lang: Add module imports and more

This enables imports in mcl code, and is one of last remaining blockers
to using mgmt. Now we can start writing standalone modules, and adding
standard library functions as needed. There's still lots to do, but this
was a big missing piece. It was much harder to get right than I had
expected, but I think it's solid!

This unfortunately large commit is the result of some wild hacking I've
been doing for the past little while. It's the result of a rebase that
broke many "wip" commits that tracked my private progress, into
something that's not gratuitously messy for our git logs. Since this was
a learning and discovery process for me, I've "erased" the confusing git
history that wouldn't have helped. I'm happy to discuss the dead-ends,
and a small portion of that code was even left in for possible future
use.

This patch includes:

* A change to the cli interface:
You now specify the front-end explicitly, instead of leaving it up to
the front-end to decide when to "activate". For example, instead of:

mgmt run --lang code.mcl

we now do:

mgmt run lang --lang code.mcl

We might rename the --lang flag in the future to avoid the awkward word
repetition. Suggestions welcome, but I'm considering "input". One
side-effect of this change, is that flags which are "engine" specific
now must be specified with "run" before the front-end name. Eg:

mgmt run --tmp-prefix lang --lang code.mcl

instead of putting --tmp-prefix at the end. We also changed the GAPI
slightly, but I've patched all code that used it. This also makes things
consistent with the "deploy" command.

* The deploys are more robust and let you deploy after a run
This has been vastly improved and let's mgmt really run as a smart
engine that can handle different workloads. If you don't want to deploy
when you've started with `run` or if one comes in, you can use the
--no-watch-deploy option to block new deploys.

* The import statement exists and works!
We now have a working `import` statement. Read the docs, and try it out.
I think it's quite elegant how it fits in with `SetScope`. Have a look.
As a result, we now have some built-in functions available in modules.
This also adds the metadata.yaml entry-point for all modules. Have a
look at the examples or the tests. The bulk of the patch is to support
this.

* Improved lang input parsing code:
I re-wrote the parsing that determined what ran when we passed different
things to --lang. Deciding between running an mcl file or raw code is
now handled in a more intelligent, and re-usable way. See the inputs.go
file if you want to have a look. One casualty is that you can't stream
code from stdin *directly* to the front-end, it's encapsulated into a
deploy first. You can still use stdin though! I doubt anyone will notice
this change.

* The scope was extended to include functions and classes:
Go forth and import lovely code. All these exist in scopes now, and can
be re-used!

* Function calls actually use the scope now. Glad I got this sorted out.

* There is import cycle detection for modules!
Yes, this is another dag. I think that's #4. I guess they're useful.

* A ton of tests and new test infra was added!
This should make it much easier to add new tests that run mcl code. Have
a look at TestAstFunc1 to see how to add more of these.

As usual, I'll try to keep these commits smaller in the future!
This commit is contained in:
James Shubin
2018-11-22 16:48:10 -05:00
parent 948a3c6d08
commit 96dccca475
146 changed files with 5301 additions and 1112 deletions

View File

@@ -20,16 +20,24 @@
package lang
import (
"bytes"
"fmt"
"io/ioutil"
"os"
"sort"
"strings"
"testing"
"github.com/purpleidea/mgmt/engine"
"github.com/purpleidea/mgmt/engine/resources"
"github.com/purpleidea/mgmt/lang/funcs"
"github.com/purpleidea/mgmt/lang/interfaces"
"github.com/purpleidea/mgmt/lang/unification"
"github.com/purpleidea/mgmt/pgraph"
"github.com/purpleidea/mgmt/util"
"github.com/kylelemons/godebug/pretty"
"github.com/spf13/afero"
)
func vertexAstCmpFn(v1, v2 pgraph.Vertex) (bool, error) {
@@ -68,6 +76,8 @@ func TestAstFunc0(t *testing.T) {
"hello": &ExprStr{V: "world"},
"answer": &ExprInt{V: 42},
},
// all the built-in top-level, core functions enter here...
Functions: funcs.LookupPrefix(""),
}
type test struct { // an individual test
@@ -183,6 +193,7 @@ func TestAstFunc0(t *testing.T) {
}
`,
fail: false,
scope: scope,
graph: graph,
})
}
@@ -211,6 +222,7 @@ func TestAstFunc0(t *testing.T) {
}
`,
fail: false,
scope: scope,
graph: graph,
})
}
@@ -242,6 +254,7 @@ func TestAstFunc0(t *testing.T) {
$i = 13
`,
fail: false,
scope: scope,
graph: graph,
})
}
@@ -368,121 +381,446 @@ func TestAstFunc0(t *testing.T) {
names := []string{}
for index, tc := range testCases { // run all the tests
name, code, fail, scope, exp := tc.name, tc.code, tc.fail, tc.scope, tc.graph
if name == "" {
name = "<sub test not named>"
}
if util.StrInList(name, names) {
t.Errorf("test #%d: duplicate sub test name of: %s", index, name)
if tc.name == "" {
t.Errorf("test #%d: not named", index)
continue
}
names = append(names, name)
if util.StrInList(tc.name, names) {
t.Errorf("test #%d: duplicate sub test name of: %s", index, tc.name)
continue
}
names = append(names, tc.name)
//if index != 3 { // hack to run a subset (useful for debugging)
//if tc.name != "simple operators" {
// continue
//}
t.Logf("\n\ntest #%d (%s) ----------------\n\n", index, name)
str := strings.NewReader(code)
ast, err := LexParse(str)
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: lex/parse failed with: %+v", index, err)
t.Run(fmt.Sprintf("test #%d (%s)", index, tc.name), func(t *testing.T) {
name, code, fail, scope, exp := tc.name, tc.code, tc.fail, tc.scope, tc.graph
t.Logf("\n\ntest #%d (%s) ----------------\n\n", index, name)
str := strings.NewReader(code)
ast, err := LexParse(str)
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: lex/parse failed with: %+v", index, err)
return
}
t.Logf("test #%d: AST: %+v", index, ast)
data := &interfaces.Data{
Debug: true,
Logf: func(format string, v ...interface{}) {
t.Logf("ast: "+format, v...)
},
}
// some of this might happen *after* interpolate in SetScope or Unify...
if err := ast.Init(data); err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not init and validate AST: %+v", index, err)
return
}
iast, err := ast.Interpolate()
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: interpolate failed with: %+v", index, err)
return
}
// propagate the scope down through the AST...
err = iast.SetScope(scope)
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not set scope: %+v", index, err)
return
}
if fail && err != nil {
return // fail happened during set scope, don't run unification!
}
// apply type unification
logf := func(format string, v ...interface{}) {
t.Logf(fmt.Sprintf("test #%d", index)+": unification: "+format, v...)
}
err = unification.Unify(iast, unification.SimpleInvariantSolverLogger(logf))
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not unify types: %+v", index, err)
return
}
// maybe it will fail during graph below instead?
//if fail && err == nil {
// t.Errorf("test #%d: FAIL", index)
// t.Errorf("test #%d: unification passed, expected fail", index)
// continue
//}
if fail && err != nil {
return // fail happened during unification, don't run Graph!
}
// build the function graph
graph, err := iast.Graph()
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions failed with: %+v", index, err)
return
}
if fail && err == nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions passed, expected fail", index)
return
}
if fail { // can't process graph if it's nil
// TODO: match against expected error
t.Logf("test #%d: error: %+v", index, err)
return
}
t.Logf("test #%d: graph: %+v", index, graph)
// TODO: improve: https://github.com/purpleidea/mgmt/issues/199
if err := graph.GraphCmp(exp, vertexAstCmpFn, edgeAstCmpFn); err != nil {
t.Errorf("test #%d: FAIL\n\n", index)
t.Logf("test #%d: actual (g1): %v%s\n\n", index, graph, fullPrint(graph))
t.Logf("test #%d: expected (g2): %v%s\n\n", index, exp, fullPrint(exp))
t.Errorf("test #%d: cmp error:\n%v", index, err)
return
}
for i, v := range graph.Vertices() {
t.Logf("test #%d: vertex(%d): %+v", index, i, v)
}
for v1 := range graph.Adjacency() {
for v2, e := range graph.Adjacency()[v1] {
t.Logf("test #%d: edge(%+v): %+v -> %+v", index, e, v1, v2)
}
}
})
}
}
// TestAstFunc1 is a more advanced version which pulls code from physical dirs.
func TestAstFunc1(t *testing.T) {
const magicError = "# err: "
const magicEmpty = "# empty!"
dir, err := util.TestDirFull()
if err != nil {
t.Errorf("FAIL: could not get tests directory: %+v", err)
return
}
t.Logf("tests directory is: %s", dir)
scope := &interfaces.Scope{ // global scope
Variables: map[string]interfaces.Expr{
"purpleidea": &ExprStr{V: "hello world!"}, // james says hi
// TODO: change to a func when we can change hostname dynamically!
"hostname": &ExprStr{V: ""}, // NOTE: empty b/c not used
},
// all the built-in top-level, core functions enter here...
Functions: funcs.LookupPrefix(""),
}
type test struct { // an individual test
name string
path string // relative sub directory path inside tests dir
fail bool
//graph *pgraph.Graph
expstr string // expected graph in string format
}
testCases := []test{}
//{
// graph, _ := pgraph.NewGraph("g")
// testCases = append(testCases, test{
// name: "simple hello world",
// path: "hello0/",
// fail: false,
// expstr: graph.Sprint(),
// })
//}
// build test array automatically from reading the dir
files, err := ioutil.ReadDir(dir)
if err != nil {
t.Errorf("FAIL: could not read through tests directory: %+v", err)
return
}
for _, f := range files {
if !f.IsDir() {
continue
}
t.Logf("test #%d: AST: %+v", index, ast)
data := &interfaces.Data{
Debug: true,
Logf: func(format string, v ...interface{}) {
t.Logf("ast: "+format, v...)
},
graphFile := f.Name() + ".graph" // expected graph file
graphFileFull := dir + graphFile
info, err := os.Stat(graphFileFull)
if err != nil || info.IsDir() {
t.Errorf("FAIL: missing: %s", graphFile)
t.Errorf("(err: %+v)", err)
continue
}
// some of this might happen *after* interpolate in SetScope or Unify...
if err := ast.Init(data); err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not init and validate AST: %+v", index, err)
content, err := ioutil.ReadFile(graphFileFull)
if err != nil {
t.Errorf("FAIL: could not read graph file: %+v", err)
return
}
str := string(content) // expected graph
iast, err := ast.Interpolate()
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: interpolate failed with: %+v", index, err)
continue
// if the graph file has a magic error string, it's a failure
errStr := ""
if strings.HasPrefix(str, magicError) {
errStr = strings.TrimPrefix(str, magicError)
str = errStr
}
// propagate the scope down through the AST...
err = iast.SetScope(scope)
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not set scope: %+v", index, err)
continue
}
if fail && err != nil {
continue // fail happened during set scope, don't run unification!
}
// add automatic test case
testCases = append(testCases, test{
name: fmt.Sprintf("dir: %s", f.Name()),
path: f.Name() + "/",
fail: errStr != "",
expstr: str,
})
//t.Logf("adding: %s", f.Name() + "/")
}
// apply type unification
logf := func(format string, v ...interface{}) {
t.Logf(fmt.Sprintf("test #%d", index)+": unification: "+format, v...)
}
err = unification.Unify(iast, unification.SimpleInvariantSolverLogger(logf))
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not unify types: %+v", index, err)
names := []string{}
for index, tc := range testCases { // run all the tests
if tc.name == "" {
t.Errorf("test #%d: not named", index)
continue
}
// maybe it will fail during graph below instead?
//if fail && err == nil {
// t.Errorf("test #%d: FAIL", index)
// t.Errorf("test #%d: unification passed, expected fail", index)
if util.StrInList(tc.name, names) {
t.Errorf("test #%d: duplicate sub test name of: %s", index, tc.name)
continue
}
names = append(names, tc.name)
//if index != 3 { // hack to run a subset (useful for debugging)
//if tc.name != "simple operators" {
// continue
//}
if fail && err != nil {
continue // fail happened during unification, don't run Graph!
}
// build the function graph
graph, err := iast.Graph()
t.Run(fmt.Sprintf("test #%d (%s)", index, tc.name), func(t *testing.T) {
name, path, fail, expstr := tc.name, tc.path, tc.fail, strings.Trim(tc.expstr, "\n")
src := dir + path // location of the test
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions failed with: %+v", index, err)
continue
}
if fail && err == nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions passed, expected fail", index)
continue
}
t.Logf("\n\ntest #%d (%s) ----------------\npath: %s\n\n", index, name, src)
if fail { // can't process graph if it's nil
// TODO: match against expected error
t.Logf("test #%d: error: %+v", index, err)
continue
}
t.Logf("test #%d: graph: %+v", index, graph)
// TODO: improve: https://github.com/purpleidea/mgmt/issues/199
if err := graph.GraphCmp(exp, vertexAstCmpFn, edgeAstCmpFn); err != nil {
t.Errorf("test #%d: FAIL\n\n", index)
t.Logf("test #%d: actual (g1): %v%s\n\n", index, graph, fullPrint(graph))
t.Logf("test #%d: expected (g2): %v%s\n\n", index, exp, fullPrint(exp))
t.Errorf("test #%d: cmp error:\n%v", index, err)
continue
}
for i, v := range graph.Vertices() {
t.Logf("test #%d: vertex(%d): %+v", index, i, v)
}
for v1 := range graph.Adjacency() {
for v2, e := range graph.Adjacency()[v1] {
t.Logf("test #%d: edge(%+v): %+v -> %+v", index, e, v1, v2)
logf := func(format string, v ...interface{}) {
t.Logf(fmt.Sprintf("test #%d", index)+": "+format, v...)
}
}
mmFs := afero.NewMemMapFs()
afs := &afero.Afero{Fs: mmFs} // wrap so that we're implementing ioutil
fs := &util.Fs{Afero: afs}
// use this variant, so that we don't copy the dir name
// this is the equivalent to running `rsync -a src/ /`
if err := util.CopyDiskContentsToFs(fs, src, "/", false); err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: CopyDiskContentsToFs failed: %+v", index, err)
return
}
// this shows us what we pulled in from the test dir:
tree0, err := util.FsTree(fs, "/")
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: FsTree failed: %+v", index, err)
return
}
logf("tree:\n%s", tree0)
input := "/"
logf("input: %s", input)
output, err := parseInput(input, fs) // raw code can be passed in
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: parseInput failed: %+v", index, err)
return
}
for _, fn := range output.Workers {
if err := fn(fs); err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: worker execution failed: %+v", index, err)
return
}
}
tree, err := util.FsTree(fs, "/")
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: FsTree failed: %+v", index, err)
return
}
logf("tree:\n%s", tree)
logf("main:\n%s", output.Main) // debug
reader := bytes.NewReader(output.Main)
ast, err := LexParse(reader)
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: lex/parse failed with: %+v", index, err)
return
}
if fail && err != nil {
// TODO: %+v instead?
s := fmt.Sprintf("%s", err) // convert to string
if s != expstr {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: expected different error", index)
t.Logf("test #%d: err: %s", index, s)
t.Logf("test #%d: exp: %s", index, expstr)
}
return // fail happened during set scope, don't run unification!
}
t.Logf("test #%d: AST: %+v", index, ast)
importGraph, err := pgraph.NewGraph("importGraph")
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not create graph: %+v", index, err)
return
}
importVertex := &pgraph.SelfVertex{
Name: "", // first node is the empty string
Graph: importGraph, // store a reference to ourself
}
importGraph.AddVertex(importVertex)
data := &interfaces.Data{
Fs: fs,
Base: output.Base, // base dir (absolute path) the metadata file is in
Files: output.Files, // no really needed here afaict
Imports: importVertex,
Metadata: output.Metadata,
Modules: "/" + interfaces.ModuleDirectory, // not really needed here afaict
Debug: true,
Logf: func(format string, v ...interface{}) {
logf("ast: "+format, v...)
},
}
// some of this might happen *after* interpolate in SetScope or Unify...
if err := ast.Init(data); err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not init and validate AST: %+v", index, err)
return
}
iast, err := ast.Interpolate()
if err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: interpolate failed with: %+v", index, err)
return
}
// propagate the scope down through the AST...
err = iast.SetScope(scope)
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not set scope: %+v", index, err)
return
}
if fail && err != nil {
// TODO: %+v instead?
s := fmt.Sprintf("%s", err) // convert to string
if s != expstr {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: expected different error", index)
t.Logf("test #%d: err: %s", index, s)
t.Logf("test #%d: exp: %s", index, expstr)
}
return // fail happened during set scope, don't run unification!
}
// apply type unification
xlogf := func(format string, v ...interface{}) {
logf("unification: "+format, v...)
}
err = unification.Unify(iast, unification.SimpleInvariantSolverLogger(xlogf))
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: could not unify types: %+v", index, err)
return
}
// maybe it will fail during graph below instead?
//if fail && err == nil {
// t.Errorf("test #%d: FAIL", index)
// t.Errorf("test #%d: unification passed, expected fail", index)
// continue
//}
if fail && err != nil {
// TODO: %+v instead?
s := fmt.Sprintf("%s", err) // convert to string
if s != expstr {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: expected different error", index)
t.Logf("test #%d: err: %s", index, s)
t.Logf("test #%d: exp: %s", index, expstr)
}
return // fail happened during unification, don't run Graph!
}
// build the function graph
graph, err := iast.Graph()
if !fail && err != nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions failed with: %+v", index, err)
return
}
if fail && err == nil {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: functions passed, expected fail", index)
return
}
if fail { // can't process graph if it's nil
// TODO: %+v instead?
s := fmt.Sprintf("%s", err) // convert to string
if s != expstr {
t.Errorf("test #%d: FAIL", index)
t.Errorf("test #%d: expected different error", index)
t.Logf("test #%d: err: %s", index, s)
t.Logf("test #%d: exp: %s", index, expstr)
}
return
}
t.Logf("test #%d: graph: %+v", index, graph)
str := strings.Trim(graph.Sprint(), "\n") // text format of graph
if expstr == magicEmpty {
expstr = ""
}
// XXX: something isn't consistent, and I can't figure
// out what, so workaround this by sorting these :(
sortHack := func(x string) string {
l := strings.Split(x, "\n")
sort.Strings(l)
return strings.Join(l, "\n")
}
str = sortHack(str)
expstr = sortHack(expstr)
if expstr != str {
t.Errorf("test #%d: FAIL\n\n", index)
t.Logf("test #%d: actual (g1):\n%s\n\n", index, str)
t.Logf("test #%d: expected (g2):\n%s\n\n", index, expstr)
diff := pretty.Compare(str, expstr)
if diff != "" { // bonus
t.Logf("test #%d: diff:\n%s", index, diff)
}
return
}
for i, v := range graph.Vertices() {
t.Logf("test #%d: vertex(%d): %+v", index, i, v)
}
for v1 := range graph.Adjacency() {
for v2, e := range graph.Adjacency()[v1] {
t.Logf("test #%d: edge(%+v): %+v -> %+v", index, e, v1, v2)
}
}
})
}
}