...
-			obj.SetState(resStateWatching)  // reset
+			obj.SetState(ResStateWatching)  // reset
 			select {
 			case event := <-subChannel:
 
@@ -178,12 +200,16 @@ func (obj *SvcRes) Watch(processChan chan Event) {
 				// NOTE: the value returned is a map for some reason...
 				if event[svc] != nil {
 					// event[svc].ActiveState is not nil
-					if event[svc].ActiveState == "active" {
+
-						log.Printf("Svc[%v]->Started()", svc)
+					switch event[svc].ActiveState {
-					} else if event[svc].ActiveState == "inactive" {
+					case "active":
-						log.Printf("Svc[%v]->Stopped!()", svc)
+						log.Printf("Svc[%v]->Started", svc)
-					} else {
+					case "inactive":
-						log.Fatal("Unknown svc state: ", event[svc].ActiveState)
+						log.Printf("Svc[%v]->Stopped", svc)
+					case "reloading":
+						log.Printf("Svc[%v]->Reloading", svc)
+					default:
+						log.Fatalf("Unknown svc state: %s", event[svc].ActiveState)
 					}
 				} else {
 					// svc stopped (and ActiveState is nil...)
@@ -193,37 +219,46 @@ func (obj *SvcRes) Watch(processChan chan Event) {
 				dirty = true
 
 			case err := <-subErrors:
-				obj.SetConvergedState(resConvergedNil) // XXX ?
+				cuid.SetConverged(false)
-				log.Printf("error: %v", err)
+				return fmt.Errorf("Unknown %s[%s] error: %v", obj.Kind(), obj.GetName(), err)
-				log.Fatal(err)
-				//vertex.events <- fmt.Sprintf("svc: %v", "error") // XXX: how should we handle errors?
 
-			case event := <-obj.events:
+			case event := <-obj.Events():
-				obj.SetConvergedState(resConvergedNil)
+				cuid.SetConverged(false)
 				if exit, send = obj.ReadEvent(&event); exit {
-					return // exit
+					return nil // exit
 				}
 				if event.GetActivity() {
 					dirty = true
 				}
+
+			case <-cuid.ConvergedTimer():
+				cuid.SetConverged(true) // converged!
+				continue
+
+			case <-Startup(startup):
+				cuid.SetConverged(false)
+				send = true
+				dirty = true
 			}
 		}
 
 		if send {
+			startup = true // startup finished
 			send = false
 			if dirty {
 				dirty = false
 				obj.isStateOK = false // something made state dirty
 			}
-			resp := NewResp()
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
-			processChan <- Event{eventNil, resp, "", true} // trigger process
+				return err // we exit or bubble up a NACK...
-			resp.ACKWait()                                 // wait for the ACK()
+			}
 		}
-
 	}
 }
 
-func (obj *SvcRes) CheckApply(apply bool) (stateok bool, err error) {
+// CheckApply checks the resource state and applies the resource if the bool
+// input is true. It returns error info and if the state check passed or not.
+func (obj *SvcRes) CheckApply(apply bool) (checkok bool, err error) {
 	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
 
 	if obj.isStateOK { // cache the state
@@ -263,7 +298,7 @@ func (obj *SvcRes) CheckApply(apply bool) (stateok bool, err error) {
 
 	var running = (activestate.Value == dbus.MakeVariant("active"))
 	var stateOK = ((obj.State == "") || (obj.State == "running" && running) || (obj.State == "stopped" && !running))
-	var startupOK = true // XXX DETECT AND SET
+	var startupOK = true // XXX: DETECT AND SET
 
 	if stateOK && startupOK {
 		return true, nil // we are in the correct state
@@ -316,32 +351,34 @@ func (obj *SvcRes) CheckApply(apply bool) (stateok bool, err error) {
 	return false, nil // success
 }
 
-type SvcUUID struct {
+// SvcUID is the UID struct for SvcRes.
-	// NOTE: there is also a name variable in the BaseUUID struct, this is
+type SvcUID struct {
-	// information about where this UUID came from, and is unrelated to the
+	// NOTE: there is also a name variable in the BaseUID struct, this is
+	// information about where this UID came from, and is unrelated to the
 	// information about the resource we're matching. That data which is
 	// used in the IFF function, is what you see in the struct fields here.
-	BaseUUID
+	BaseUID
 	name string // the svc name
 }
 
-// if and only if they are equivalent, return true
+// IFF aka if and only if they are equivalent, return true. If not, false.
-// if they are not equivalent, return false
+func (obj *SvcUID) IFF(uid ResUID) bool {
-func (obj *SvcUUID) IFF(uuid ResUUID) bool {
+	res, ok := uid.(*SvcUID)
-	res, ok := uuid.(*SvcUUID)
 	if !ok {
 		return false
 	}
 	return obj.name == res.name
 }
 
+// SvcResAutoEdges holds the state of the auto edge generator.
 type SvcResAutoEdges struct {
-	data    []ResUUID
+	data    []ResUID
 	pointer int
 	found   bool
 }
 
-func (obj *SvcResAutoEdges) Next() []ResUUID {
+// Next returns the next automatic edge.
+func (obj *SvcResAutoEdges) Next() []ResUID {
 	if obj.found {
 		log.Fatal("Shouldn't be called anymore!")
 	}
@@ -350,10 +387,10 @@ func (obj *SvcResAutoEdges) Next() []ResUUID {
 	}
 	value := obj.data[obj.pointer]
 	obj.pointer++
-	return []ResUUID{value} // we return one, even though api supports N
+	return []ResUID{value} // we return one, even though api supports N
 }
 
-// get results of the earlier Next() call, return if we should continue!
+// Test gets results of the earlier Next() call, & returns if we should continue!
 func (obj *SvcResAutoEdges) Test(input []bool) bool {
 	// if there aren't any more remaining
 	if len(obj.data) <= obj.pointer {
@@ -372,16 +409,17 @@ func (obj *SvcResAutoEdges) Test(input []bool) bool {
 	return true // keep going
 }
 
+// AutoEdges returns the AutoEdge interface. In this case the systemd units.
 func (obj *SvcRes) AutoEdges() AutoEdge {
-	var data []ResUUID
+	var data []ResUID
 	svcFiles := []string{
 		fmt.Sprintf("/etc/systemd/system/%s.service", obj.Name),     // takes precedence
 		fmt.Sprintf("/usr/lib/systemd/system/%s.service", obj.Name), // pkg default
 	}
 	for _, x := range svcFiles {
 		var reversed = true
-		data = append(data, &FileUUID{
+		data = append(data, &FileUID{
-			BaseUUID: BaseUUID{
+			BaseUID: BaseUID{
 				name:     obj.GetName(),
 				kind:     obj.Kind(),
 				reversed: &reversed,
@@ -396,15 +434,17 @@ func (obj *SvcRes) AutoEdges() AutoEdge {
 	}
 }
 
-// include all params to make a unique identification of this object
+// GetUIDs includes all params to make a unique identification of this object.
-func (obj *SvcRes) GetUUIDs() []ResUUID {
+// Most resources only return one, although some resources can return multiple.
-	x := &SvcUUID{
+func (obj *SvcRes) GetUIDs() []ResUID {
-		BaseUUID: BaseUUID{name: obj.GetName(), kind: obj.Kind()},
+	x := &SvcUID{
-		name:     obj.Name, // svc name
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
+		name:    obj.Name, // svc name
 	}
-	return []ResUUID{x}
+	return []ResUID{x}
 }
 
+// GroupCmp returns whether two resources can be grouped together or not.
 func (obj *SvcRes) GroupCmp(r Res) bool {
 	_, ok := r.(*SvcRes)
 	if !ok {
@@ -416,10 +456,15 @@ func (obj *SvcRes) GroupCmp(r Res) bool {
 	return false // not possible atm
 }
 
+// Compare two resources and return if they are equivalent.
 func (obj *SvcRes) Compare(res Res) bool {
 	switch res.(type) {
 	case *SvcRes:
 		res := res.(*SvcRes)
+		if !obj.BaseRes.Compare(res) { // call base Compare
+			return false
+		}
+
 		if obj.Name != res.Name {
 			return false
 		}

resources/timer.go

@@ -0,0 +1,166 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package resources
+
+import (
+	"encoding/gob"
+	"log"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+)
+
+func init() {
+	gob.Register(&TimerRes{})
+}
+
+// TimerRes is a timer resource for time based events.
+type TimerRes struct {
+	BaseRes  `yaml:",inline"`
+	Interval int `yaml:"interval"` // Interval : Interval between runs
+}
+
+// TimerUID is the UID struct for TimerRes.
+type TimerUID struct {
+	BaseUID
+	name string
+}
+
+// NewTimerRes is a constructor for this resource. It also calls Init() for you.
+func NewTimerRes(name string, interval int) (*TimerRes, error) {
+	obj := &TimerRes{
+		BaseRes: BaseRes{
+			Name: name,
+		},
+		Interval: interval,
+	}
+	return obj, obj.Init()
+}
+
+// Init runs some startup code for this resource.
+func (obj *TimerRes) Init() error {
+	obj.BaseRes.kind = "Timer"
+	return obj.BaseRes.Init() // call base init, b/c we're overrriding
+}
+
+// Validate the params that are passed to TimerRes
+// Currently we are getting only an interval in seconds
+// which gets validated by go compiler
+func (obj *TimerRes) Validate() error {
+	return nil
+}
+
+// Watch is the primary listener for this resource and it outputs events.
+func (obj *TimerRes) Watch(processChan chan event.Event) error {
+	if obj.IsWatching() {
+		return nil
+	}
+	obj.SetWatching(true)
+	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
+
+	// Create a time.Ticker for the given interval
+	ticker := time.NewTicker(time.Duration(obj.Interval) * time.Second)
+	defer ticker.Stop()
+
+	var send = false
+
+	for {
+		obj.SetState(ResStateWatching)
+		select {
+		case <-ticker.C: // received the timer event
+			send = true
+			log.Printf("%v[%v]: received tick", obj.Kind(), obj.GetName())
+
+		case event := <-obj.Events():
+			cuid.SetConverged(false)
+			if exit, _ := obj.ReadEvent(&event); exit {
+				return nil
+			}
+		case <-cuid.ConvergedTimer():
+			cuid.SetConverged(true)
+			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
+		}
+		if send {
+			startup = true // startup finished
+			send = false
+			obj.isStateOK = false
+			if exit, err := obj.DoSend(processChan, "timer ticked"); exit || err != nil {
+				return err // we exit or bubble up a NACK...
+			}
+		}
+	}
+}
+
+// GetUIDs includes all params to make a unique identification of this object.
+// Most resources only return one, although some resources can return multiple.
+func (obj *TimerRes) GetUIDs() []ResUID {
+	x := &TimerUID{
+		BaseUID: BaseUID{
+			name: obj.GetName(),
+			kind: obj.Kind(),
+		},
+		name: obj.Name,
+	}
+	return []ResUID{x}
+}
+
+// AutoEdges returns the AutoEdge interface. In this case no autoedges are used.
+func (obj *TimerRes) AutoEdges() AutoEdge {
+	return nil
+}
+
+// Compare two resources and return if they are equivalent.
+func (obj *TimerRes) Compare(res Res) bool {
+	switch res.(type) {
+	case *TimerRes:
+		res := res.(*TimerRes)
+		if !obj.BaseRes.Compare(res) {
+			return false
+		}
+		if obj.Name != res.Name {
+			return false
+		}
+		if obj.Interval != res.Interval {
+			return false
+		}
+	default:
+		return false
+	}
+	return true
+}
+
+// CheckApply method for Timer resource. Does nothing, returns happy!
+func (obj *TimerRes) CheckApply(apply bool) (bool, error) {
+	log.Printf("%v[%v]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
+	return true, nil // state is always okay
+}

resources/virt.go

@@ -0,0 +1,729 @@
+// Mgmt
+// Copyright (C) 2013-2016+ James Shubin and the project contributors
+// Written by James Shubin <james@shubin.ca> and the project contributors
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package resources
+
+import (
+	"encoding/gob"
+	"fmt"
+	"log"
+	"math/rand"
+	"time"
+
+	"github.com/purpleidea/mgmt/event"
+	"github.com/purpleidea/mgmt/global"
+
+	errwrap "github.com/pkg/errors"
+	"github.com/rgbkrk/libvirt-go"
+)
+
+func init() {
+	gob.Register(&VirtRes{})
+}
+
+var (
+	libvirtInitialized = false
+)
+
+// VirtRes is a libvirt resource. A transient virt resource, which has its state
+// set to `shutoff` is one which does not exist. The parallel equivalent is a
+// file resource which removes a particular path.
+type VirtRes struct {
+	BaseRes    `yaml:",inline"`
+	URI        string             `yaml:"uri"`       // connection uri, eg: qemu:///session
+	State      string             `yaml:"state"`     // running, paused, shutoff
+	Transient  bool               `yaml:"transient"` // defined (false) or undefined (true)
+	CPUs       uint16             `yaml:"cpus"`
+	Memory     uint64             `yaml:"memory"` // in KBytes
+	Boot       []string           `yaml:"boot"`   // boot order. values: fd, hd, cdrom, network
+	Disk       []diskDevice       `yaml:"disk"`
+	CDRom      []cdRomDevice      `yaml:"cdrom"`
+	Network    []networkDevice    `yaml:"network"`
+	Filesystem []filesystemDevice `yaml:"filesystem"`
+
+	conn   libvirt.VirConnection
+	absent bool // cached state
+}
+
+// NewVirtRes is a constructor for this resource. It also calls Init() for you.
+func NewVirtRes(name string, uri, state string, transient bool, cpus uint16, memory uint64) (*VirtRes, error) {
+	obj := &VirtRes{
+		BaseRes: BaseRes{
+			Name: name,
+		},
+		URI:       uri,
+		State:     state,
+		Transient: transient,
+		CPUs:      cpus,
+		Memory:    memory,
+	}
+	return obj, obj.Init()
+}
+
+// Init runs some startup code for this resource.
+func (obj *VirtRes) Init() error {
+	if !libvirtInitialized {
+		if err := libvirt.EventRegisterDefaultImpl(); err != nil {
+			return errwrap.Wrapf(err, "EventRegisterDefaultImpl failed")
+		}
+		libvirtInitialized = true
+	}
+
+	obj.absent = (obj.Transient && obj.State == "shutoff") // machine shouldn't exist
+
+	obj.BaseRes.kind = "Virt"
+	return obj.BaseRes.Init() // call base init, b/c we're overriding
+}
+
+// Validate if the params passed in are valid data.
+func (obj *VirtRes) Validate() error {
+	return nil
+}
+
+// Watch is the primary listener for this resource and it outputs events.
+func (obj *VirtRes) Watch(processChan chan event.Event) error {
+	if obj.IsWatching() {
+		return nil
+	}
+	obj.SetWatching(true)
+	defer obj.SetWatching(false)
+	cuid := obj.converger.Register()
+	defer cuid.Unregister()
+
+	var startup bool
+	Startup := func(block bool) <-chan time.Time {
+		if block {
+			return nil // blocks forever
+			//return make(chan time.Time) // blocks forever
+		}
+		return time.After(time.Duration(500) * time.Millisecond) // 1/2 the resolution of converged timeout
+	}
+
+	conn, err := libvirt.NewVirConnection(obj.URI)
+	if err != nil {
+		return fmt.Errorf("Connection to libvirt failed with: %s", err)
+	}
+
+	eventChan := make(chan int) // TODO: do we need to buffer this?
+	errorChan := make(chan error)
+	exitChan := make(chan struct{})
+	defer close(exitChan)
+
+	// run libvirt event loop
+	// TODO: *trigger* EventRunDefaultImpl to unblock so it can shut down...
+	// at the moment this isn't a major issue because it seems to unblock in
+	// bursts every 5 seconds! we can do this by writing to an event handler
+	// in the meantime, terminating the program causes it to exit anyways...
+	go func() {
+		for {
+			// TODO: can we merge this into our main for loop below?
+			select {
+			case <-exitChan:
+				log.Printf("EventRunDefaultImpl exited!")
+				return
+			default:
+			}
+			//log.Printf("EventRunDefaultImpl started!")
+			if err := libvirt.EventRunDefaultImpl(); err != nil {
+				errorChan <- errwrap.Wrapf(err, "EventRunDefaultImpl failed")
+				return
+			}
+			//log.Printf("EventRunDefaultImpl looped!")
+		}
+	}()
+
+	callback := libvirt.DomainEventCallback(
+		func(c *libvirt.VirConnection, d *libvirt.VirDomain, eventDetails interface{}, f func()) int {
+			if lifecycleEvent, ok := eventDetails.(libvirt.DomainLifecycleEvent); ok {
+				domName, _ := d.GetName()
+				if domName == obj.GetName() {
+					eventChan <- lifecycleEvent.Event
+				}
+			} else if global.DEBUG {
+				log.Printf("%s[%s]: Event details isn't DomainLifecycleEvent", obj.Kind(), obj.GetName())
+			}
+			return 0
+		},
+	)
+	callbackID := conn.DomainEventRegister(
+		libvirt.VirDomain{},
+		libvirt.VIR_DOMAIN_EVENT_ID_LIFECYCLE,
+		&callback,
+		nil,
+	)
+	defer conn.DomainEventDeregister(callbackID)
+
+	var send = false
+	var exit = false
+	var dirty = false
+
+	for {
+		select {
+		case event := <-eventChan:
+			// TODO: shouldn't we do these checks in CheckApply ?
+			switch event {
+			case libvirt.VIR_DOMAIN_EVENT_DEFINED:
+				if obj.Transient {
+					dirty = true
+					send = true
+				}
+			case libvirt.VIR_DOMAIN_EVENT_UNDEFINED:
+				if !obj.Transient {
+					dirty = true
+					send = true
+				}
+			case libvirt.VIR_DOMAIN_EVENT_STARTED:
+				fallthrough
+			case libvirt.VIR_DOMAIN_EVENT_RESUMED:
+				if obj.State != "running" {
+					dirty = true
+					send = true
+				}
+			case libvirt.VIR_DOMAIN_EVENT_SUSPENDED:
+				if obj.State != "paused" {
+					dirty = true
+					send = true
+				}
+			case libvirt.VIR_DOMAIN_EVENT_STOPPED:
+				fallthrough
+			case libvirt.VIR_DOMAIN_EVENT_SHUTDOWN:
+				if obj.State != "shutoff" {
+					dirty = true
+					send = true
+				}
+			case libvirt.VIR_DOMAIN_EVENT_PMSUSPENDED:
+				fallthrough
+			case libvirt.VIR_DOMAIN_EVENT_CRASHED:
+				dirty = true
+				send = true
+			}
+
+		case err := <-errorChan:
+			cuid.SetConverged(false)
+			return fmt.Errorf("Unknown %s[%s] libvirt error: %s", obj.Kind(), obj.GetName(), err)
+
+		case event := <-obj.Events():
+			cuid.SetConverged(false)
+			if exit, send = obj.ReadEvent(&event); exit {
+				return nil // exit
+			}
+
+		case <-cuid.ConvergedTimer():
+			cuid.SetConverged(true) // converged!
+			continue
+
+		case <-Startup(startup):
+			cuid.SetConverged(false)
+			send = true
+			dirty = true
+		}
+
+		if send {
+			startup = true // startup finished
+			send = false
+			// only invalid state on certain types of events
+			if dirty {
+				dirty = false
+				obj.isStateOK = false // something made state dirty
+			}
+			if exit, err := obj.DoSend(processChan, ""); exit || err != nil {
+				return err // we exit or bubble up a NACK...
+			}
+		}
+	}
+}
+
+// attrCheckApply performs the CheckApply functions for CPU, Memory and others.
+// This shouldn't be called when the machine is absent; it won't be found!
+func (obj *VirtRes) attrCheckApply(apply bool) (bool, error) {
+	var checkOK = true
+
+	dom, err := obj.conn.LookupDomainByName(obj.GetName())
+	if err != nil {
+		return false, errwrap.Wrapf(err, "conn.LookupDomainByName failed")
+	}
+
+	domInfo, err := dom.GetInfo()
+	if err != nil {
+		// we don't know if the state is ok
+		return false, errwrap.Wrapf(err, "domain.GetInfo failed")
+	}
+
+	// check memory
+	if domInfo.GetMemory() != obj.Memory {
+		checkOK = false
+		if !apply {
+			return false, nil
+		}
+		if err := dom.SetMemory(obj.Memory); err != nil {
+			return false, errwrap.Wrapf(err, "domain.SetMemory failed")
+		}
+		log.Printf("%s[%s]: Memory changed", obj.Kind(), obj.GetName())
+	}
+
+	// check cpus
+	if domInfo.GetNrVirtCpu() != obj.CPUs {
+		checkOK = false
+		if !apply {
+			return false, nil
+		}
+		if err := dom.SetVcpus(obj.CPUs); err != nil {
+			return false, errwrap.Wrapf(err, "domain.SetVcpus failed")
+		}
+		log.Printf("%s[%s]: CPUs changed", obj.Kind(), obj.GetName())
+	}
+
+	return checkOK, nil
+}
+
+// domainCreate creates a transient or persistent domain in the correct state. It
+// doesn't check the state before hand, as it is a simple helper function.
+func (obj *VirtRes) domainCreate() (libvirt.VirDomain, bool, error) {
+
+	if obj.Transient {
+		var flag uint32
+		var state string
+		switch obj.State {
+		case "running":
+			flag = libvirt.VIR_DOMAIN_NONE
+			state = "started"
+		case "paused":
+			flag = libvirt.VIR_DOMAIN_START_PAUSED
+			state = "paused"
+		case "shutoff":
+			// a transient, shutoff machine, means machine is absent
+			return libvirt.VirDomain{}, true, nil // returned dom is invalid
+		}
+		dom, err := obj.conn.DomainCreateXML(obj.getDomainXML(), flag)
+		if err != nil {
+			return dom, false, err // returned dom is invalid
+		}
+		log.Printf("%s[%s]: Domain transient %s", state, obj.Kind(), obj.GetName())
+		return dom, false, nil
+	}
+
+	dom, err := obj.conn.DomainDefineXML(obj.getDomainXML())
+	if err != nil {
+		return dom, false, err // returned dom is invalid
+	}
+	log.Printf("%s[%s]: Domain defined", obj.Kind(), obj.GetName())
+
+	if obj.State == "running" {
+		if err := dom.Create(); err != nil {
+			return dom, false, err
+		}
+		log.Printf("%s[%s]: Domain started", obj.Kind(), obj.GetName())
+	}
+
+	if obj.State == "paused" {
+		if err := dom.CreateWithFlags(libvirt.VIR_DOMAIN_START_PAUSED); err != nil {
+			return dom, false, err
+		}
+		log.Printf("%s[%s]: Domain created paused", obj.Kind(), obj.GetName())
+	}
+
+	return dom, false, nil
+}
+
+// CheckApply checks the resource state and applies the resource if the bool
+// input is true. It returns error info and if the state check passed or not.
+func (obj *VirtRes) CheckApply(apply bool) (bool, error) {
+	log.Printf("%s[%s]: CheckApply(%t)", obj.Kind(), obj.GetName(), apply)
+
+	if obj.isStateOK { // cache the state
+		return true, nil
+	}
+
+	var err error
+	obj.conn, err = libvirt.NewVirConnection(obj.URI)
+	if err != nil {
+		return false, fmt.Errorf("Connection to libvirt failed with: %s", err)
+	}
+
+	var checkOK = true
+
+	dom, err := obj.conn.LookupDomainByName(obj.GetName())
+	if err == nil {
+		// pass
+	} else if virErr, ok := err.(libvirt.VirError); ok && virErr.Domain == libvirt.VIR_FROM_QEMU && virErr.Code == libvirt.VIR_ERR_NO_DOMAIN {
+		// domain not found
+		if obj.absent {
+			obj.isStateOK = true
+			return true, nil
+		}
+
+		if !apply {
+			return false, nil
+		}
+
+		var c = true
+		dom, c, err = obj.domainCreate() // create the domain
+		if err != nil {
+			return false, errwrap.Wrapf(err, "domainCreate failed")
+		} else if !c {
+			checkOK = false
+		}
+
+	} else {
+		return false, errwrap.Wrapf(err, "LookupDomainByName failed")
+	}
+	defer dom.Free()
+	// domain exists
+
+	domInfo, err := dom.GetInfo()
+	if err != nil {
+		// we don't know if the state is ok
+		return false, errwrap.Wrapf(err, "domain.GetInfo failed")
+	}
+	isPersistent, err := dom.IsPersistent()
+	if err != nil {
+		// we don't know if the state is ok
+		return false, errwrap.Wrapf(err, "domain.IsPersistent failed")
+	}
+	isActive, err := dom.IsActive()
+	if err != nil {
+		// we don't know if the state is ok
+		return false, errwrap.Wrapf(err, "domain.IsActive failed")
+	}
+
+	// check for persistence
+	if isPersistent == obj.Transient { // if they're different!
+		if !apply {
+			return false, nil
+		}
+		if isPersistent {
+			if err := dom.Undefine(); err != nil {
+				return false, errwrap.Wrapf(err, "domain.Undefine failed")
+			}
+			log.Printf("%s[%s]: Domain undefined", obj.Kind(), obj.GetName())
+		} else {
+			domXML, err := dom.GetXMLDesc(libvirt.VIR_DOMAIN_XML_INACTIVE)
+			if err != nil {
+				return false, errwrap.Wrapf(err, "domain.GetXMLDesc failed")
+			}
+			if _, err = obj.conn.DomainDefineXML(domXML); err != nil {
+				return false, errwrap.Wrapf(err, "conn.DomainDefineXML failed")
+			}
+			log.Printf("%s[%s]: Domain defined", obj.Kind(), obj.GetName())
+		}
+		checkOK = false
+	}
+
+	// check for valid state
+	domState := domInfo.GetState()
+	switch obj.State {
+	case "running":
+		if domState == libvirt.VIR_DOMAIN_RUNNING {
+			break
+		}
+		if domState == libvirt.VIR_DOMAIN_BLOCKED {
+			// TODO: what should happen?
+			return false, fmt.Errorf("Domain %s is blocked!", obj.GetName())
+		}
+		if !apply {
+			return false, nil
+		}
+		if isActive { // domain must be paused ?
+			if err := dom.Resume(); err != nil {
+				return false, errwrap.Wrapf(err, "domain.Resume failed")
+			}
+			checkOK = false
+			log.Printf("%s[%s]: Domain resumed", obj.Kind(), obj.GetName())
+			break
+		}
+		if err := dom.Create(); err != nil {
+			return false, errwrap.Wrapf(err, "domain.Create failed")
+		}
+		checkOK = false
+		log.Printf("%s[%s]: Domain created", obj.Kind(), obj.GetName())
+
+	case "paused":
+		if domState == libvirt.VIR_DOMAIN_PAUSED {
+			break
+		}
+		if !apply {
+			return false, nil
+		}
+		if isActive { // domain must be running ?
+			if err := dom.Suspend(); err != nil {
+				return false, errwrap.Wrapf(err, "domain.Suspend failed")
+			}
+			checkOK = false
+			log.Printf("%s[%s]: Domain paused", obj.Kind(), obj.GetName())
+			break
+		}
+		if err := dom.CreateWithFlags(libvirt.VIR_DOMAIN_START_PAUSED); err != nil {
+			return false, errwrap.Wrapf(err, "domain.CreateWithFlags failed")
+		}
+		checkOK = false
+		log.Printf("%s[%s]: Domain created paused", obj.Kind(), obj.GetName())
+
+	case "shutoff":
+		if domState == libvirt.VIR_DOMAIN_SHUTOFF || domState == libvirt.VIR_DOMAIN_SHUTDOWN {
+			break
+		}
+		if !apply {
+			return false, nil
+		}
+
+		if err := dom.Destroy(); err != nil {
+			return false, errwrap.Wrapf(err, "domain.Destroy failed")
+		}
+		checkOK = false
+		log.Printf("%s[%s]: Domain destroyed", obj.Kind(), obj.GetName())
+	}
+
+	if !apply {
+		return false, nil
+	}
+	// remaining apply portion
+
+	// mem & cpu checks...
+	if !obj.absent {
+		if c, err := obj.attrCheckApply(apply); err != nil {
+			return false, errwrap.Wrapf(err, "attrCheckApply failed")
+		} else if !c {
+			checkOK = false
+		}
+	}
+
+	if apply || checkOK {
+		obj.isStateOK = true
+	}
+	return checkOK, nil // w00t
+}
+
+func (obj *VirtRes) getDomainXML() string {
+	var b string
+	b += "<domain type='kvm'>" // start domain
+
+	b += fmt.Sprintf("<name>%s</name>", obj.GetName())
+	b += fmt.Sprintf("<memory unit='KiB'>%d</memory>", obj.Memory)
+	b += fmt.Sprintf("<vcpu>%d</vcpu>", obj.CPUs)
+
+	b += "<os>"
+	b += "<type>hvm</type>"
+	if obj.Boot != nil {
+		for _, boot := range obj.Boot {
+			b += fmt.Sprintf("<boot dev='%s'/>", boot)
+		}
+	}
+	b += fmt.Sprintf("</os>")
+
+	b += fmt.Sprintf("<devices>") // start devices
+
+	if obj.Disk != nil {
+		for i, disk := range obj.Disk {
+			b += fmt.Sprintf(disk.GetXML(i))
+		}
+	}
+
+	if obj.CDRom != nil {
+		for i, cdrom := range obj.CDRom {
+			b += fmt.Sprintf(cdrom.GetXML(i))
+		}
+	}
+
+	if obj.Network != nil {
+		for i, net := range obj.Network {
+			b += fmt.Sprintf(net.GetXML(i))
+		}
+	}
+
+	if obj.Filesystem != nil {
+		for i, fs := range obj.Filesystem {
+			b += fmt.Sprintf(fs.GetXML(i))
+		}
+	}
+
+	b += "<serial type='pty'><target port='0'/></serial>"
+	b += "<console type='pty'><target type='serial' port='0'/></console>"
+	b += "</devices>" // end devices
+	b += "</domain>"  // end domain
+	return b
+}
+
+type virtDevice interface {
+	GetXML(idx int) string
+}
+
+type diskDevice struct {
+	Source string `yaml:"source"`
+	Type   string `yaml:"type"`
+}
+
+type cdRomDevice struct {
+	Source string `yaml:"source"`
+	Type   string `yaml:"type"`
+}
+
+type networkDevice struct {
+	Name string `yaml:"name"`
+	MAC  string `yaml:"mac"`
+}
+
+type filesystemDevice struct {
+	Access   string `yaml:"access"`
+	Source   string `yaml:"source"`
+	Target   string `yaml:"target"`
+	ReadOnly bool   `yaml:"read_only"`
+}
+
+func (d *diskDevice) GetXML(idx int) string {
+	var b string
+	b += "<disk type='file' device='disk'>"
+	b += fmt.Sprintf("<driver name='qemu' type='%s'/>", d.Type)
+	b += fmt.Sprintf("<source file='%s'/>", d.Source)
+	b += fmt.Sprintf("<target dev='vd%s' bus='virtio'/>", (string)(idx+97)) // TODO: 26, 27... should be 'aa', 'ab'...
+	b += "</disk>"
+	return b
+}
+
+func (d *cdRomDevice) GetXML(idx int) string {
+	var b string
+	b += "<disk type='file' device='cdrom'>"
+	b += fmt.Sprintf("<driver name='qemu' type='%s'/>", d.Type)
+	b += fmt.Sprintf("<source file='%s'/>", d.Source)
+	b += fmt.Sprintf("<target dev='hd%s' bus='ide'/>", (string)(idx+97)) // TODO: 26, 27... should be 'aa', 'ab'...
+	b += "<readonly/>"
+	b += "</disk>"
+	return b
+}
+
+func (d *networkDevice) GetXML(idx int) string {
+	if d.MAC == "" {
+		d.MAC = randMAC()
+	}
+	var b string
+	b += "<interface type='network'>"
+	b += fmt.Sprintf("<mac address='%s'/>", d.MAC)
+	b += fmt.Sprintf("<source network='%s'/>", d.Name)
+	b += "</interface>"
+	return b
+}
+
+func (d *filesystemDevice) GetXML(idx int) string {
+	var b string
+	b += "<filesystem" // open
+	if d.Access != "" {
+		b += fmt.Sprintf(" accessmode='%s'", d.Access)
+	}
+	b += ">" // close
+	b += fmt.Sprintf("<source dir='%s'/>", d.Source)
+	b += fmt.Sprintf("<target dir='%s'/>", d.Target)
+	if d.ReadOnly {
+		b += "<readonly/>"
+	}
+	b += "</filesystem>"
+	return b
+}
+
+// VirtUID is the UID struct for FileRes.
+type VirtUID struct {
+	BaseUID
+}
+
+// GetUIDs includes all params to make a unique identification of this object.
+// Most resources only return one, although some resources can return multiple.
+func (obj *VirtRes) GetUIDs() []ResUID {
+	x := &VirtUID{
+		BaseUID: BaseUID{name: obj.GetName(), kind: obj.Kind()},
+		// TODO: add more properties here so we can link to vm dependencies
+	}
+	return []ResUID{x}
+}
+
+// GroupCmp returns whether two resources can be grouped together or not.
+func (obj *VirtRes) GroupCmp(r Res) bool {
+	_, ok := r.(*VirtRes)
+	if !ok {
+		return false
+	}
+	return false // not possible atm
+}
+
+// AutoEdges returns the AutoEdge interface. In this case no autoedges are used.
+func (obj *VirtRes) AutoEdges() AutoEdge {
+	return nil
+}
+
+// Compare two resources and return if they are equivalent.
+func (obj *VirtRes) Compare(res Res) bool {
+	switch res.(type) {
+	case *VirtRes:
+		res := res.(*VirtRes)
+		if !obj.BaseRes.Compare(res) { // call base Compare
+			return false
+		}
+
+		if obj.Name != res.Name {
+			return false
+		}
+		if obj.URI != res.URI {
+			return false
+		}
+		if obj.State != res.State {
+			return false
+		}
+		if obj.Transient != res.Transient {
+			return false
+		}
+		if obj.CPUs != res.CPUs {
+			return false
+		}
+		// TODO: can we skip the compare of certain properties such as
+		// Memory because this object (but with different memory) can be
+		// *converted* into the new version that has more/less memory?
+		// We would need to run some sort of "old struct update", to get
+		// the new values, but that's easy to add.
+		if obj.Memory != res.Memory {
+			return false
+		}
+		// TODO:
+		//if obj.Boot != res.Boot {
+		//	return false
+		//}
+		//if obj.Disk != res.Disk {
+		//	return false
+		//}
+		//if obj.CDRom != res.CDRom {
+		//	return false
+		//}
+		//if obj.Network != res.Network {
+		//	return false
+		//}
+		//if obj.Filesystem != res.Filesystem {
+		//	return false
+		//}
+	default:
+		return false
+	}
+	return true
+}
+
+// CollectPattern applies the pattern for collection resources.
+func (obj *VirtRes) CollectPattern(string) {
+}
+
+// randMAC returns a random mac address in the libvirt range.
+func randMAC() string {
+	rand.Seed(time.Now().UnixNano())
+	return "52:54:00" +
+		fmt.Sprintf(":%x", rand.Intn(255)) +
+		fmt.Sprintf(":%x", rand.Intn(255)) +
+		fmt.Sprintf(":%x", rand.Intn(255))
+}

spec.in

@@ -12,11 +12,14 @@ Source0:	https://dl.fedoraproject.org/pub/alt/purpleidea/__PROGRAM__/SOURCES/__P
 # graphviz should really be a "suggests", since technically it's optional
 Requires: graphviz
 
-BuildRequires:	golang
+# If go_compiler is not set to 1, there is no virtual provide. Use golang instead.
+BuildRequires:  %{?go_compiler:compiler(go-compiler)}%{!?go_compiler:golang}
 BuildRequires:	golang-googlecode-tools-stringer
 BuildRequires:	git-core
 BuildRequires:	mercurial
 
+ExclusiveArch:  %{go_arches}
+
 %description
 A next generation config management prototype!
 
@@ -30,16 +33,15 @@ export GOPATH=`pwd`/vendor/
 go get github.com/coreos/etcd/client
 go get gopkg.in/yaml.v2
 go get gopkg.in/fsnotify.v1
-go get github.com/codegangsta/cli
+go get github.com/urfave/cli
 go get github.com/coreos/go-systemd/dbus
 go get github.com/coreos/go-systemd/util
 make build
 
 %install
 rm -rf %{buildroot}
-# _datadir is typically /usr/share/
+mkdir -p %{buildroot}/%{_unitdir}/
-install -d -m 0755 %{buildroot}/%{_datadir}/__PROGRAM__/
+install -pm 0644 misc/__PROGRAM__.service %{buildroot}/%{_unitdir}/
-cp -a AUTHORS COPYING COPYRIGHT DOCUMENTATION.md README.md THANKS examples/ %{buildroot}/%{_datadir}/__PROGRAM__/
 
 # install the binary
 mkdir -p %{buildroot}/%{_bindir}
@@ -55,9 +57,19 @@ install -m 0644 misc/example.conf %{buildroot}%{_sysconfdir}/__PROGRAM__/__PROGR
 
 %files
 %attr(0755, root, root) %{_sysconfdir}/profile.d/__PROGRAM__.sh
-%{_datadir}/__PROGRAM__/*
 %{_bindir}/__PROGRAM__
 %{_sysconfdir}/__PROGRAM__/*
+%{_unitdir}/__PROGRAM__.service
+
+# https://fedoraproject.org/wiki/Packaging:Guidelines?rd=Packaging/Guidelines#Documentation
+# Please add docs one per line in alpha order to avoid diff churn.
+%doc AUTHORS
+%doc COPYING
+%doc COPYRIGHT
+%doc DOCUMENTATION.md
+%doc README.md
+%doc THANKS
+%doc examples/*
 
 # this changelog is auto-generated by git log
 %changelog

test.sh

@@ -4,31 +4,45 @@ echo running test.sh
 echo "ENV:"
 env
 
+failures=''
+function run-test()
+{
+	$@ || failures=$( [ -n "$failures" ] && echo "$failures\\n$@" || echo "$@" )
+}
+
 # ensure there is no trailing whitespace or other whitespace errors
-git diff-tree --check $(git hash-object -t tree /dev/null) HEAD
+run-test git diff-tree --check $(git hash-object -t tree /dev/null) HEAD
 
 # ensure entries to authors file are sorted
 start=$(($(grep -n '^[[:space:]]*$' AUTHORS | awk -F ':' '{print $1}' | head -1) + 1))
-diff <(tail -n +$start AUTHORS | sort) <(tail -n +$start AUTHORS)
+run-test diff <(tail -n +$start AUTHORS | sort) <(tail -n +$start AUTHORS)
 
-./test/test-gofmt.sh
+run-test ./test/test-gofmt.sh
-./test/test-yamlfmt.sh
+run-test ./test/test-yamlfmt.sh
-./test/test-bashfmt.sh
+run-test ./test/test-bashfmt.sh
-./test/test-headerfmt.sh
+run-test ./test/test-headerfmt.sh
-go test
+run-test go test
-./test/test-govet.sh
+run-test ./test/test-govet.sh
 
 # do these longer tests only when running on ci
 if env | grep -q -e '^TRAVIS=true$' -e '^JENKINS_URL=' -e '^BUILD_TAG=jenkins'; then
-	go test -race
+	run-test go test -race
-	./test/test-shell.sh
+	run-test ./test/test-shell.sh
-else
-	# FIXME: this fails on travis for some reason
-	./test/test-reproducible.sh
 fi
 
+# FIXME: this now fails everywhere :(
+#run-test ./test/test-reproducible.sh
+
 # run omv tests on jenkins physical hosts only
 if env | grep -q -e '^JENKINS_URL=' -e '^BUILD_TAG=jenkins'; then
-	./test/test-omv.sh
+	run-test ./test/test-omv.sh
 fi
-./test/test-golint.sh	# test last, because this test is somewhat arbitrary
+run-test ./test/test-golint.sh	# test last, because this test is somewhat arbitrary
+
+if [[ -n "$failures" ]]; then
+	echo 'FAIL'
+	echo 'The following tests have failed:'
+	echo -e "$failures"
+	exit 1
+fi
+echo 'ALL PASSED'

test/omv/helloworld.yaml

@@ -27,12 +27,11 @@
 :ansible_extras: {}
 :cachier: false
 :vms:
-- :name: etcd
+- :name: mgmt0
   :shell:
   - iptables -F
   - cd /vagrant/mgmt/ && make path
   - cd /vagrant/mgmt/ && make deps && make build && cp mgmt ~/bin/
-  - etcd -bind-addr "`hostname --ip-address`:2379" &
   - cd && mgmt --help
 :namespace: omv
 :count: 0
@@ -45,8 +44,7 @@
 :unsafe: false
 :nested: false
 :tests:
-- omv up etcd
+- omv up mgmt0
-- vssh root@etcd -c pidof etcd
 - omv destroy
 :comment: simple hello world test case for mgmt
 :reallyrm: false

test/omv/pkg1a.yaml

@@ -32,8 +32,7 @@
   - iptables -F
   - cd /vagrant/mgmt/ && make path
   - cd /vagrant/mgmt/ && make deps && make build && cp mgmt ~/bin/
-  - etcd -bind-addr "`hostname --ip-address`:2379" &
+  - cd && mgmt run --yaml /vagrant/mgmt/examples/pkg1.yaml --converged-timeout=5
-  - cd && mgmt run --file /vagrant/mgmt/examples/pkg1.yaml --converged-timeout=5
 :namespace: omv
 :count: 0
 :username: ''

test/omv/pkg1b.yaml

@@ -33,8 +33,7 @@
   - iptables -F
   - cd /vagrant/mgmt/ && make path
   - cd /vagrant/mgmt/ && make deps && make build && cp mgmt ~/bin/
-  - etcd -bind-addr "`hostname --ip-address`:2379" &
+  - cd && mgmt run --yaml /vagrant/mgmt/examples/pkg1.yaml --converged-timeout=5
-  - cd && mgmt run --file /vagrant/mgmt/examples/pkg1.yaml --converged-timeout=5
 :namespace: omv
 :count: 0
 :username: ''

test/shell/etcd.sh

@@ -1,23 +0,0 @@
-# NOTE: boiler plate to run etcd; source with: . etcd.sh; should NOT be +x
-cleanup ()
-{
-	echo "cleanup: $1"
-	killall etcd || killall -9 etcd || true	# kill etcd
-	rm -rf /tmp/etcd/
-}
-
-trap_with_arg() {
-	func="$1"
-	shift
-	for sig in "$@"
-	do
-		trap "$func $sig" "$sig"
-	done
-}
-
-trap_with_arg cleanup INT QUIT TERM EXIT	#	ERR
-mkdir -p /tmp/etcd/
-cd /tmp/etcd/ >/dev/null	# shush the cd operation
-etcd &	# start etcd as job # 1
-sleep 1s	# let etcd startup
-cd - >/dev/null

test/shell/t1.sh

@@ -5,10 +5,11 @@
 #	* it is recommended that you run mgmt with --no-watch
 #	* it is recommended that you run mgmt --converged-timeout=<seconds>
 #	* you can run mgmt with --max-runtime=<seconds> in special scenarios
-#	* you can get a separate etcd going by sourcing etcd.sh: . etcd.sh
 
 set -o errexit
-set -o nounset
 set -o pipefail
 
 timeout --kill-after=3s 1s ./mgmt --help # hello world!
+pid=$!
+wait $pid	# get exit status
+exit $?

test/shell/t2.sh

@@ -6,14 +6,15 @@ if env | grep -q -e '^TRAVIS=true$'; then
 	exit
 fi
 
-. etcd.sh	# start etcd as job # 1
-
 # run till completion
-timeout --kill-after=15s 10s ./mgmt run --file t2.yaml --converged-timeout=5 --no-watch &
+timeout --kill-after=15s 10s ./mgmt run --yaml t2.yaml --converged-timeout=5 --no-watch --tmp-prefix &
-
+pid=$!
-. wait.sh	# wait for everything except etcd
+wait $pid	# get exit status
+e=$?
 
 test -e /tmp/mgmt/f1
 test -e /tmp/mgmt/f2
 test -e /tmp/mgmt/f3
 test ! -e /tmp/mgmt/f4
+
+exit $e

test/shell/t2.yaml

@@ -27,15 +27,15 @@ resources:
 edges:
 - name: e1
   from:
-    res: file
+    kind: file
     name: file1
   to:
-    res: file
+    kind: file
     name: file2
 - name: e2
   from:
-    res: file
+    kind: file
     name: file2
   to:
-    res: file
+    kind: file
     name: file3

test/shell/t3-a.yaml

@@ -23,6 +23,6 @@ resources:
       i am f4, exported from host A
     state: exists
 collect:
-- res: file
+- kind: file
   pattern: "/tmp/mgmt/mgmtA/"
 edges: []

test/shell/t3-b.yaml

@@ -23,6 +23,6 @@ resources:
       i am f4, exported from host B
     state: exists
 collect:
-- res: file
+- kind: file
   pattern: "/tmp/mgmt/mgmtB/"
 edges: []

test/shell/t3-c.yaml

@@ -23,6 +23,6 @@ resources:
       i am f4, exported from host C
     state: exists
 collect:
-- res: file
+- kind: file
   pattern: "/tmp/mgmt/mgmtC/"
 edges: []

test/shell/t3.sh

@@ -6,17 +6,23 @@ if env | grep -q -e '^TRAVIS=true$'; then
 	exit
 fi
 
-. etcd.sh	# start etcd as job # 1
-
 # setup
 mkdir -p "${MGMT_TMPDIR}"mgmt{A..C}
 
 # run till completion
-timeout --kill-after=15s 10s ./mgmt run --file t3-a.yaml --converged-timeout=5 --no-watch &
+timeout --kill-after=15s 10s ./mgmt run --yaml t3-a.yaml --converged-timeout=5 --no-watch --tmp-prefix &
-timeout --kill-after=15s 10s ./mgmt run --file t3-b.yaml --converged-timeout=5 --no-watch &
+pid1=$!
-timeout --kill-after=15s 10s ./mgmt run --file t3-c.yaml --converged-timeout=5 --no-watch &
+timeout --kill-after=15s 10s ./mgmt run --yaml t3-b.yaml --converged-timeout=5 --no-watch --tmp-prefix &
+pid2=$!
+timeout --kill-after=15s 10s ./mgmt run --yaml t3-c.yaml --converged-timeout=5 --no-watch --tmp-prefix &
+pid3=$!
 
-. wait.sh	# wait for everything except etcd
+wait $pid1	# get exit status
+e1=$?
+wait $pid2	# get exit status
+e2=$?
+wait $pid3	# get exit status
+e3=$?
 
 # A: collected
 test -e "${MGMT_TMPDIR}"mgmtA/f3b
@@ -71,3 +77,5 @@ test ! -e "${MGMT_TMPDIR}"mgmtC/f1a
 test ! -e "${MGMT_TMPDIR}"mgmtC/f2a
 test ! -e "${MGMT_TMPDIR}"mgmtC/f1b
 test ! -e "${MGMT_TMPDIR}"mgmtC/f2b
+
+exit $(($e1+$e2+$e3))

test/shell/t4.sh

@@ -1,10 +1,7 @@
 #!/bin/bash -e
 
-. etcd.sh	# start etcd as job # 1
-
 # should take slightly more than 25s, but fail if we take 35s)
-timeout --kill-after=35s 30s ./mgmt run --file t4.yaml --converged-timeout=5 --no-watch &
+timeout --kill-after=35s 30s ./mgmt run --yaml t4.yaml --converged-timeout=5 --no-watch --tmp-prefix &
-
+pid=$!
-#jobs	# etcd is 1
+wait $pid	# get exit status
-#wait -n 2	# wait for mgmt to exit
+exit $?
-. wait.sh	# wait for everything except etcd

test/shell/t4.yaml

@@ -56,22 +56,22 @@ resources:
 edges:
 - name: e1
   from:
-    res: exec
+    kind: exec
     name: exec1
   to:
-    res: exec
+    kind: exec
     name: exec5
 - name: e2
   from:
-    res: exec
+    kind: exec
     name: exec2
   to:
-    res: exec
+    kind: exec
     name: exec5
 - name: e3
   from:
-    res: exec
+    kind: exec
     name: exec3
   to:
-    res: exec
+    kind: exec
     name: exec5