godoc: support for file systems stored in .zip files

Instead of serving files of the underlying OS file system,
a .zip file may be provided to godoc containing the files
to serve; for instance:

   godoc -http=:6060 -zip=go.zip

using a .zip file created from a clean tree as follows:

   zip -r go.zip $GOROOT

R=rsc
CC=golang-dev
https://golang.org/cl/4670053

src/cmd/godoc/doc.go

@@ -73,6 +73,8 @@ The flags are:
 		filter file containing permitted package directory paths
 	-filter_minutes=0
 		filter file update interval in minutes; update is disabled if <= 0
+	-zip=""
+		zip file providing the file system to serve; disabled if empty
 
 The -path flag accepts a list of colon-separated paths; unrooted paths are relative
 to the current working directory. Each path is considered as an additional root for

src/cmd/godoc/filesystem.go

@@ -2,11 +2,14 @@
 // Use of this source code is governed by a BSD-style
 // license that can be found in the LICENSE file.
 
-// This file defines abstract file system access.
+// This file defines types for abstract file system access and
+// provides an implementation accessing the file system of the
+// underlying OS.
 
 package main
 
 import (
+	"fmt"
 	"io"
 	"io/ioutil"
 	"os"
@@ -62,7 +65,18 @@ func (fi osFI) Size() int64 {
 type osFS struct{}
 
 func (osFS) Open(path string) (io.ReadCloser, os.Error) {
-	return os.Open(path)
+	f, err := os.Open(path)
+	if err != nil {
+		return nil, err
+	}
+	fi, err := f.Stat()
+	if err != nil {
+		return nil, err
+	}
+	if fi.IsDirectory() {
+		return nil, fmt.Errorf("Open: %s is a directory", path)
+	}
+	return f, nil
 }
 
 
@@ -79,7 +93,7 @@ func (osFS) Stat(path string) (FileInfo, os.Error) {
 
 
 func (osFS) ReadDir(path string) ([]FileInfo, os.Error) {
-	l0, err := ioutil.ReadDir(path)
+	l0, err := ioutil.ReadDir(path) // l0 is sorted
 	if err != nil {
 		return nil, err
 	}

src/cmd/godoc/main.go

@@ -26,6 +26,7 @@
 package main
 
 import (
+	"archive/zip"
 	"bytes"
 	_ "expvar" // to serve /debug/vars
 	"flag"
@@ -47,6 +48,10 @@ import (
 const defaultAddr = ":6060" // default webserver address
 
 var (
+	// file system to serve
+	// (with e.g.: zip -r go.zip $GOROOT -i \*.go -i \*.html -i \*.css -i \*.js -i \*.txt -i \*.c -i \*.h -i \*.s -i \*.png -i \*.jpg -i \*.sh)
+	zipfile = flag.String("zip", "", "zip file providing the file system to serve; disabled if empty")
+
 	// periodic sync
 	syncCmd   = flag.String("sync", "", "sync command; disabled if empty")
 	syncMin   = flag.Int("sync_minutes", 0, "sync interval in minutes; disabled if <= 0")
@@ -223,13 +228,19 @@ func main() {
 	flag.Usage = usage
 	flag.Parse()
 
-	// Determine file system to use.
-	// TODO(gri) Complete this - for now we only have one.
-	fs = OS
-
 	// Clean goroot: normalize path separator.
 	*goroot = filepath.Clean(*goroot)
 
+	// Determine file system to use.
+	fs = OS
+	if *zipfile != "" {
+		rc, err := zip.OpenReader(*zipfile)
+		if err != nil {
+			log.Fatalf("%s: %s\n", *zipfile, err)
+		}
+		fs = NewZipFS(rc)
+	}
+
 	// Check usage: either server and no args, or command line and args
 	if (*httpAddr != "") != (flag.NArg() == 0) {
 		usage()

src/cmd/godoc/zip.go

@@ -0,0 +1,199 @@
+// 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.
+
+// This file provides an implementation of the FileSystem
+// interface based on the contents of a .zip file.
+//
+// Assumptions:
+//
+// - The file paths stored in the zip file must use a slash ('/') as path
+//   separator; and they must be relative (i.e., they must not start with
+//   a '/' - this is usually the case if the file was created w/o special
+//   options).
+// - The zip file system treats the file paths found in the zip internally
+//   like absolute paths w/o a leading '/'; i.e., the paths are considered
+//   relative to the root of the file system.
+// - All path arguments to file system methods must be absolute paths.
+
+package main
+
+import (
+	"archive/zip"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"path"
+	"sort"
+	"strings"
+)
+
+
+// zipFI is the zip-file based implementation of FileInfo
+type zipFI struct {
+	name string    // directory-local name
+	file *zip.File // nil for a directory
+}
+
+
+func (fi zipFI) Name() string {
+	return fi.name
+}
+
+
+func (fi zipFI) Size() int64 {
+	if fi.file != nil {
+		return int64(fi.file.UncompressedSize)
+	}
+	return 0 // directory
+}
+
+
+func (fi zipFI) IsDirectory() bool {
+	return fi.file == nil
+}
+
+
+func (fi zipFI) IsRegular() bool {
+	return fi.file != nil
+}
+
+
+// zipFS is the zip-file based implementation of FileSystem
+type zipFS struct {
+	*zip.ReadCloser
+	list zipList
+}
+
+
+func (fs *zipFS) Close() os.Error {
+	fs.list = nil
+	return fs.ReadCloser.Close()
+}
+
+
+func zipPath(name string) string {
+	if !path.IsAbs(name) {
+		panic(fmt.Sprintf("stat: not an absolute path: %s", name))
+	}
+	return name[1:] // strip '/'
+}
+
+
+func (fs *zipFS) stat(abspath string) (int, zipFI, os.Error) {
+	i := fs.list.lookup(abspath)
+	if i < 0 {
+		return -1, zipFI{}, fmt.Errorf("file not found: %s", abspath)
+	}
+	var file *zip.File
+	if abspath == fs.list[i].Name {
+		file = fs.list[i] // exact match found - must be a file
+	}
+	_, name := path.Split(abspath)
+	return i, zipFI{name, file}, nil
+}
+
+
+func (fs *zipFS) Open(abspath string) (io.ReadCloser, os.Error) {
+	_, fi, err := fs.stat(zipPath(abspath))
+	if err != nil {
+		return nil, err
+	}
+	if fi.IsDirectory() {
+		return nil, fmt.Errorf("Open: %s is a directory", abspath)
+	}
+	return fi.file.Open()
+}
+
+
+func (fs *zipFS) Lstat(abspath string) (FileInfo, os.Error) {
+	_, fi, err := fs.stat(zipPath(abspath))
+	return fi, err
+}
+
+
+func (fs *zipFS) Stat(abspath string) (FileInfo, os.Error) {
+	_, fi, err := fs.stat(zipPath(abspath))
+	return fi, err
+}
+
+
+func (fs *zipFS) ReadDir(abspath string) ([]FileInfo, os.Error) {
+	path := zipPath(abspath)
+	i, fi, err := fs.stat(path)
+	if err != nil {
+		return nil, err
+	}
+	if !fi.IsDirectory() {
+		return nil, fmt.Errorf("ReadDir: %s is not a directory", abspath)
+	}
+
+	var list []FileInfo
+	dirname := path + "/"
+	prevname := ""
+	for _, e := range fs.list[i:] {
+		if !strings.HasPrefix(e.Name, dirname) {
+			break // not in the same directory anymore
+		}
+		name := e.Name[len(dirname):] // local name
+		file := e
+		if i := strings.IndexRune(name, '/'); i >= 0 {
+			// We infer directories from files in subdirectories.
+			// If we have x/y, return a directory entry for x.
+			name = name[0:i] // keep local directory name only
+			file = nil
+		}
+		// If we have x/y and x/z, don't return two directory entries for x.
+		// TODO(gri): It should be possible to do this more efficiently
+		// by determining the (fs.list) range of local directory entries
+		// (via two binary searches).
+		if name != prevname {
+			list = append(list, zipFI{name, file})
+			prevname = name
+		}
+	}
+
+	return list, nil
+}
+
+
+func (fs *zipFS) ReadFile(abspath string) ([]byte, os.Error) {
+	rc, err := fs.Open(abspath)
+	if err != nil {
+		return nil, err
+	}
+	return ioutil.ReadAll(rc)
+}
+
+
+func NewZipFS(rc *zip.ReadCloser) FileSystem {
+	list := make(zipList, len(rc.File))
+	copy(list, rc.File) // sort a copy of rc.File
+	sort.Sort(list)
+	return &zipFS{rc, list}
+}
+
+
+type zipList []*zip.File
+
+// zipList implements sort.Interface
+func (z zipList) Len() int           { return len(z) }
+func (z zipList) Less(i, j int) bool { return z[i].Name < z[j].Name }
+func (z zipList) Swap(i, j int)      { z[i], z[j] = z[j], z[i] }
+
+
+// lookup returns the first index in the zipList
+// of a path equal to name or beginning with name/.
+func (z zipList) lookup(name string) int {
+	i := sort.Search(len(z), func(i int) bool {
+		return name <= z[i].Name
+	})
+	if i >= 0 {
+		iname := z[i].Name
+		if strings.HasPrefix(iname, name) && (len(name) == len(iname) || iname[len(name)] == '/') {
+			return i
+		}
+	}
+	return -1 // no match
+}