io/fs: add Glob and GlobFS

Add Glob helper function, GlobFS interface, and test.
Add Glob method to fstest.MapFS.
Add testing of Glob method to fstest.TestFS.

For #41190.

Change-Id: If89dd7f63e310ba5ca2651340267a9ff39fcc0c7
Reviewed-on: https://go-review.googlesource.com/c/go/+/243915
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>

src/go/build/deps_test.go

@@ -123,7 +123,7 @@ var depsRules = `
 	< context
 	< TIME;
 
-	TIME, io, sort
+	TIME, io, path, sort
 	< io/fs;
 
 	# MATH is RUNTIME plus the basic math packages.

src/io/fs/glob.go

@@ -0,0 +1,116 @@
+// Copyright 2010 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 fs
+
+import (
+	"path"
+	"runtime"
+)
+
+// A GlobFS is a file system with a Glob method.
+type GlobFS interface {
+	FS
+
+	// Glob returns the names of all files matching pattern,
+	// providing an implementation of the top-level
+	// Glob function.
+	Glob(pattern string) ([]string, error)
+}
+
+// Glob returns the names of all files matching pattern or nil
+// if there is no matching file. The syntax of patterns is the same
+// as in path.Match. The pattern may describe hierarchical names such as
+// /usr/*/bin/ed (assuming the Separator is '/').
+//
+// Glob ignores file system errors such as I/O errors reading directories.
+// The only possible returned error is path.ErrBadPattern, reporting that
+// the pattern is malformed.
+//
+// If fs implements GlobFS, Glob calls fs.Glob.
+// Otherwise, Glob uses ReadDir to traverse the directory tree
+// and look for matches for the pattern.
+func Glob(fsys FS, pattern string) (matches []string, err error) {
+	if fsys, ok := fsys.(GlobFS); ok {
+		return fsys.Glob(pattern)
+	}
+
+	if !hasMeta(pattern) {
+		if _, err = Stat(fsys, pattern); err != nil {
+			return nil, nil
+		}
+		return []string{pattern}, nil
+	}
+
+	dir, file := path.Split(pattern)
+	dir = cleanGlobPath(dir)
+
+	if !hasMeta(dir) {
+		return glob(fsys, dir, file, nil)
+	}
+
+	// Prevent infinite recursion. See issue 15879.
+	if dir == pattern {
+		return nil, path.ErrBadPattern
+	}
+
+	var m []string
+	m, err = Glob(fsys, dir)
+	if err != nil {
+		return
+	}
+	for _, d := range m {
+		matches, err = glob(fsys, d, file, matches)
+		if err != nil {
+			return
+		}
+	}
+	return
+}
+
+// cleanGlobPath prepares path for glob matching.
+func cleanGlobPath(path string) string {
+	switch path {
+	case "":
+		return "."
+	default:
+		return path[0 : len(path)-1] // chop off trailing separator
+	}
+}
+
+// glob searches for files matching pattern in the directory dir
+// and appends them to matches, returning the updated slice.
+// If the directory cannot be opened, glob returns the existing matches.
+// New matches are added in lexicographical order.
+func glob(fs FS, dir, pattern string, matches []string) (m []string, e error) {
+	m = matches
+	infos, err := ReadDir(fs, dir)
+	if err != nil {
+		return // ignore I/O error
+	}
+
+	for _, info := range infos {
+		n := info.Name()
+		matched, err := path.Match(pattern, n)
+		if err != nil {
+			return m, err
+		}
+		if matched {
+			m = append(m, path.Join(dir, n))
+		}
+	}
+	return
+}
+
+// hasMeta reports whether path contains any of the magic characters
+// recognized by path.Match.
+func hasMeta(path string) bool {
+	for i := 0; i < len(path); i++ {
+		c := path[i]
+		if c == '*' || c == '?' || c == '[' || runtime.GOOS == "windows" && c == '\\' {
+			return true
+		}
+	}
+	return false
+}

src/io/fs/glob_test.go

@@ -0,0 +1,82 @@
+// Copyright 2009 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 fs_test
+
+import (
+	. "io/fs"
+	"os"
+	"testing"
+)
+
+var globTests = []struct {
+	fs              FS
+	pattern, result string
+}{
+	{os.DirFS("."), "glob.go", "glob.go"},
+	{os.DirFS("."), "gl?b.go", "glob.go"},
+	{os.DirFS("."), "*", "glob.go"},
+	{os.DirFS(".."), "*/glob.go", "fs/glob.go"},
+}
+
+func TestGlob(t *testing.T) {
+	for _, tt := range globTests {
+		matches, err := Glob(tt.fs, tt.pattern)
+		if err != nil {
+			t.Errorf("Glob error for %q: %s", tt.pattern, err)
+			continue
+		}
+		if !contains(matches, tt.result) {
+			t.Errorf("Glob(%#q) = %#v want %v", tt.pattern, matches, tt.result)
+		}
+	}
+	for _, pattern := range []string{"no_match", "../*/no_match"} {
+		matches, err := Glob(os.DirFS("."), pattern)
+		if err != nil {
+			t.Errorf("Glob error for %q: %s", pattern, err)
+			continue
+		}
+		if len(matches) != 0 {
+			t.Errorf("Glob(%#q) = %#v want []", pattern, matches)
+		}
+	}
+}
+
+func TestGlobError(t *testing.T) {
+	_, err := Glob(os.DirFS("."), "[]")
+	if err == nil {
+		t.Error("expected error for bad pattern; got none")
+	}
+}
+
+// contains reports whether vector contains the string s.
+func contains(vector []string, s string) bool {
+	for _, elem := range vector {
+		if elem == s {
+			return true
+		}
+	}
+	return false
+}
+
+type globOnly struct{ GlobFS }
+
+func (globOnly) Open(name string) (File, error) { return nil, ErrNotExist }
+
+func TestGlobMethod(t *testing.T) {
+	check := func(desc string, names []string, err error) {
+		t.Helper()
+		if err != nil || len(names) != 1 || names[0] != "hello.txt" {
+			t.Errorf("Glob(%s) = %v, %v, want %v, nil", desc, names, err, []string{"hello.txt"})
+		}
+	}
+
+	// Test that ReadDir uses the method when present.
+	names, err := Glob(globOnly{testFsys}, "*.txt")
+	check("readDirOnly", names, err)
+
+	// Test that ReadDir uses Open when the method is not present.
+	names, err = Glob(openOnly{testFsys}, "*.txt")
+	check("openOnly", names, err)
+}

src/testing/fstest/mapfs.go

@@ -128,6 +128,10 @@ func (fsys MapFS) ReadDir(name string) ([]fs.DirEntry, error) {
 	return fs.ReadDir(fsOnly{fsys}, name)
 }
 
+func (fsys MapFS) Glob(pattern string) ([]string, error) {
+	return fs.Glob(fsOnly{fsys}, pattern)
+}
+
 // A mapFileInfo implements fs.FileInfo and fs.DirEntry for a given map file.
 type mapFileInfo struct {
 	name string

src/testing/fstest/testfs.go

@@ -11,6 +11,8 @@ import (
 	"io"
 	"io/fs"
 	"io/ioutil"
+	"path"
+	"reflect"
 	"sort"
 	"strings"
 	"testing/iotest"
@@ -226,6 +228,8 @@ func (t *fsTester) checkDir(dir string) {
 			t.errorf("%s: fs.ReadDir: list not sorted: %s before %s", dir, list2[i].Name(), list2[i+1].Name())
 		}
 	}
+
+	t.checkGlob(dir, list)
 }
 
 // formatEntry formats an fs.DirEntry into a string for error messages and comparison.
@@ -243,6 +247,98 @@ func formatInfo(info fs.FileInfo) string {
 	return fmt.Sprintf("%s IsDir=%v Mode=%v Size=%d ModTime=%v", info.Name(), info.IsDir(), info.Mode(), info.Size(), info.ModTime())
 }
 
+// checkGlob checks that various glob patterns work if the file system implements GlobFS.
+func (t *fsTester) checkGlob(dir string, list []fs.DirEntry) {
+	if _, ok := t.fsys.(fs.GlobFS); !ok {
+		return
+	}
+
+	// Make a complex glob pattern prefix that only matches dir.
+	var glob string
+	if dir != "." {
+		elem := strings.Split(dir, "/")
+		for i, e := range elem {
+			var pattern []rune
+			for j, r := range e {
+				if r == '*' || r == '?' || r == '\\' || r == '[' {
+					pattern = append(pattern, '\\', r)
+					continue
+				}
+				switch (i + j) % 5 {
+				case 0:
+					pattern = append(pattern, r)
+				case 1:
+					pattern = append(pattern, '[', r, ']')
+				case 2:
+					pattern = append(pattern, '[', r, '-', r, ']')
+				case 3:
+					pattern = append(pattern, '[', '\\', r, ']')
+				case 4:
+					pattern = append(pattern, '[', '\\', r, '-', '\\', r, ']')
+				}
+			}
+			elem[i] = string(pattern)
+		}
+		glob = strings.Join(elem, "/") + "/"
+	}
+
+	// Try to find a letter that appears in only some of the final names.
+	c := rune('a')
+	for ; c <= 'z'; c++ {
+		have, haveNot := false, false
+		for _, d := range list {
+			if strings.ContainsRune(d.Name(), c) {
+				have = true
+			} else {
+				haveNot = true
+			}
+		}
+		if have && haveNot {
+			break
+		}
+	}
+	if c > 'z' {
+		c = 'a'
+	}
+	glob += "*" + string(c) + "*"
+
+	var want []string
+	for _, d := range list {
+		if strings.ContainsRune(d.Name(), c) {
+			want = append(want, path.Join(dir, d.Name()))
+		}
+	}
+
+	names, err := t.fsys.(fs.GlobFS).Glob(glob)
+	if err != nil {
+		t.errorf("%s: Glob(%#q): %v", dir, glob, err)
+		return
+	}
+	if reflect.DeepEqual(want, names) {
+		return
+	}
+
+	if !sort.StringsAreSorted(names) {
+		t.errorf("%s: Glob(%#q): unsorted output:\n%s", dir, glob, strings.Join(names, "\n"))
+		sort.Strings(names)
+	}
+
+	var problems []string
+	for len(want) > 0 || len(names) > 0 {
+		switch {
+		case len(want) > 0 && len(names) > 0 && want[0] == names[0]:
+			want, names = want[1:], names[1:]
+		case len(want) > 0 && (len(names) == 0 || want[0] < names[0]):
+			problems = append(problems, "missing: "+want[0])
+			want = want[1:]
+		default:
+			problems = append(problems, "extra: "+names[0])
+			names = names[1:]
+		}
+	}
+	t.errorf("%s: Glob(%#q): wrong output:\n%s", dir, glob, strings.Join(problems, "\n"))
+}
+
 // checkStat checks that a direct stat of path matches entry,
 // which was found in the parent's directory listing.
 func (t *fsTester) checkStat(path string, entry fs.DirEntry) {