fix(ingestion): brain/raw/ readers go through an os.Root; symlinks are inert (#153) (#156)

Co-authored-by: Mathias <mthbqv@gmail.com>

ingestion/internal/api/ingestpath_root.go

@@ -2,6 +2,9 @@ package api
 
 import (
 	"errors"
+	"os"
+
+	"github.com/mathiasbq/hyperguild/ingestion/internal/extract"
 	"path/filepath"
 	"strings"
 )
@@ -19,6 +22,11 @@ func ResolveIngestPath(brainDir, p string) (string, error) {
 	if strings.TrimSpace(p) == "" {
 		return "", ErrOutsideIngestRoot
 	}
+	// raw/ itself must be a real directory: if it were a symlink, containment
+	// inside it would mean nothing (hyperguild#153 item 4).
+	if fi, lerr := os.Lstat(filepath.Join(brainDir, "raw")); lerr != nil || fi.Mode()&os.ModeSymlink != 0 || !fi.IsDir() {
+		return "", ErrOutsideIngestRoot
+	}
 	root, err := filepath.EvalSymlinks(filepath.Join(brainDir, "raw"))
 	if err != nil {
 		return "", ErrOutsideIngestRoot
@@ -36,3 +44,24 @@ func ResolveIngestPath(brainDir, p string) (string, error) {
 	}
 	return real, nil
 }
+
+// ReadIngestFile extracts text from an ingest path already confined by
+// ResolveIngestPath, reading it through the raw/ root rather than reopening
+// the path by name (hyperguild#153 item 3): if the file is swapped for a
+// symlink after the check, the read still cannot leave raw/.
+func ReadIngestFile(brainDir, resolved string) (string, error) {
+	root, rawPath, err := OpenRawRoot(brainDir)
+	if err != nil {
+		return "", ErrOutsideIngestRoot
+	}
+	defer func() { _ = root.Close() }()
+	realRoot, err := filepath.EvalSymlinks(rawPath)
+	if err != nil {
+		return "", ErrOutsideIngestRoot
+	}
+	rel, err := filepath.Rel(realRoot, resolved)
+	if err != nil || rel == ".." || strings.HasPrefix(rel, ".."+string(filepath.Separator)) || filepath.IsAbs(rel) {
+		return "", ErrOutsideIngestRoot
+	}
+	return extract.TextIn(root, rel)
+}