Accelerating go-extractor by using 'go list -deps' instead of just 'go list'

Change-Id: Icc77214809a0bb8536d751f21194690d58663dc5

extractor/extractor.go

@@ -109,6 +109,25 @@ func ExtractWithFlags(buildFlags []string, patterns []string) error {
 	// root directories of packages that we want to extract
 	wantedRoots := make(map[string]bool)
 
+	log.Printf("Running go list to resolve package and module directories.")
+	// get all packages information
+	pkgsInfo, err := util.GetPkgsInfo(patterns, true, modFlags...)
+	if err != nil {
+		log.Fatalf("Error getting dependency package or module directories: %v.", err)
+	}
+	for pkgPath, pkgInfo := range pkgsInfo {
+		mdir := pkgInfo.ModDir
+		pdir := pkgInfo.PkgDir
+		// GetPkgsInfo returns the empty string if the module directory cannot be determined, e.g. if the package
+		// is not using modules. If this is the case, fall back to the package directory
+		if mdir == "" {
+			mdir = pdir
+		}
+		pkgRoots[pkgPath] = mdir
+		pkgDirs[pkgPath] = pdir
+	}
+	log.Printf("Done running go list deps: resolved %d packages.", len(pkgsInfo))
+
 	// recursively visit all packages in depth-first order;
 	// on the way down, associate each package scope with its corresponding package,
 	// and on the way up extract the package's scope
@@ -117,18 +136,6 @@ func ExtractWithFlags(buildFlags []string, patterns []string) error {
 	}, func(pkg *packages.Package) {
 		log.Printf("Processing package %s.", pkg.PkgPath)
 
-		if _, ok := pkgRoots[pkg.PkgPath]; !ok {
-			mdir := util.GetModDir(pkg.PkgPath, modFlags...)
-			pdir := util.GetPkgDir(pkg.PkgPath, modFlags...)
-			// GetModDir returns the empty string if the module directory cannot be determined, e.g. if the package
-			// is not using modules. If this is the case, fall back to the package directory
-			if mdir == "" {
-				mdir = pdir
-			}
-			pkgRoots[pkg.PkgPath] = mdir
-			pkgDirs[pkg.PkgPath] = pdir
-		}
-
 		log.Printf("Extracting types for package %s.", pkg.PkgPath)
 
 		tw, err := trap.NewWriter(pkg.PkgPath, pkg)

extractor/util/util.go

@@ -1,7 +1,9 @@
 package util
 
 import (
+	"encoding/json"
 	"errors"
+	"io"
 	"log"
 	"os"
 	"os/exec"
@@ -32,14 +34,17 @@ func Getenv(key string, aliases ...string) string {
 // runGoList is a helper function for running go list with format `format` and flags `flags` on
 // package `pkgpath`.
 func runGoList(format string, pkgpath string, flags ...string) (string, error) {
-	return runGoListWithEnv(format, pkgpath, nil, flags...)
+	return runGoListWithEnv(format, []string{pkgpath}, nil, flags...)
 }
 
-func runGoListWithEnv(format string, pkgpath string, additionalEnv []string, flags ...string) (string, error) {
+// runGoListWithEnv is a helper function for running go list with format `format` and flags `flags` on
+// pattern `patterns`, with environment variables `additionalEnv`, which is a slice of strings in the format `NAME=VALUE`.
+func runGoListWithEnv(format string, patterns []string, additionalEnv []string, flags ...string) (string, error) {
 	args := append([]string{"list", "-e", "-f", format}, flags...)
-	args = append(args, pkgpath)
+	args = append(args, patterns...)
 	cmd := exec.Command("go", args...)
 	cmd.Env = append(os.Environ(), additionalEnv...)
+	log.Printf("Running cmd: %s", cmd.String())
 	out, err := cmd.Output()
 
 	if err != nil {
@@ -54,44 +59,66 @@ func runGoListWithEnv(format string, pkgpath string, additionalEnv []string, fla
 	return strings.TrimSpace(string(out)), nil
 }
 
-// GetModDir gets the absolute directory of the module containing the package with path
-// `pkgpath`. It passes the `go list` the flags specified by `flags`.
-func GetModDir(pkgpath string, flags ...string) string {
-	// enable module mode so that we can find a module root if it exists, even if go module support is
-	// disabled by a build
-	mod, err := runGoListWithEnv("{{.Module}}", pkgpath, []string{"GO111MODULE=on"}, flags...)
-	if err != nil || mod == "<nil>" {
-		// if the command errors or modules aren't being used, return the empty string
-		return ""
-	}
-
-	modDir, err := runGoListWithEnv("{{.Module.Dir}}", pkgpath, []string{"GO111MODULE=on"}, flags...)
-	if err != nil {
-		return ""
-	}
-
-	abs, err := filepath.Abs(modDir)
-	if err != nil {
-		log.Printf("Warning: unable to make %s absolute: %s", modDir, err.Error())
-		return ""
-	}
-	return abs
+// PkgInfo holds package directory and module directory(if any) for a package
+type PkgInfo struct {
+	PkgDir string // the directory directly containing source code of this pacakge
+	ModDir string // the module directory containing this package, empty if not a module
 }
 
-// GetPkgDir gets the absolute directory containing the package with path `pkgpath`. It passes the
-// `go list` command the flags specified by `flags`.
-func GetPkgDir(pkgpath string, flags ...string) string {
-	pkgDir, err := runGoList("{{.Dir}}", pkgpath, flags...)
-	if err != nil {
-		return ""
+// GetPkgsInfo gets absolute directories of module and package root, where packages are denoted by `patterns`.
+// It passes the `go list` the flags specified by `flags`.
+// If `includingDeps`, all dependencies will also be included. see `go help list` for `-deps` flag
+func GetPkgsInfo(patterns []string, includingDeps bool, flags ...string) (map[string]*PkgInfo, error) {
+	// enable module mode so that we can find a module root if it exists, even if go module support is
+	// disabled by a build
+	if includingDeps {
+		// the flag `-deps` causes all dependencies to be retrieved
+		flags = append(flags, "-deps")
 	}
 
-	abs, err := filepath.Abs(pkgDir)
+	// using -json overrides -f format
+	output, err := runGoListWithEnv("", patterns, []string{"GO111MODULE=on"}, append(flags, "-json")...)
 	if err != nil {
-		log.Printf("Warning: unable to make %s absolute: %s", pkgDir, err.Error())
-		return ""
+		return nil, err
 	}
-	return abs
+
+	// the output of `go list -json` is a stream of json object
+	type goListPkgInfo struct {
+		ImportPath string
+		Dir        string
+		Module     *struct {
+			Dir string
+		}
+	}
+	pkgInfoMapping := make(map[string]*PkgInfo)
+	streamDecoder := json.NewDecoder(strings.NewReader(output))
+	for {
+		var pkgInfo goListPkgInfo
+		decErr := streamDecoder.Decode(&pkgInfo)
+		if decErr == io.EOF {
+			break
+		}
+		if decErr != nil {
+			log.Printf("Error decoding output of go list -json: %s", err.Error())
+			return nil, decErr
+		}
+		pkgAbsDir, err := filepath.Abs(pkgInfo.Dir)
+		if err != nil {
+			log.Printf("Unable to make package dir %s absolute: %s", pkgInfo.Dir, err.Error())
+		}
+		var modAbsDir string
+		if pkgInfo.Module != nil {
+			modAbsDir, err = filepath.Abs(pkgInfo.Module.Dir)
+			if err != nil {
+				log.Printf("Unable to make module dir %s absolute: %s", pkgInfo.Module.Dir, err.Error())
+			}
+		}
+		pkgInfoMapping[pkgInfo.ImportPath] = &PkgInfo{
+			PkgDir: pkgAbsDir,
+			ModDir: modAbsDir,
+		}
+	}
+	return pkgInfoMapping, nil
 }
 
 // DepErrors checks there are any errors resolving dependencies for `pkgpath`. It passes the `go