-
Notifications
You must be signed in to change notification settings - Fork 570
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Signed-off-by: Alex Goodman <[email protected]>
- Loading branch information
Showing
7 changed files
with
642 additions
and
8 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,220 @@ | ||
| package file | ||
|
|
||
| import ( | ||
| "archive/tar" | ||
| "fmt" | ||
| "io" | ||
| "os" | ||
| "path/filepath" | ||
| "strings" | ||
| "time" | ||
|
|
||
| "github.com/spf13/afero" | ||
| "github.com/xi2/xz" | ||
| ) | ||
|
|
||
| // Note: this is a copy of the TarXzDecompressor from https://github.com/hashicorp/go-getter/blob/v2.2.3/decompress_txz.go | ||
| // with the xz lib swapped out (for performance). A few adjustments were made: | ||
| // - refactored to use afero filesystem abstraction | ||
| // - fixed some linting issues | ||
|
|
||
| // TarXzDecompressor is an implementation of Decompressor that can | ||
| // decompress tar.xz files. | ||
| type tarXzDecompressor struct { | ||
| // FileSizeLimit limits the total size of all | ||
| // decompressed files. | ||
| // | ||
| // The zero value means no limit. | ||
| FileSizeLimit int64 | ||
|
|
||
| // FilesLimit limits the number of files that are | ||
| // allowed to be decompressed. | ||
| // | ||
| // The zero value means no limit. | ||
| FilesLimit int | ||
|
|
||
| Fs afero.Fs | ||
| } | ||
|
|
||
| func (d *tarXzDecompressor) Decompress(dst, src string, dir bool, umask os.FileMode) error { | ||
| // If we're going into a directory we should make that first | ||
| mkdir := dst | ||
| if !dir { | ||
| mkdir = filepath.Dir(dst) | ||
| } | ||
| if err := d.Fs.MkdirAll(mkdir, mode(0755, umask)); err != nil { | ||
| return err | ||
| } | ||
|
|
||
| // File first | ||
| f, err := d.Fs.Open(src) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| defer f.Close() | ||
|
|
||
| // xz compression is second | ||
| txzR, err := xz.NewReader(f, 0) | ||
| if err != nil { | ||
| return fmt.Errorf("error opening an xz reader for %s: %s", src, err) | ||
| } | ||
|
|
||
| return untar(d.Fs, txzR, dst, src, dir, umask, d.FileSizeLimit, d.FilesLimit) | ||
| } | ||
|
|
||
| // untar is a shared helper for untarring an archive. The reader should provide | ||
| // an uncompressed view of the tar archive. | ||
| func untar(fs afero.Fs, input io.Reader, dst, src string, dir bool, umask os.FileMode, fileSizeLimit int64, filesLimit int) error { // nolint:funlen,gocognit | ||
| tarR := tar.NewReader(input) | ||
| done := false | ||
| dirHdrs := []*tar.Header{} | ||
| now := time.Now() | ||
|
|
||
| var ( | ||
| fileSize int64 | ||
| filesCount int | ||
| ) | ||
|
|
||
| for { | ||
| if filesLimit > 0 { | ||
| filesCount++ | ||
| if filesCount > filesLimit { | ||
| return fmt.Errorf("tar archive contains too many files: %d > %d", filesCount, filesLimit) | ||
| } | ||
| } | ||
|
|
||
| hdr, err := tarR.Next() | ||
| if err == io.EOF { | ||
| if !done { | ||
| // Empty archive | ||
| return fmt.Errorf("empty archive: %s", src) | ||
| } | ||
|
|
||
| break | ||
| } | ||
| if err != nil { | ||
| return err | ||
| } | ||
|
|
||
| switch hdr.Typeflag { | ||
| case tar.TypeSymlink, tar.TypeLink: | ||
| // to prevent any potential indirect traversal attacks | ||
| continue | ||
| case tar.TypeXGlobalHeader, tar.TypeXHeader: | ||
| // don't unpack extended headers as files | ||
| continue | ||
| } | ||
|
|
||
| path := dst | ||
| if dir { | ||
| // Disallow parent traversal | ||
| if containsDotDot(hdr.Name) { | ||
| return fmt.Errorf("entry contains '..': %s", hdr.Name) | ||
| } | ||
|
|
||
| path = filepath.Join(path, hdr.Name) // nolint:gosec // hdr.Name is checked above | ||
| } | ||
|
|
||
| fileInfo := hdr.FileInfo() | ||
|
|
||
| fileSize += fileInfo.Size() | ||
|
|
||
| if fileSizeLimit > 0 && fileSize > fileSizeLimit { | ||
| return fmt.Errorf("tar archive larger than limit: %d", fileSizeLimit) | ||
| } | ||
|
|
||
| if fileInfo.IsDir() { | ||
| if !dir { | ||
| return fmt.Errorf("expected a single file: %s", src) | ||
| } | ||
|
|
||
| // A directory, just make the directory and continue unarchiving... | ||
| if err := fs.MkdirAll(path, mode(0755, umask)); err != nil { | ||
| return err | ||
| } | ||
|
|
||
| // Record the directory information so that we may set its attributes | ||
| // after all files have been extracted | ||
| dirHdrs = append(dirHdrs, hdr) | ||
|
|
||
| continue | ||
| } | ||
| // There is no ordering guarantee that a file in a directory is | ||
| // listed before the directory | ||
| dstPath := filepath.Dir(path) | ||
|
|
||
| // Check that the directory exists, otherwise create it | ||
| if _, err := fs.Stat(dstPath); os.IsNotExist(err) { | ||
| if err := fs.MkdirAll(dstPath, mode(0755, umask)); err != nil { | ||
| return err | ||
| } | ||
| } | ||
|
|
||
| // We have a file. If we already decoded, then it is an error | ||
| if !dir && done { | ||
| return fmt.Errorf("expected a single file, got multiple: %s", src) | ||
| } | ||
|
|
||
| // Mark that we're done so future in single file mode errors | ||
| done = true | ||
|
|
||
| // Size limit is tracked using the returned file info. | ||
| err = copyReader(fs, path, tarR, hdr.FileInfo().Mode(), umask, 0) | ||
| if err != nil { | ||
| return err | ||
| } | ||
|
|
||
| // Set the access and modification time if valid, otherwise default to current time | ||
| aTime := now | ||
| mTime := now | ||
| if hdr.AccessTime.Unix() > 0 { | ||
| aTime = hdr.AccessTime | ||
| } | ||
| if hdr.ModTime.Unix() > 0 { | ||
| mTime = hdr.ModTime | ||
| } | ||
| if err := fs.Chtimes(path, aTime, mTime); err != nil { | ||
| return err | ||
| } | ||
| } | ||
|
|
||
| // Perform a final pass over extracted directories to update metadata | ||
| for _, dirHdr := range dirHdrs { | ||
| path := filepath.Join(dst, dirHdr.Name) // nolint:gosec // hdr.Name is checked above | ||
| // Chmod the directory since they might be created before we know the mode flags | ||
| if err := fs.Chmod(path, mode(dirHdr.FileInfo().Mode(), umask)); err != nil { | ||
| return err | ||
| } | ||
| // Set the mtime/atime attributes since they would have been changed during extraction | ||
| aTime := now | ||
| mTime := now | ||
| if dirHdr.AccessTime.Unix() > 0 { | ||
| aTime = dirHdr.AccessTime | ||
| } | ||
| if dirHdr.ModTime.Unix() > 0 { | ||
| mTime = dirHdr.ModTime | ||
| } | ||
| if err := fs.Chtimes(path, aTime, mTime); err != nil { | ||
| return err | ||
| } | ||
| } | ||
|
|
||
| return nil | ||
| } | ||
|
|
||
| // containsDotDot checks if the filepath value v contains a ".." entry. | ||
| // This will check filepath components by splitting along / or \. This | ||
| // function is copied directly from the Go net/http implementation. | ||
| func containsDotDot(v string) bool { | ||
| if !strings.Contains(v, "..") { | ||
| return false | ||
| } | ||
| for _, ent := range strings.FieldsFunc(v, isSlashRune) { | ||
| if ent == ".." { | ||
| return true | ||
| } | ||
| } | ||
| return false | ||
| } | ||
|
|
||
| func isSlashRune(r rune) bool { return r == '/' || r == '\\' } |
Oops, something went wrong.