build.go

// Copyright 2011 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 build

import (
	"bytes"
	"errors"
	"fmt"
	"go/ast"
	"go/doc"
	"go/parser"
	"go/token"
	"internal/goroot"
	"io"
	"io/ioutil"
	"log"
	"os"
	"os/exec"
	pathpkg "path"
	"path/filepath"
	"runtime"
	"sort"
	"strconv"
	"strings"
	"unicode"
	"unicode/utf8"
)

// A Context specifies the supporting context for a build.
type Context struct {
	GOARCH      string // target architecture
	GOOS        string // target operating system
	GOROOT      string // Go root
	GOPATH      string // Go path
	CgoEnabled  bool   // whether cgo files are included
	UseAllFiles bool   // use files regardless of +build lines, file names
	Compiler    string // compiler to assume when computing target paths

	// The build and release tags specify build constraints
	// that should be considered satisfied when processing +build lines.
	// Clients creating a new context may customize BuildTags, which
	// defaults to empty, but it is usually an error to customize ReleaseTags,
	// which defaults to the list of Go releases the current release is compatible with.
	// BuildTags is not set for the Default build Context.
	// In addition to the BuildTags and ReleaseTags, build constraints
	// consider the values of GOARCH and GOOS as satisfied tags.
	// The last element in ReleaseTags is assumed to be the current release.
	BuildTags   []string
	ReleaseTags []string

	// The install suffix specifies a suffix to use in the name of the installation
	// directory. By default it is empty, but custom builds that need to keep
	// their outputs separate can set InstallSuffix to do so. For example, when
	// using the race detector, the go command uses InstallSuffix = "race", so
	// that on a Linux/386 system, packages are written to a directory named
	// "linux_386_race" instead of the usual "linux_386".
	InstallSuffix string

	// By default, Import uses the operating system's file system calls
	// to read directories and files. To read from other sources,
	// callers can set the following functions. They all have default
	// behaviors that use the local file system, so clients need only set
	// the functions whose behaviors they wish to change.

	// JoinPath joins the sequence of path fragments into a single path.
	// If JoinPath is nil, Import uses filepath.Join.
	JoinPath func(elem ...string) string

	// SplitPathList splits the path list into a slice of individual paths.
	// If SplitPathList is nil, Import uses filepath.SplitList.
	SplitPathList func(list string) []string

	// IsAbsPath reports whether path is an absolute path.
	// If IsAbsPath is nil, Import uses filepath.IsAbs.
	IsAbsPath func(path string) bool

	// IsDir reports whether the path names a directory.
	// If IsDir is nil, Import calls os.Stat and uses the result's IsDir method.
	IsDir func(path string) bool

	// HasSubdir reports whether dir is lexically a subdirectory of
	// root, perhaps multiple levels below. It does not try to check
	// whether dir exists.
	// If so, HasSubdir sets rel to a slash-separated path that
	// can be joined to root to produce a path equivalent to dir.
	// If HasSubdir is nil, Import uses an implementation built on
	// filepath.EvalSymlinks.
	HasSubdir func(root, dir string) (rel string, ok bool)

	// ReadDir returns a slice of os.FileInfo, sorted by Name,
	// describing the content of the named directory.
	// If ReadDir is nil, Import uses ioutil.ReadDir.
	ReadDir func(dir string) ([]os.FileInfo, error)

	// OpenFile opens a file (not a directory) for reading.
	// If OpenFile is nil, Import uses os.Open.
	OpenFile func(path string) (io.ReadCloser, error)
}

// joinPath calls ctxt.JoinPath (if not nil) or else filepath.Join.
func (ctxt *Context) joinPath(elem ...string) string {
	if f := ctxt.JoinPath; f != nil {
		return f(elem...)
	}
	return filepath.Join(elem...)
}

// splitPathList calls ctxt.SplitPathList (if not nil) or else filepath.SplitList.
func (ctxt *Context) splitPathList(s string) []string {
	if f := ctxt.SplitPathList; f != nil {
		return f(s)
	}
	return filepath.SplitList(s)
}

// isAbsPath calls ctxt.IsAbsPath (if not nil) or else filepath.IsAbs.
func (ctxt *Context) isAbsPath(path string) bool {
	if f := ctxt.IsAbsPath; f != nil {
		return f(path)
	}
	return filepath.IsAbs(path)
}

// isDir calls ctxt.IsDir (if not nil) or else uses os.Stat.
func (ctxt *Context) isDir(path string) bool {
	if f := ctxt.IsDir; f != nil {
		return f(path)
	}
	fi, err := os.Stat(path)
	return err == nil && fi.IsDir()
}

// hasSubdir calls ctxt.HasSubdir (if not nil) or else uses
// the local file system to answer the question.
func (ctxt *Context) hasSubdir(root, dir string) (rel string, ok bool) {
	if f := ctxt.HasSubdir; f != nil {
		return f(root, dir)
	}

	// Try using paths we received.
	if rel, ok = hasSubdir(root, dir); ok {
		return
	}

	// Try expanding symlinks and comparing
	// expanded against unexpanded and
	// expanded against expanded.
	rootSym, _ := filepath.EvalSymlinks(root)
	dirSym, _ := filepath.EvalSymlinks(dir)

	if rel, ok = hasSubdir(rootSym, dir); ok {
		return
	}
	if rel, ok = hasSubdir(root, dirSym); ok {
		return
	}
	return hasSubdir(rootSym, dirSym)
}

// hasSubdir reports if dir is within root by performing lexical analysis only.
func hasSubdir(root, dir string) (rel string, ok bool) {
	const sep = string(filepath.Separator)
	root = filepath.Clean(root)
	if !strings.HasSuffix(root, sep) {
		root += sep
	}
	dir = filepath.Clean(dir)
	if !strings.HasPrefix(dir, root) {
		return "", false
	}
	return filepath.ToSlash(dir[len(root):]), true
}

// readDir calls ctxt.ReadDir (if not nil) or else ioutil.ReadDir.
func (ctxt *Context) readDir(path string) ([]os.FileInfo, error) {
	if f := ctxt.ReadDir; f != nil {
		return f(path)
	}
	return ioutil.ReadDir(path)
}

// openFile calls ctxt.OpenFile (if not nil) or else os.Open.
func (ctxt *Context) openFile(path string) (io.ReadCloser, error) {
	if fn := ctxt.OpenFile; fn != nil {
		return fn(path)
	}

	f, err := os.Open(path)
	if err != nil {
		return nil, err // nil interface
	}
	return f, nil
}

// isFile determines whether path is a file by trying to open it.
// It reuses openFile instead of adding another function to the
// list in Context.
func (ctxt *Context) isFile(path string) bool {
	f, err := ctxt.openFile(path)
	if err != nil {
		return false
	}
	f.Close()
	return true
}

// gopath returns the list of Go path directories.
func (ctxt *Context) gopath() []string {
	var all []string
	for _, p := range ctxt.splitPathList(ctxt.GOPATH) {
		if p == "" || p == ctxt.GOROOT {
			// Empty paths are uninteresting.
			// If the path is the GOROOT, ignore it.
			// People sometimes set GOPATH=$GOROOT.
			// Do not get confused by this common mistake.
			continue
		}
		if strings.HasPrefix(p, "~") {
			// Path segments starting with ~ on Unix are almost always
			// users who have incorrectly quoted ~ while setting GOPATH,
			// preventing it from expanding to $HOME.
			// The situation is made more confusing by the fact that
			// bash allows quoted ~ in $PATH (most shells do not).
			// Do not get confused by this, and do not try to use the path.
			// It does not exist, and printing errors about it confuses
			// those users even more, because they think "sure ~ exists!".
			// The go command diagnoses this situation and prints a
			// useful error.
			// On Windows, ~ is used in short names, such as c:\progra~1
			// for c:\program files.
			continue
		}
		all = append(all, p)
	}
	return all
}

// SrcDirs returns a list of package source root directories.
// It draws from the current Go root and Go path but omits directories
// that do not exist.
func (ctxt *Context) SrcDirs() []string {
	var all []string
	if ctxt.GOROOT != "" && ctxt.Compiler != "gccgo" {
		dir := ctxt.joinPath(ctxt.GOROOT, "src")
		if ctxt.isDir(dir) {
			all = append(all, dir)
		}
	}
	for _, p := range ctxt.gopath() {
		dir := ctxt.joinPath(p, "src")
		if ctxt.isDir(dir) {
			all = append(all, dir)
		}
	}
	return all
}

// Default is the default Context for builds.
// It uses the GOARCH, GOOS, GOROOT, and GOPATH environment variables
// if set, or else the compiled code's GOARCH, GOOS, and GOROOT.
var Default Context = defaultContext()

func defaultGOPATH() string {
	env := "HOME"
	if runtime.GOOS == "windows" {
		env = "USERPROFILE"
	} else if runtime.GOOS == "plan9" {
		env = "home"
	}
	if home := os.Getenv(env); home != "" {
		def := filepath.Join(home, "go")
		if filepath.Clean(def) == filepath.Clean(runtime.GOROOT()) {
			// Don't set the default GOPATH to GOROOT,
			// as that will trigger warnings from the go tool.
			return ""
		}
		return def
	}
	return ""
}

var defaultReleaseTags []string

func defaultContext() Context {
	var c Context

	c.GOARCH = envOr("GOARCH", runtime.GOARCH)
	c.GOOS = envOr("GOOS", runtime.GOOS)
	c.GOROOT = pathpkg.Clean(runtime.GOROOT())
	c.GOPATH = envOr("GOPATH", defaultGOPATH())
	c.Compiler = runtime.Compiler

	// Each major Go release in the Go 1.x series should add a tag here.
	// Old tags should not be removed. That is, the go1.x tag is present
	// in all releases >= Go 1.x. Code that requires Go 1.x or later should
	// say "+build go1.x", and code that should only be built before Go 1.x
	// (perhaps it is the stub to use in that case) should say "+build !go1.x".
	// NOTE: If you add to this list, also update the doc comment in doc.go.
	// NOTE: The last element in ReleaseTags should be the current release.
	const version = 13 // go1.13
	for i := 1; i <= version; i++ {
		c.ReleaseTags = append(c.ReleaseTags, "go1."+strconv.Itoa(i))
	}

	defaultReleaseTags = append([]string{}, c.ReleaseTags...) // our own private copy

	env := os.Getenv("CGO_ENABLED")
	if env == "" {
		env = defaultCGO_ENABLED
	}
	switch env {
	case "1":
		c.CgoEnabled = true
	case "0":
		c.CgoEnabled = false
	default:
		// cgo must be explicitly enabled for cross compilation builds
		if runtime.GOARCH == c.GOARCH && runtime.GOOS == c.GOOS {
			c.CgoEnabled = cgoEnabled[c.GOOS+"/"+c.GOARCH]
			break
		}
		c.CgoEnabled = false
	}

	return c
}

func envOr(name, def string) string {
	s := os.Getenv(name)
	if s == "" {
		return def
	}
	return s
}

// An ImportMode controls the behavior of the Import method.
type ImportMode uint

const (
	// If FindOnly is set, Import stops after locating the directory
	// that should contain the sources for a package. It does not
	// read any files in the directory.
	FindOnly ImportMode = 1 << iota

	// If AllowBinary is set, Import can be satisfied by a compiled
	// package object without corresponding sources.
	//
	// Deprecated:
	// The supported way to create a compiled-only package is to
	// write source code containing a //go:binary-only-package comment at
	// the top of the file. Such a package will be recognized
	// regardless of this flag setting (because it has source code)
	// and will have BinaryOnly set to true in the returned Package.
	AllowBinary

	// If ImportComment is set, parse import comments on package statements.
	// Import returns an error if it finds a comment it cannot understand
	// or finds conflicting comments in multiple source files.
	// See golang.org/s/go14customimport for more information.
	ImportComment

	// By default, Import searches vendor directories
	// that apply in the given source directory before searching
	// the GOROOT and GOPATH roots.
	// If an Import finds and returns a package using a vendor
	// directory, the resulting ImportPath is the complete path
	// to the package, including the path elements leading up
	// to and including "vendor".
	// For example, if Import("y", "x/subdir", 0) finds
	// "x/vendor/y", the returned package's ImportPath is "x/vendor/y",
	// not plain "y".
	// See golang.org/s/go15vendor for more information.
	//
	// Setting IgnoreVendor ignores vendor directories.
	//
	// In contrast to the package's ImportPath,
	// the returned package's Imports, TestImports, and XTestImports
	// are always the exact import paths from the source files:
	// Import makes no attempt to resolve or check those paths.
	IgnoreVendor
)

// A Package describes the Go package found in a directory.
type Package struct {
	Dir           string   // directory containing package sources
	Name          string   // package name
	ImportComment string   // path in import comment on package statement
	Doc           string   // documentation synopsis
	ImportPath    string   // import path of package ("" if unknown)
	Root          string   // root of Go tree where this package lives
	SrcRoot       string   // package source root directory ("" if unknown)
	PkgRoot       string   // package install root directory ("" if unknown)
	PkgTargetRoot string   // architecture dependent install root directory ("" if unknown)
	BinDir        string   // command install directory ("" if unknown)
	Goroot        bool     // package found in Go root
	PkgObj        string   // installed .a file
	AllTags       []string // tags that can influence file selection in this directory
	ConflictDir   string   // this directory shadows Dir in $GOPATH
	BinaryOnly    bool     // cannot be rebuilt from source (has //go:binary-only-package comment)

	// Source files
	GoFiles        []string // .go source files (excluding CgoFiles, TestGoFiles, XTestGoFiles)
	CgoFiles       []string // .go source files that import "C"
	IgnoredGoFiles []string // .go source files ignored for this build
	InvalidGoFiles []string // .go source files with detected problems (parse error, wrong package name, and so on)
	CFiles         []string // .c source files
	CXXFiles       []string // .cc, .cpp and .cxx source files
	MFiles         []string // .m (Objective-C) source files
	HFiles         []string // .h, .hh, .hpp and .hxx source files
	FFiles         []string // .f, .F, .for and .f90 Fortran source files
	SFiles         []string // .s source files
	SwigFiles      []string // .swig files
	SwigCXXFiles   []string // .swigcxx files
	SysoFiles      []string // .syso system object files to add to archive

	// Cgo directives
	CgoCFLAGS    []string // Cgo CFLAGS directives
	CgoCPPFLAGS  []string // Cgo CPPFLAGS directives
	CgoCXXFLAGS  []string // Cgo CXXFLAGS directives
	CgoFFLAGS    []string // Cgo FFLAGS directives
	CgoLDFLAGS   []string // Cgo LDFLAGS directives
	CgoPkgConfig []string // Cgo pkg-config directives

	// Dependency information
	Imports   []string                    // import paths from GoFiles, CgoFiles
	ImportPos map[string][]token.Position // line information for Imports

	// Test information
	TestGoFiles    []string                    // _test.go files in package
	TestImports    []string                    // import paths from TestGoFiles
	TestImportPos  map[string][]token.Position // line information for TestImports
	XTestGoFiles   []string                    // _test.go files outside package
	XTestImports   []string                    // import paths from XTestGoFiles
	XTestImportPos map[string][]token.Position // line information for XTestImports
}

// IsCommand reports whether the package is considered a
// command to be installed (not just a library).
// Packages named "main" are treated as commands.
func (p *Package) IsCommand() bool {
	return p.Name == "main"
}

// ImportDir is like Import but processes the Go package found in
// the named directory.
func (ctxt *Context) ImportDir(dir string, mode ImportMode) (*Package, error) {
	return ctxt.Import(".", dir, mode)
}

// NoGoError is the error used by Import to describe a directory
// containing no buildable Go source files. (It may still contain
// test files, files hidden by build tags, and so on.)
type NoGoError struct {
	Dir string
}

func (e *NoGoError) Error() string {
	return "no buildable Go source files in " + e.Dir
}

// MultiplePackageError describes a directory containing
// multiple buildable Go source files for multiple packages.
type MultiplePackageError struct {
	Dir      string   // directory containing files
	Packages []string // package names found
	Files    []string // corresponding files: Files[i] declares package Packages[i]
}

func (e *MultiplePackageError) Error() string {
	// Error string limited to two entries for compatibility.
	return fmt.Sprintf("found packages %s (%s) and %s (%s) in %s", e.Packages[0], e.Files[0], e.Packages[1], e.Files[1], e.Dir)
}

func nameExt(name string) string {
	i := strings.LastIndex(name, ".")
	if i < 0 {
		return ""
	}
	return name[i:]
}

// Import returns details about the Go package named by the import path,
// interpreting local import paths relative to the srcDir directory.
// If the path is a local import path naming a package that can be imported
// using a standard import path, the returned package will set p.ImportPath
// to that path.
//
// In the directory containing the package, .go, .c, .h, and .s files are
// considered part of the package except for:
//
//	- .go files in package documentation
//	- files starting with _ or . (likely editor temporary files)
//	- files with build constraints not satisfied by the context
//
// If an error occurs, Import returns a non-nil error and a non-nil
// *Package containing partial information.
//
func (ctxt *Context) Import(path string, srcDir string, mode ImportMode) (*Package, error) {
	p := &Package{
		ImportPath: path,
	}
	if path == "" {
		return p, fmt.Errorf("import %q: invalid import path", path)
	}

	var pkgtargetroot string
	var pkga string
	var pkgerr error
	suffix := ""
	if ctxt.InstallSuffix != "" {
		suffix = "_" + ctxt.InstallSuffix
	}
	switch ctxt.Compiler {
	case "gccgo":
		pkgtargetroot = "pkg/gccgo_" + ctxt.GOOS + "_" + ctxt.GOARCH + suffix
	case "gc":
		pkgtargetroot = "pkg/" + ctxt.GOOS + "_" + ctxt.GOARCH + suffix
	default:
		// Save error for end of function.
		pkgerr = fmt.Errorf("import %q: unknown compiler %q", path, ctxt.Compiler)
	}
	setPkga := func() {
		switch ctxt.Compiler {
		case "gccgo":
			dir, elem := pathpkg.Split(p.ImportPath)
			pkga = pkgtargetroot + "/" + dir + "lib" + elem + ".a"
		case "gc":
			pkga = pkgtargetroot + "/" + p.ImportPath + ".a"
		}
	}
	setPkga()

	binaryOnly := false
	if IsLocalImport(path) {
		pkga = "" // local imports have no installed path
		if srcDir == "" {
			return p, fmt.Errorf("import %q: import relative to unknown directory", path)
		}
		if !ctxt.isAbsPath(path) {
			p.Dir = ctxt.joinPath(srcDir, path)
		}
		// p.Dir directory may or may not exist. Gather partial information first, check if it exists later.
		// Determine canonical import path, if any.
		// Exclude results where the import path would include /testdata/.
		inTestdata := func(sub string) bool {
			return strings.Contains(sub, "/testdata/") || strings.HasSuffix(sub, "/testdata") || strings.HasPrefix(sub, "testdata/") || sub == "testdata"
		}
		if ctxt.GOROOT != "" {
			root := ctxt.joinPath(ctxt.GOROOT, "src")
			if sub, ok := ctxt.hasSubdir(root, p.Dir); ok && !inTestdata(sub) {
				p.Goroot = true
				p.ImportPath = sub
				p.Root = ctxt.GOROOT
				setPkga() // p.ImportPath changed
				goto Found
			}
		}
		all := ctxt.gopath()
		for i, root := range all {
			rootsrc := ctxt.joinPath(root, "src")
			if sub, ok := ctxt.hasSubdir(rootsrc, p.Dir); ok && !inTestdata(sub) {
				// We found a potential import path for dir,
				// but check that using it wouldn't find something
				// else first.
				if ctxt.GOROOT != "" && ctxt.Compiler != "gccgo" {
					if dir := ctxt.joinPath(ctxt.GOROOT, "src", sub); ctxt.isDir(dir) {
						p.ConflictDir = dir
						goto Found
					}
				}
				for _, earlyRoot := range all[:i] {
					if dir := ctxt.joinPath(earlyRoot, "src", sub); ctxt.isDir(dir) {
						p.ConflictDir = dir
						goto Found
					}
				}

				// sub would not name some other directory instead of this one.
				// Record it.
				p.ImportPath = sub
				p.Root = root
				setPkga() // p.ImportPath changed
				goto Found
			}
		}
		// It's okay that we didn't find a root containing dir.
		// Keep going with the information we have.
	} else {
		if strings.HasPrefix(path, "/") {
			return p, fmt.Errorf("import %q: cannot import absolute path", path)
		}

		gopath := ctxt.gopath() // needed by both importGo and below; avoid computing twice
		if err := ctxt.importGo(p, path, srcDir, mode, gopath); err == nil {
			goto Found
		} else if err != errNoModules {
			return p, err
		}

		// tried records the location of unsuccessful package lookups
		var tried struct {
			vendor []string
			goroot string
			gopath []string
		}

		// Vendor directories get first chance to satisfy import.
		if mode&IgnoreVendor == 0 && srcDir != "" {
			searchVendor := func(root string, isGoroot bool) bool {
				sub, ok := ctxt.hasSubdir(root, srcDir)
				if !ok || !strings.HasPrefix(sub, "src/") || strings.Contains(sub, "/testdata/") {
					return false
				}
				for {
					vendor := ctxt.joinPath(root, sub, "vendor")
					if ctxt.isDir(vendor) {
						dir := ctxt.joinPath(vendor, path)
						if ctxt.isDir(dir) && hasGoFiles(ctxt, dir) {
							p.Dir = dir
							p.ImportPath = strings.TrimPrefix(pathpkg.Join(sub, "vendor", path), "src/")
							p.Goroot = isGoroot
							p.Root = root
							setPkga() // p.ImportPath changed
							return true
						}
						tried.vendor = append(tried.vendor, dir)
					}
					i := strings.LastIndex(sub, "/")
					if i < 0 {
						break
					}
					sub = sub[:i]
				}
				return false
			}
			if ctxt.Compiler != "gccgo" && searchVendor(ctxt.GOROOT, true) {
				goto Found
			}
			for _, root := range gopath {
				if searchVendor(root, false) {
					goto Found
				}
			}
		}

		// Determine directory from import path.
		if ctxt.GOROOT != "" {
			// If the package path starts with "vendor/", only search GOROOT before
			// GOPATH if the importer is also within GOROOT. That way, if the user has
			// vendored in a package that is subsequently included in the standard
			// distribution, they'll continue to pick up their own vendored copy.
			gorootFirst := srcDir == "" || !strings.HasPrefix(path, "vendor/")
			if !gorootFirst {
				_, gorootFirst = ctxt.hasSubdir(ctxt.GOROOT, srcDir)
			}
			if gorootFirst {
				dir := ctxt.joinPath(ctxt.GOROOT, "src", path)
				if ctxt.Compiler != "gccgo" {
					isDir := ctxt.isDir(dir)
					binaryOnly = !isDir && mode&AllowBinary != 0 && pkga != "" && ctxt.isFile(ctxt.joinPath(ctxt.GOROOT, pkga))
					if isDir || binaryOnly {
						p.Dir = dir
						p.Goroot = true
						p.Root = ctxt.GOROOT
						goto Found
					}
				}
				tried.goroot = dir
			}
		}
		if ctxt.Compiler == "gccgo" && goroot.IsStandardPackage(ctxt.GOROOT, ctxt.Compiler, path) {
			p.Dir = ctxt.joinPath(ctxt.GOROOT, "src", path)
			p.Goroot = true
			p.Root = ctxt.GOROOT
			goto Found
		}
		for _, root := range gopath {
			dir := ctxt.joinPath(root, "src", path)
			isDir := ctxt.isDir(dir)
			binaryOnly = !isDir && mode&AllowBinary != 0 && pkga != "" && ctxt.isFile(ctxt.joinPath(root, pkga))
			if isDir || binaryOnly {
				p.Dir = dir
				p.Root = root
				goto Found
			}
			tried.gopath = append(tried.gopath, dir)
		}

		// If we tried GOPATH first due to a "vendor/" prefix, fall back to GOPATH.
		// That way, the user can still get useful results from 'go list' for
		// standard-vendored paths passed on the command line.
		if ctxt.GOROOT != "" && tried.goroot == "" {
			dir := ctxt.joinPath(ctxt.GOROOT, "src", path)
			if ctxt.Compiler != "gccgo" {
				isDir := ctxt.isDir(dir)
				binaryOnly = !isDir && mode&AllowBinary != 0 && pkga != "" && ctxt.isFile(ctxt.joinPath(ctxt.GOROOT, pkga))
				if isDir || binaryOnly {
					p.Dir = dir
					p.Goroot = true
					p.Root = ctxt.GOROOT
					goto Found
				}
			}
			tried.goroot = dir
		}

		// package was not found
		var paths []string
		format := "\t%s (vendor tree)"
		for _, dir := range tried.vendor {
			paths = append(paths, fmt.Sprintf(format, dir))
			format = "\t%s"
		}
		if tried.goroot != "" {
			paths = append(paths, fmt.Sprintf("\t%s (from $GOROOT)", tried.goroot))
		} else {
			paths = append(paths, "\t($GOROOT not set)")
		}
		format = "\t%s (from $GOPATH)"
		for _, dir := range tried.gopath {
			paths = append(paths, fmt.Sprintf(format, dir))
			format = "\t%s"
		}
		if len(tried.gopath) == 0 {
			paths = append(paths, "\t($GOPATH not set. For more details see: 'go help gopath')")
		}
		return p, fmt.Errorf("cannot find package %q in any of:\n%s", path, strings.Join(paths, "\n"))
	}

Found:
	if p.Root != "" {
		p.SrcRoot = ctxt.joinPath(p.Root, "src")
		p.PkgRoot = ctxt.joinPath(p.Root, "pkg")
		p.BinDir = ctxt.joinPath(p.Root, "bin")
		if pkga != "" {
			p.PkgTargetRoot = ctxt.joinPath(p.Root, pkgtargetroot)
			p.PkgObj = ctxt.joinPath(p.Root, pkga)
		}
	}

	// If it's a local import path, by the time we get here, we still haven't checked
	// that p.Dir directory exists. This is the right time to do that check.
	// We can't do it earlier, because we want to gather partial information for the
	// non-nil *Package returned when an error occurs.
	// We need to do this before we return early on FindOnly flag.
	if IsLocalImport(path) && !ctxt.isDir(p.Dir) {
		if ctxt.Compiler == "gccgo" && p.Goroot {
			// gccgo has no sources for GOROOT packages.
			return p, nil
		}

		// package was not found
		return p, fmt.Errorf("cannot find package %q in:\n\t%s", path, p.Dir)
	}

	if mode&FindOnly != 0 {
		return p, pkgerr
	}
	if binaryOnly && (mode&AllowBinary) != 0 {
		return p, pkgerr
	}

	if ctxt.Compiler == "gccgo" && p.Goroot {
		// gccgo has no sources for GOROOT packages.
		return p, nil
	}

	dirs, err := ctxt.readDir(p.Dir)
	if err != nil {
		return p, err
	}

	var badGoError error
	var Sfiles []string // files with ".S" (capital S)
	var firstFile, firstCommentFile string
	imported := make(map[string][]token.Position)
	testImported := make(map[string][]token.Position)
	xTestImported := make(map[string][]token.Position)
	allTags := make(map[string]bool)
	fset := token.NewFileSet()
	for _, d := range dirs {
		if d.IsDir() {
			continue
		}

		name := d.Name()
		ext := nameExt(name)

		badFile := func(err error) {
			if badGoError == nil {
				badGoError = err
			}
			p.InvalidGoFiles = append(p.InvalidGoFiles, name)
		}

		match, data, filename, err := ctxt.matchFile(p.Dir, name, allTags, &p.BinaryOnly)
		if err != nil {
			badFile(err)
			continue
		}
		if !match {
			if ext == ".go" {
				p.IgnoredGoFiles = append(p.IgnoredGoFiles, name)
			}
			continue
		}

		// Going to save the file. For non-Go files, can stop here.
		switch ext {
		case ".c":
			p.CFiles = append(p.CFiles, name)
			continue
		case ".cc", ".cpp", ".cxx":
			p.CXXFiles = append(p.CXXFiles, name)
			continue
		case ".m":
			p.MFiles = append(p.MFiles, name)
			continue
		case ".h", ".hh", ".hpp", ".hxx":
			p.HFiles = append(p.HFiles, name)
			continue
		case ".f", ".F", ".for", ".f90":
			p.FFiles = append(p.FFiles, name)
			continue
		case ".s":
			p.SFiles = append(p.SFiles, name)
			continue
		case ".S":
			Sfiles = append(Sfiles, name)
			continue
		case ".swig":
			p.SwigFiles = append(p.SwigFiles, name)
			continue
		case ".swigcxx":
			p.SwigCXXFiles = append(p.SwigCXXFiles, name)
			continue
		case ".syso":
			// binary objects to add to package archive
			// Likely of the form foo_windows.syso, but
			// the name was vetted above with goodOSArchFile.
			p.SysoFiles = append(p.SysoFiles, name)
			continue
		}

		pf, err := parser.ParseFile(fset, filename, data, parser.ImportsOnly|parser.ParseComments)
		if err != nil {
			badFile(err)
			continue
		}

		pkg := pf.Name.Name
		if pkg == "documentation" {
			p.IgnoredGoFiles = append(p.IgnoredGoFiles, name)
			continue
		}

		isTest := strings.HasSuffix(name, "_test.go")
		isXTest := false
		if isTest && strings.HasSuffix(pkg, "_test") {
			isXTest = true
			pkg = pkg[:len(pkg)-len("_test")]
		}

		if p.Name == "" {
			p.Name = pkg
			firstFile = name
		} else if pkg != p.Name {
			badFile(&MultiplePackageError{
				Dir:      p.Dir,
				Packages: []string{p.Name, pkg},
				Files:    []string{firstFile, name},
			})
			p.InvalidGoFiles = append(p.InvalidGoFiles, name)
		}
		// Grab the first package comment as docs, provided it is not from a test file.
		if pf.Doc != nil && p.Doc == "" && !isTest && !isXTest {
			p.Doc = doc.Synopsis(pf.Doc.Text())
		}

		if mode&ImportComment != 0 {
			qcom, line := findImportComment(data)
			if line != 0 {
				com, err := strconv.Unquote(qcom)
				if err != nil {
					badFile(fmt.Errorf("%s:%d: cannot parse import comment", filename, line))
				} else if p.ImportComment == "" {
					p.ImportComment = com
					firstCommentFile = name
				} else if p.ImportComment != com {
					badFile(fmt.Errorf("found import comments %q (%s) and %q (%s) in %s", p.ImportComment, firstCommentFile, com, name, p.Dir))
				}
			}
		}

		// Record imports and information about cgo.
		isCgo := false
		for _, decl := range pf.Decls {
			d, ok := decl.(*ast.GenDecl)
			if !ok {
				continue
			}
			for _, dspec := range d.Specs {
				spec, ok := dspec.(*ast.ImportSpec)
				if !ok {
					continue
				}
				quoted := spec.Path.Value
				path, err := strconv.Unquote(quoted)
				if err != nil {
					log.Panicf("%s: parser returned invalid quoted string: <%s>", filename, quoted)
				}
				if isXTest {
					xTestImported[path] = append(xTestImported[path], fset.Position(spec.Pos()))
				} else if isTest {
					testImported[path] = append(testImported[path], fset.Position(spec.Pos()))
				} else {
					imported[path] = append(imported[path], fset.Position(spec.Pos()))
				}
				if path == "C" {
					if isTest {
						badFile(fmt.Errorf("use of cgo in test %s not supported", filename))
					} else {
						cg := spec.Doc
						if cg == nil && len(d.Specs) == 1 {
							cg = d.Doc
						}
						if cg != nil {
							if err := ctxt.saveCgo(filename, p, cg); err != nil {
								badFile(err)
							}
						}
						isCgo = true
					}
				}
			}
		}
		if isCgo {
			allTags["cgo"] = true
			if ctxt.CgoEnabled {
				p.CgoFiles = append(p.CgoFiles, name)
			} else {
				p.IgnoredGoFiles = append(p.IgnoredGoFiles, name)
			}
		} else if isXTest {
			p.XTestGoFiles = append(p.XTestGoFiles, name)
		} else if isTest {
			p.TestGoFiles = append(p.TestGoFiles, name)
		} else {
			p.GoFiles = append(p.GoFiles, name)
		}
	}
	if badGoError != nil {
		return p, badGoError
	}
	if len(p.GoFiles)+len(p.CgoFiles)+len(p.TestGoFiles)+len(p.XTestGoFiles) == 0 {
		return p, &NoGoError{p.Dir}
	}

	for tag := range allTags {
		p.AllTags = append(p.AllTags, tag)
	}
	sort.Strings(p.AllTags)

	p.Imports, p.ImportPos = cleanImports(imported)
	p.TestImports, p.TestImportPos = cleanImports(testImported)
	p.XTestImports, p.XTestImportPos = cleanImports(xTestImported)

	// add the .S files only if we are using cgo
	// (which means gcc will compile them).
	// The standard assemblers expect .s files.
	if len(p.CgoFiles) > 0 {
		p.SFiles = append(p.SFiles, Sfiles...)
		sort.Strings(p.SFiles)
	}

	return p, pkgerr
}

var errNoModules = errors.New("not using modules")

// importGo checks whether it can use the go command to find the directory for path.
// If using the go command is not appopriate, importGo returns errNoModules.
// Otherwise, importGo tries using the go command and reports whether that succeeded.
// Using the go command lets build.Import and build.Context.Import find code
// in Go modules. In the long term we want tools to use go/packages (currently golang.org/x/tools/go/packages),
// which will also use the go command.
// Invoking the go command here is not very efficient in that it computes information
// about the requested package and all dependencies and then only reports about the requested package.
// Then we reinvoke it for every dependency. But this is still better than not working at all.
// See golang.org/issue/26504.
func (ctxt *Context) importGo(p *Package, path, srcDir string, mode ImportMode, gopath []string) error {
	const debugImportGo = false

	// To invoke the go command, we must know the source directory,
	// we must not being doing special things like AllowBinary or IgnoreVendor,
	// and all the file system callbacks must be nil (we're meant to use the local file system).
	if srcDir == "" || mode&AllowBinary != 0 || mode&IgnoreVendor != 0 ||
		ctxt.JoinPath != nil || ctxt.SplitPathList != nil || ctxt.IsAbsPath != nil || ctxt.IsDir != nil || ctxt.HasSubdir != nil || ctxt.ReadDir != nil || ctxt.OpenFile != nil || !equal(ctxt.ReleaseTags, defaultReleaseTags) {
		return errNoModules
	}

	// If modules are not enabled, then the in-process code works fine and we should keep using it.
	// TODO(bcmills): This assumes that the default is "auto" instead of "on".
	switch os.Getenv("GO111MODULE") {
	case "off":
		return errNoModules
	case "on":
		// ok
	default: // "", "auto", anything else
		// Automatic mode: no module use in $GOPATH/src.
		for _, root := range gopath {
			sub, ok := ctxt.hasSubdir(root, srcDir)
			if ok && strings.HasPrefix(sub, "src/") {
				return errNoModules
			}
		}
	}

	// If the source directory is in GOROOT, then the in-process code works fine
	// and we should keep using it. Moreover, the 'go list' approach below doesn't
	// take standard-library vendoring into account and will fail.
	if _, ok := ctxt.hasSubdir(filepath.Join(ctxt.GOROOT, "src"), srcDir); ok {
		return errNoModules
	}

	// For efficiency, if path is a standard library package, let the usual lookup code handle it.
	if ctxt.GOROOT != "" {
		dir := ctxt.joinPath(ctxt.GOROOT, "src", path)
		if ctxt.isDir(dir) {
			return errNoModules
		}
	}

	// Look to see if there is a go.mod.
	abs, err := filepath.Abs(srcDir)
	if err != nil {
		return errNoModules
	}
	for {
		info, err := os.Stat(filepath.Join(abs, "go.mod"))
		if err == nil && !info.IsDir() {
			break
		}
		d := filepath.Dir(abs)
		if len(d) >= len(abs) {
			return errNoModules // reached top of file system, no go.mod
		}
		abs = d
	}

	cmd := exec.Command("go", "list", "-compiler="+ctxt.Compiler, "-tags="+strings.Join(ctxt.BuildTags, ","), "-installsuffix="+ctxt.InstallSuffix, "-f={{.Dir}}\n{{.ImportPath}}\n{{.Root}}\n{{.Goroot}}\n", path)

	// TODO(bcmills): This is wrong if srcDir is in a vendor directory, or if
	// srcDir is in some module dependency of the main module. The main module
	// chooses what the import paths mean: individual packages don't.
	cmd.Dir = srcDir

	var stdout, stderr strings.Builder
	cmd.Stdout = &stdout
	cmd.Stderr = &stderr

	cgo := "0"
	if ctxt.CgoEnabled {
		cgo = "1"
	}
	cmd.Env = append(os.Environ(),
		"GOOS="+ctxt.GOOS,
		"GOARCH="+ctxt.GOARCH,
		"GOROOT="+ctxt.GOROOT,
		"GOPATH="+ctxt.GOPATH,
		"CGO_ENABLED="+cgo,
	)

	if err := cmd.Run(); err != nil {
		return fmt.Errorf("go/build: importGo %s: %v\n%s\n", path, err, stderr.String())
	}

	f := strings.Split(stdout.String(), "\n")
	if len(f) != 5 || f[4] != "" {
		return fmt.Errorf("go/build: importGo %s: unexpected output:\n%s\n", path, stdout.String())
	}

	p.Dir = f[0]
	p.ImportPath = f[1]
	p.Root = f[2]
	p.Goroot = f[3] == "true"
	return nil
}

func equal(x, y []string) bool {
	if len(x) != len(y) {
		return false
	}
	for i, xi := range x {
		if xi != y[i] {
			return false
		}
	}
	return true
}

// hasGoFiles reports whether dir contains any files with names ending in .go.
// For a vendor check we must exclude directories that contain no .go files.
// Otherwise it is not possible to vendor just a/b/c and still import the
// non-vendored a/b. See golang.org/issue/13832.
func hasGoFiles(ctxt *Context, dir string) bool {
	ents, _ := ctxt.readDir(dir)
	for _, ent := range ents {
		if !ent.IsDir() && strings.HasSuffix(ent.Name(), ".go") {
			return true
		}
	}
	return false
}

func findImportComment(data []byte) (s string, line int) {
	// expect keyword package
	word, data := parseWord(data)
	if string(word) != "package" {
		return "", 0
	}

	// expect package name
	_, data = parseWord(data)

	// now ready for import comment, a // or /* */ comment
	// beginning and ending on the current line.
	for len(data) > 0 && (data[0] == ' ' || data[0] == '\t' || data[0] == '\r') {
		data = data[1:]
	}

	var comment []byte
	switch {
	case bytes.HasPrefix(data, slashSlash):
		i := bytes.Index(data, newline)
		if i < 0 {
			i = len(data)
		}
		comment = data[2:i]
	case bytes.HasPrefix(data, slashStar):
		data = data[2:]
		i := bytes.Index(data, starSlash)
		if i < 0 {
			// malformed comment
			return "", 0
		}
		comment = data[:i]
		if bytes.Contains(comment, newline) {
			return "", 0
		}
	}
	comment = bytes.TrimSpace(comment)

	// split comment into `import`, `"pkg"`
	word, arg := parseWord(comment)
	if string(word) != "import" {
		return "", 0
	}

	line = 1 + bytes.Count(data[:cap(data)-cap(arg)], newline)
	return strings.TrimSpace(string(arg)), line
}

var (
	slashSlash = []byte("//")
	slashStar  = []byte("/*")
	starSlash  = []byte("*/")
	newline    = []byte("\n")
)

// skipSpaceOrComment returns data with any leading spaces or comments removed.
func skipSpaceOrComment(data []byte) []byte {
	for len(data) > 0 {
		switch data[0] {
		case ' ', '\t', '\r', '\n':
			data = data[1:]
			continue
		case '/':
			if bytes.HasPrefix(data, slashSlash) {
				i := bytes.Index(data, newline)
				if i < 0 {
					return nil
				}
				data = data[i+1:]
				continue
			}
			if bytes.HasPrefix(data, slashStar) {
				data = data[2:]
				i := bytes.Index(data, starSlash)
				if i < 0 {
					return nil
				}
				data = data[i+2:]
				continue
			}
		}
		break
	}
	return data
}

// parseWord skips any leading spaces or comments in data
// and then parses the beginning of data as an identifier or keyword,
// returning that word and what remains after the word.
func parseWord(data []byte) (word, rest []byte) {
	data = skipSpaceOrComment(data)

	// Parse past leading word characters.
	rest = data
	for {
		r, size := utf8.DecodeRune(rest)
		if unicode.IsLetter(r) || '0' <= r && r <= '9' || r == '_' {
			rest = rest[size:]
			continue
		}
		break
	}

	word = data[:len(data)-len(rest)]
	if len(word) == 0 {
		return nil, nil
	}

	return word, rest
}

// MatchFile reports whether the file with the given name in the given directory
// matches the context and would be included in a Package created by ImportDir
// of that directory.
//
// MatchFile considers the name of the file and may use ctxt.OpenFile to
// read some or all of the file's content.
func (ctxt *Context) MatchFile(dir, name string) (match bool, err error) {
	match, _, _, err = ctxt.matchFile(dir, name, nil, nil)
	return
}

// matchFile determines whether the file with the given name in the given directory
// should be included in the package being constructed.
// It returns the data read from the file.
// If name denotes a Go program, matchFile reads until the end of the
// imports (and returns that data) even though it only considers text
// until the first non-comment.
// If allTags is non-nil, matchFile records any encountered build tag
// by setting allTags[tag] = true.
func (ctxt *Context) matchFile(dir, name string, allTags map[string]bool, binaryOnly *bool) (match bool, data []byte, filename string, err error) {
	if strings.HasPrefix(name, "_") ||
		strings.HasPrefix(name, ".") {
		return
	}

	i := strings.LastIndex(name, ".")
	if i < 0 {
		i = len(name)
	}
	ext := name[i:]

	if !ctxt.goodOSArchFile(name, allTags) && !ctxt.UseAllFiles {
		return
	}

	switch ext {
	case ".go", ".c", ".cc", ".cxx", ".cpp", ".m", ".s", ".h", ".hh", ".hpp", ".hxx", ".f", ".F", ".f90", ".S", ".swig", ".swigcxx":
		// tentatively okay - read to make sure
	case ".syso":
		// binary, no reading
		match = true
		return
	default:
		// skip
		return
	}

	filename = ctxt.joinPath(dir, name)
	f, err := ctxt.openFile(filename)
	if err != nil {
		return
	}

	if strings.HasSuffix(filename, ".go") {
		data, err = readImports(f, false, nil)
		if strings.HasSuffix(filename, "_test.go") {
			binaryOnly = nil // ignore //go:binary-only-package comments in _test.go files
		}
	} else {
		binaryOnly = nil // ignore //go:binary-only-package comments in non-Go sources
		data, err = readComments(f)
	}
	f.Close()
	if err != nil {
		err = fmt.Errorf("read %s: %v", filename, err)
		return
	}

	// Look for +build comments to accept or reject the file.
	var sawBinaryOnly bool
	if !ctxt.shouldBuild(data, allTags, &sawBinaryOnly) && !ctxt.UseAllFiles {
		return
	}

	if binaryOnly != nil && sawBinaryOnly {
		*binaryOnly = true
	}
	match = true
	return
}

func cleanImports(m map[string][]token.Position) ([]string, map[string][]token.Position) {
	all := make([]string, 0, len(m))
	for path := range m {
		all = append(all, path)
	}
	sort.Strings(all)
	return all, m
}

// Import is shorthand for Default.Import.
func Import(path, srcDir string, mode ImportMode) (*Package, error) {
	return Default.Import(path, srcDir, mode)
}

// ImportDir is shorthand for Default.ImportDir.
func ImportDir(dir string, mode ImportMode) (*Package, error) {
	return Default.ImportDir(dir, mode)
}

var slashslash = []byte("//")

// Special comment denoting a binary-only package.
// See https://golang.org/design/2775-binary-only-packages
// for more about the design of binary-only packages.
var binaryOnlyComment = []byte("//go:binary-only-package")

// shouldBuild reports whether it is okay to use this file,
// The rule is that in the file's leading run of // comments
// and blank lines, which must be followed by a blank line
// (to avoid including a Go package clause doc comment),
// lines beginning with '// +build' are taken as build directives.
//
// The file is accepted only if each such line lists something
// matching the file. For example:
//
//	// +build windows linux
//
// marks the file as applicable only on Windows and Linux.
//
// If shouldBuild finds a //go:binary-only-package comment in the file,
// it sets *binaryOnly to true. Otherwise it does not change *binaryOnly.
//
func (ctxt *Context) shouldBuild(content []byte, allTags map[string]bool, binaryOnly *bool) bool {
	sawBinaryOnly := false

	// Pass 1. Identify leading run of // comments and blank lines,
	// which must be followed by a blank line.
	end := 0
	p := content
	for len(p) > 0 {
		line := p
		if i := bytes.IndexByte(line, '\n'); i >= 0 {
			line, p = line[:i], p[i+1:]
		} else {
			p = p[len(p):]
		}
		line = bytes.TrimSpace(line)
		if len(line) == 0 { // Blank line
			end = len(content) - len(p)
			continue
		}
		if !bytes.HasPrefix(line, slashslash) { // Not comment line
			break
		}
	}
	content = content[:end]

	// Pass 2.  Process each line in the run.
	p = content
	allok := true
	for len(p) > 0 {
		line := p
		if i := bytes.IndexByte(line, '\n'); i >= 0 {
			line, p = line[:i], p[i+1:]
		} else {
			p = p[len(p):]
		}
		line = bytes.TrimSpace(line)
		if !bytes.HasPrefix(line, slashslash) {
			continue
		}
		if bytes.Equal(line, binaryOnlyComment) {
			sawBinaryOnly = true
		}
		line = bytes.TrimSpace(line[len(slashslash):])
		if len(line) > 0 && line[0] == '+' {
			// Looks like a comment +line.
			f := strings.Fields(string(line))
			if f[0] == "+build" {
				ok := false
				for _, tok := range f[1:] {
					if ctxt.match(tok, allTags) {
						ok = true
					}
				}
				if !ok {
					allok = false
				}
			}
		}
	}

	if binaryOnly != nil && sawBinaryOnly {
		*binaryOnly = true
	}

	return allok
}

// saveCgo saves the information from the #cgo lines in the import "C" comment.
// These lines set CFLAGS, CPPFLAGS, CXXFLAGS and LDFLAGS and pkg-config directives
// that affect the way cgo's C code is built.
func (ctxt *Context) saveCgo(filename string, di *Package, cg *ast.CommentGroup) error {
	text := cg.Text()
	for _, line := range strings.Split(text, "\n") {
		orig := line

		// Line is
		//	#cgo [GOOS/GOARCH...] LDFLAGS: stuff
		//
		line = strings.TrimSpace(line)
		if len(line) < 5 || line[:4] != "#cgo" || (line[4] != ' ' && line[4] != '\t') {
			continue
		}

		// Split at colon.
		line = strings.TrimSpace(line[4:])
		i := strings.Index(line, ":")
		if i < 0 {
			return fmt.Errorf("%s: invalid #cgo line: %s", filename, orig)
		}
		line, argstr := line[:i], line[i+1:]

		// Parse GOOS/GOARCH stuff.
		f := strings.Fields(line)
		if len(f) < 1 {
			return fmt.Errorf("%s: invalid #cgo line: %s", filename, orig)
		}

		cond, verb := f[:len(f)-1], f[len(f)-1]
		if len(cond) > 0 {
			ok := false
			for _, c := range cond {
				if ctxt.match(c, nil) {
					ok = true
					break
				}
			}
			if !ok {
				continue
			}
		}

		args, err := splitQuoted(argstr)
		if err != nil {
			return fmt.Errorf("%s: invalid #cgo line: %s", filename, orig)
		}
		var ok bool
		for i, arg := range args {
			if arg, ok = expandSrcDir(arg, di.Dir); !ok {
				return fmt.Errorf("%s: malformed #cgo argument: %s", filename, arg)
			}
			args[i] = arg
		}

		switch verb {
		case "CFLAGS", "CPPFLAGS", "CXXFLAGS", "FFLAGS", "LDFLAGS":
			// Change relative paths to absolute.
			ctxt.makePathsAbsolute(args, di.Dir)
		}

		switch verb {
		case "CFLAGS":
			di.CgoCFLAGS = append(di.CgoCFLAGS, args...)
		case "CPPFLAGS":
			di.CgoCPPFLAGS = append(di.CgoCPPFLAGS, args...)
		case "CXXFLAGS":
			di.CgoCXXFLAGS = append(di.CgoCXXFLAGS, args...)
		case "FFLAGS":
			di.CgoFFLAGS = append(di.CgoFFLAGS, args...)
		case "LDFLAGS":
			di.CgoLDFLAGS = append(di.CgoLDFLAGS, args...)
		case "pkg-config":
			di.CgoPkgConfig = append(di.CgoPkgConfig, args...)
		default:
			return fmt.Errorf("%s: invalid #cgo verb: %s", filename, orig)
		}
	}
	return nil
}

// expandSrcDir expands any occurrence of ${SRCDIR}, making sure
// the result is safe for the shell.
func expandSrcDir(str string, srcdir string) (string, bool) {
	// "\" delimited paths cause safeCgoName to fail
	// so convert native paths with a different delimiter
	// to "/" before starting (eg: on windows).
	srcdir = filepath.ToSlash(srcdir)

	chunks := strings.Split(str, "${SRCDIR}")
	if len(chunks) < 2 {
		return str, safeCgoName(str)
	}
	ok := true
	for _, chunk := range chunks {
		ok = ok && (chunk == "" || safeCgoName(chunk))
	}
	ok = ok && (srcdir == "" || safeCgoName(srcdir))
	res := strings.Join(chunks, srcdir)
	return res, ok && res != ""
}

// makePathsAbsolute looks for compiler options that take paths and
// makes them absolute. We do this because through the 1.8 release we
// ran the compiler in the package directory, so any relative -I or -L
// options would be relative to that directory. In 1.9 we changed to
// running the compiler in the build directory, to get consistent
// build results (issue #19964). To keep builds working, we change any
// relative -I or -L options to be absolute.
//
// Using filepath.IsAbs and filepath.Join here means the results will be
// different on different systems, but that's OK: -I and -L options are
// inherently system-dependent.
func (ctxt *Context) makePathsAbsolute(args []string, srcDir string) {
	nextPath := false
	for i, arg := range args {
		if nextPath {
			if !filepath.IsAbs(arg) {
				args[i] = filepath.Join(srcDir, arg)
			}
			nextPath = false
		} else if strings.HasPrefix(arg, "-I") || strings.HasPrefix(arg, "-L") {
			if len(arg) == 2 {
				nextPath = true
			} else {
				if !filepath.IsAbs(arg[2:]) {
					args[i] = arg[:2] + filepath.Join(srcDir, arg[2:])
				}
			}
		}
	}
}

// NOTE: $ is not safe for the shell, but it is allowed here because of linker options like -Wl,$ORIGIN.
// We never pass these arguments to a shell (just to programs we construct argv for), so this should be okay.
// See golang.org/issue/6038.
// The @ is for OS X. See golang.org/issue/13720.
// The % is for Jenkins. See golang.org/issue/16959.
// The ! is because module paths may use them. See golang.org/issue/26716.
const safeString = "+-.,/0123456789=ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz:$@%! "

func safeCgoName(s string) bool {
	if s == "" {
		return false
	}
	for i := 0; i < len(s); i++ {
		if c := s[i]; c < utf8.RuneSelf && strings.IndexByte(safeString, c) < 0 {
			return false
		}
	}
	return true
}

// splitQuoted splits the string s around each instance of one or more consecutive
// white space characters while taking into account quotes and escaping, and
// returns an array of substrings of s or an empty list if s contains only white space.
// Single quotes and double quotes are recognized to prevent splitting within the
// quoted region, and are removed from the resulting substrings. If a quote in s
// isn't closed err will be set and r will have the unclosed argument as the
// last element. The backslash is used for escaping.
//
// For example, the following string:
//
//     a b:"c d" 'e''f'  "g\""
//
// Would be parsed as:
//
//     []string{"a", "b:c d", "ef", `g"`}
//
func splitQuoted(s string) (r []string, err error) {
	var args []string
	arg := make([]rune, len(s))
	escaped := false
	quoted := false
	quote := '\x00'
	i := 0
	for _, rune := range s {
		switch {
		case escaped:
			escaped = false
		case rune == '\\':
			escaped = true
			continue
		case quote != '\x00':
			if rune == quote {
				quote = '\x00'
				continue
			}
		case rune == '"' || rune == '\'':
			quoted = true
			quote = rune
			continue
		case unicode.IsSpace(rune):
			if quoted || i > 0 {
				quoted = false
				args = append(args, string(arg[:i]))
				i = 0
			}
			continue
		}
		arg[i] = rune
		i++
	}
	if quoted || i > 0 {
		args = append(args, string(arg[:i]))
	}
	if quote != 0 {
		err = errors.New("unclosed quote")
	} else if escaped {
		err = errors.New("unfinished escaping")
	}
	return args, err
}

// match reports whether the name is one of:
//
//	$GOOS
//	$GOARCH
//	cgo (if cgo is enabled)
//	!cgo (if cgo is disabled)
//	ctxt.Compiler
//	!ctxt.Compiler
//	tag (if tag is listed in ctxt.BuildTags or ctxt.ReleaseTags)
//	!tag (if tag is not listed in ctxt.BuildTags or ctxt.ReleaseTags)
//	a comma-separated list of any of these
//
func (ctxt *Context) match(name string, allTags map[string]bool) bool {
	if name == "" {
		if allTags != nil {
			allTags[name] = true
		}
		return false
	}
	if i := strings.Index(name, ","); i >= 0 {
		// comma-separated list
		ok1 := ctxt.match(name[:i], allTags)
		ok2 := ctxt.match(name[i+1:], allTags)
		return ok1 && ok2
	}
	if strings.HasPrefix(name, "!!") { // bad syntax, reject always
		return false
	}
	if strings.HasPrefix(name, "!") { // negation
		return len(name) > 1 && !ctxt.match(name[1:], allTags)
	}

	if allTags != nil {
		allTags[name] = true
	}

	// Tags must be letters, digits, underscores or dots.
	// Unlike in Go identifiers, all digits are fine (e.g., "386").
	for _, c := range name {
		if !unicode.IsLetter(c) && !unicode.IsDigit(c) && c != '_' && c != '.' {
			return false
		}
	}

	// special tags
	if ctxt.CgoEnabled && name == "cgo" {
		return true
	}
	if name == ctxt.GOOS || name == ctxt.GOARCH || name == ctxt.Compiler {
		return true
	}
	if ctxt.GOOS == "android" && name == "linux" {
		return true
	}

	// other tags
	for _, tag := range ctxt.BuildTags {
		if tag == name {
			return true
		}
	}
	for _, tag := range ctxt.ReleaseTags {
		if tag == name {
			return true
		}
	}

	return false
}

// goodOSArchFile returns false if the name contains a $GOOS or $GOARCH
// suffix which does not match the current system.
// The recognized name formats are:
//
//     name_$(GOOS).*
//     name_$(GOARCH).*
//     name_$(GOOS)_$(GOARCH).*
//     name_$(GOOS)_test.*
//     name_$(GOARCH)_test.*
//     name_$(GOOS)_$(GOARCH)_test.*
//
// An exception: if GOOS=android, then files with GOOS=linux are also matched.
func (ctxt *Context) goodOSArchFile(name string, allTags map[string]bool) bool {
	if dot := strings.Index(name, "."); dot != -1 {
		name = name[:dot]
	}

	// Before Go 1.4, a file called "linux.go" would be equivalent to having a
	// build tag "linux" in that file. For Go 1.4 and beyond, we require this
	// auto-tagging to apply only to files with a non-empty prefix, so
	// "foo_linux.go" is tagged but "linux.go" is not. This allows new operating
	// systems, such as android, to arrive without breaking existing code with
	// innocuous source code in "android.go". The easiest fix: cut everything
	// in the name before the initial _.
	i := strings.Index(name, "_")
	if i < 0 {
		return true
	}
	name = name[i:] // ignore everything before first _

	l := strings.Split(name, "_")
	if n := len(l); n > 0 && l[n-1] == "test" {
		l = l[:n-1]
	}
	n := len(l)
	if n >= 2 && knownOS[l[n-2]] && knownArch[l[n-1]] {
		return ctxt.match(l[n-1], allTags) && ctxt.match(l[n-2], allTags)
	}
	if n >= 1 && (knownOS[l[n-1]] || knownArch[l[n-1]]) {
		return ctxt.match(l[n-1], allTags)
	}
	return true
}

var knownOS = make(map[string]bool)
var knownArch = make(map[string]bool)

func init() {
	for _, v := range strings.Fields(goosList) {
		knownOS[v] = true
	}
	for _, v := range strings.Fields(goarchList) {
		knownArch[v] = true
	}
}

// ToolDir is the directory containing build tools.
var ToolDir = getToolDir()

// IsLocalImport reports whether the import path is
// a local import path, like ".", "..", "./foo", or "../foo".
func IsLocalImport(path string) bool {
	return path == "." || path == ".." ||
		strings.HasPrefix(path, "./") || strings.HasPrefix(path, "../")
}

// ArchChar returns "?" and an error.
// In earlier versions of Go, the returned string was used to derive
// the compiler and linker tool names, the default object file suffix,
// and the default linker output name. As of Go 1.5, those strings
// no longer vary by architecture; they are compile, link, .o, and a.out, respectively.
func ArchChar(goarch string) (string, error) {
	return "?", errors.New("architecture letter no longer used")
}