mithril-go/internal/artifact/extract.go
Kayos e557d85d5a download + extract pipeline
- artifact.Download: resumable HTTP with optional SHA256 check + progress cb
- artifact.ExtractZstdTar: streamed zstd+tar with tar-slip defense
- aggregator client matches real API shape (digests/immutables/ancillary blocks
  with URIHolder polymorphism for templated immutable URIs)
- cmd: show + download subcommands wired up
- end-to-end verified against preprod: digests archive pulls cleanly, yields
  16836-entry SHA manifest ready for verification sprint

deps: github.com/klauspost/compress (pure-go zstd)
2026-04-23 15:16:48 -07:00

91 lines
2.3 KiB
Go

package artifact
import (
"archive/tar"
"context"
"fmt"
"io"
"os"
"path/filepath"
"strings"
"github.com/klauspost/compress/zstd"
)
// ExtractZstdTar decompresses a .tar.zst archive into targetDir, streaming
// through the reader without buffering the full archive. Refuses entries
// with ".." in the path or absolute paths (tar-slip defense).
func ExtractZstdTar(ctx context.Context, archivePath, targetDir string) error {
f, err := os.Open(archivePath)
if err != nil {
return fmt.Errorf("open archive: %w", err)
}
defer f.Close()
zr, err := zstd.NewReader(f)
if err != nil {
return fmt.Errorf("zstd reader: %w", err)
}
defer zr.Close()
tr := tar.NewReader(zr)
if err := os.MkdirAll(targetDir, 0o755); err != nil {
return fmt.Errorf("mkdir target: %w", err)
}
cleanTarget, err := filepath.Abs(targetDir)
if err != nil {
return err
}
for {
if err := ctx.Err(); err != nil {
return err
}
hdr, err := tr.Next()
if err == io.EOF {
break
}
if err != nil {
return fmt.Errorf("tar next: %w", err)
}
// tar-slip defense
clean := filepath.Clean(hdr.Name)
if strings.HasPrefix(clean, "..") || filepath.IsAbs(clean) {
return fmt.Errorf("refusing suspicious archive path: %s", hdr.Name)
}
outPath := filepath.Join(cleanTarget, clean)
if !strings.HasPrefix(filepath.Clean(outPath)+string(os.PathSeparator), cleanTarget+string(os.PathSeparator)) &&
filepath.Clean(outPath) != cleanTarget {
return fmt.Errorf("refusing path outside target: %s", hdr.Name)
}
switch hdr.Typeflag {
case tar.TypeDir:
if err := os.MkdirAll(outPath, os.FileMode(hdr.Mode)); err != nil {
return err
}
case tar.TypeReg:
if err := os.MkdirAll(filepath.Dir(outPath), 0o755); err != nil {
return err
}
out, err := os.OpenFile(outPath, os.O_CREATE|os.O_WRONLY|os.O_TRUNC, os.FileMode(hdr.Mode))
if err != nil {
return fmt.Errorf("create %s: %w", outPath, err)
}
if _, err := io.Copy(out, tr); err != nil {
out.Close()
return fmt.Errorf("write %s: %w", outPath, err)
}
if err := out.Close(); err != nil {
return err
}
case tar.TypeSymlink, tar.TypeLink:
// Refuse links for safety — a Mithril archive has no legitimate reason to contain them.
return fmt.Errorf("refusing link entry: %s", hdr.Name)
default:
// Silently skip unknown types.
}
}
return nil
}