Run Format

Source file src/go/internal/srcimporter/srcimporter.go

Documentation: go/internal/srcimporter

  // Copyright 2017 The Go Authors. All rights reserved.
  // Use of this source code is governed by a BSD-style
  // license that can be found in the LICENSE file.
  // Package srcimporter implements importing directly
  // from source files rather than installed packages.
  package srcimporter // import "go/internal/srcimporter"
  import (
  // An Importer provides the context for importing packages from source code.
  type Importer struct {
  	ctxt     *build.Context
  	fset     *token.FileSet
  	sizes    types.Sizes
  	packages map[string]*types.Package
  // NewImporter returns a new Importer for the given context, file set, and map
  // of packages. The context is used to resolve import paths to package paths,
  // and identifying the files belonging to the package. If the context provides
  // non-nil file system functions, they are used instead of the regular package
  // os functions. The file set is used to track position information of package
  // files; and imported packages are added to the packages map.
  func New(ctxt *build.Context, fset *token.FileSet, packages map[string]*types.Package) *Importer {
  	return &Importer{
  		ctxt:     ctxt,
  		fset:     fset,
  		sizes:    types.SizesFor(ctxt.Compiler, ctxt.GOARCH), // uses go/types default if GOARCH not found
  		packages: packages,
  // Importing is a sentinel taking the place in Importer.packages
  // for a package that is in the process of being imported.
  var importing types.Package
  // Import(path) is a shortcut for ImportFrom(path, ".", 0).
  func (p *Importer) Import(path string) (*types.Package, error) {
  	return p.ImportFrom(path, ".", 0) // use "." rather than "" (see issue #24441)
  // ImportFrom imports the package with the given import path resolved from the given srcDir,
  // adds the new package to the set of packages maintained by the importer, and returns the
  // package. Package path resolution and file system operations are controlled by the context
  // maintained with the importer. The import mode must be zero but is otherwise ignored.
  // Packages that are not comprised entirely of pure Go files may fail to import because the
  // type checker may not be able to determine all exported entities (e.g. due to cgo dependencies).
  func (p *Importer) ImportFrom(path, srcDir string, mode types.ImportMode) (*types.Package, error) {
  	if mode != 0 {
  		panic("non-zero import mode")
  	if abs, err := p.absPath(srcDir); err == nil { // see issue #14282
  		srcDir = abs
  	bp, err := p.ctxt.Import(path, srcDir, 0)
  	if err != nil {
  		return nil, err // err may be *build.NoGoError - return as is
  	// package unsafe is known to the type checker
  	if bp.ImportPath == "unsafe" {
  		return types.Unsafe, nil
  	// no need to re-import if the package was imported completely before
  	pkg := p.packages[bp.ImportPath]
  	if pkg != nil {
  		if pkg == &importing {
  			return nil, fmt.Errorf("import cycle through package %q", bp.ImportPath)
  		if !pkg.Complete() {
  			// Package exists but is not complete - we cannot handle this
  			// at the moment since the source importer replaces the package
  			// wholesale rather than augmenting it (see #19337 for details).
  			// Return incomplete package with error (see #16088).
  			return pkg, fmt.Errorf("reimported partially imported package %q", bp.ImportPath)
  		return pkg, nil
  	p.packages[bp.ImportPath] = &importing
  	defer func() {
  		// clean up in case of error
  		// TODO(gri) Eventually we may want to leave a (possibly empty)
  		// package in the map in all cases (and use that package to
  		// identify cycles). See also issue 16088.
  		if p.packages[bp.ImportPath] == &importing {
  			p.packages[bp.ImportPath] = nil
  	var filenames []string
  	filenames = append(filenames, bp.GoFiles...)
  	filenames = append(filenames, bp.CgoFiles...)
  	files, err := p.parseFiles(bp.Dir, filenames)
  	if err != nil {
  		return nil, err
  	// type-check package files
  	var firstHardErr error
  	conf := types.Config{
  		IgnoreFuncBodies: true,
  		FakeImportC:      true,
  		// continue type-checking after the first error
  		Error: func(err error) {
  			if firstHardErr == nil && !err.(types.Error).Soft {
  				firstHardErr = err
  		Importer: p,
  		Sizes:    p.sizes,
  	pkg, err = conf.Check(bp.ImportPath, p.fset, files, nil)
  	if err != nil {
  		// If there was a hard error it is possibly unsafe
  		// to use the package as it may not be fully populated.
  		// Do not return it (see also #20837, #20855).
  		if firstHardErr != nil {
  			pkg = nil
  			err = firstHardErr // give preference to first hard error over any soft error
  		return pkg, fmt.Errorf("type-checking package %q failed (%v)", bp.ImportPath, err)
  	if firstHardErr != nil {
  		// this can only happen if we have a bug in go/types
  		panic("package is not safe yet no error was returned")
  	p.packages[bp.ImportPath] = pkg
  	return pkg, nil
  func (p *Importer) parseFiles(dir string, filenames []string) ([]*ast.File, error) {
  	open := p.ctxt.OpenFile // possibly nil
  	files := make([]*ast.File, len(filenames))
  	errors := make([]error, len(filenames))
  	var wg sync.WaitGroup
  	for i, filename := range filenames {
  		go func(i int, filepath string) {
  			defer wg.Done()
  			if open != nil {
  				src, err := open(filepath)
  				if err != nil {
  					errors[i] = fmt.Errorf("opening package file %s failed (%v)", filepath, err)
  				files[i], errors[i] = parser.ParseFile(p.fset, filepath, src, 0)
  				src.Close() // ignore Close error - parsing may have succeeded which is all we need
  			} else {
  				// Special-case when ctxt doesn't provide a custom OpenFile and use the
  				// parser's file reading mechanism directly. This appears to be quite a
  				// bit faster than opening the file and providing an io.ReaderCloser in
  				// both cases.
  				// TODO(gri) investigate performance difference (issue #19281)
  				files[i], errors[i] = parser.ParseFile(p.fset, filepath, nil, 0)
  		}(i, p.joinPath(dir, filename))
  	// if there are errors, return the first one for deterministic results
  	for _, err := range errors {
  		if err != nil {
  			return nil, err
  	return files, nil
  // context-controlled file system operations
  func (p *Importer) absPath(path string) (string, error) {
  	// TODO(gri) This should be using p.ctxt.AbsPath which doesn't
  	// exist but probably should. See also issue #14282.
  	return filepath.Abs(path)
  func (p *Importer) isAbsPath(path string) bool {
  	if f := p.ctxt.IsAbsPath; f != nil {
  		return f(path)
  	return filepath.IsAbs(path)
  func (p *Importer) joinPath(elem ...string) string {
  	if f := p.ctxt.JoinPath; f != nil {
  		return f(elem...)
  	return filepath.Join(elem...)

View as plain text