Resolve committed module identity from Git objects

internal/git/path.go

@@ -0,0 +1,155 @@
+package git
+
+import (
+	"context"
+	"fmt"
+	"path"
+	"path/filepath"
+	"strings"
+)
+
+// PathResolver traverses committed paths, sharing reference rules between
+// release identity and validation. Lookups include both files and directories.
+type PathResolver struct {
+	RepositoryDir string
+	Lookup        func(context.Context, string) (TreeEntry, bool, error)
+	ReadLink      func(context.Context, TreeEntry) (string, error)
+}
+
+// DestinationKind classifies a committed local destination.
+type DestinationKind string
+
+const (
+	DestinationEntry           DestinationKind = "entry"
+	DestinationDirectory       DestinationKind = "directory"
+	DestinationMissing         DestinationKind = "missing"
+	DestinationOutsideCheckout DestinationKind = "outside-checkout"
+	DestinationSubmodule       DestinationKind = "submodule"
+	DestinationSymlinkCycle    DestinationKind = "symlink-cycle"
+)
+
+// Destination records the resolved identity and its classification.
+type Destination struct {
+	Kind DestinationKind
+	Path string
+}
+
+// ResolveDestination resolves local destinations using the captured Git tree.
+// Relative destinations start at the source directory; absolute destinations
+// within the repository directory map back to committed paths.
+func (resolver PathResolver) ResolveDestination(ctx context.Context, sourcePath, destination string) (Destination, error) {
+	if err := ctx.Err(); err != nil {
+		return Destination{}, err
+	}
+	directory := path.Dir(sourcePath)
+	if path.IsAbs(destination) {
+		relative, inside, err := resolver.relativeAbsoluteDestination(destination)
+		if err != nil {
+			return Destination{}, err
+		}
+		if !inside {
+			return Destination{Kind: DestinationOutsideCheckout, Path: destination}, nil
+		}
+		directory, destination = ".", relative
+	}
+	return resolver.resolveDestination(ctx, directory, strings.Split(destination, "/"), make(map[string]bool))
+}
+
+func (resolver PathResolver) resolveDestination(ctx context.Context, directory string, components []string, active map[string]bool) (Destination, error) {
+	for index, component := range components {
+		if err := ctx.Err(); err != nil {
+			return Destination{}, err
+		}
+		remaining := components[index+1:]
+		switch component {
+		case "", ".":
+			continue
+		case "..":
+			if directory == "." {
+				return Destination{Kind: DestinationOutsideCheckout, Path: unresolvedPath("..", remaining)}, nil
+			}
+			directory = path.Dir(directory)
+			continue
+		}
+		partial := path.Join(directory, component)
+		entry, exists, err := resolver.Lookup(ctx, partial)
+		if err != nil {
+			return Destination{}, err
+		}
+		if !exists {
+			return Destination{Kind: DestinationMissing, Path: unresolvedPath(partial, remaining)}, nil
+		}
+		if entry.Type == "tree" {
+			directory = partial
+			continue
+		}
+		if entry.Mode == "160000" {
+			return Destination{Kind: DestinationSubmodule, Path: unresolvedPath(partial, remaining)}, nil
+		}
+		if entry.Mode == "120000" {
+			if active[partial] {
+				return Destination{Kind: DestinationSymlinkCycle, Path: partial}, nil
+			}
+			linkText, err := resolver.ReadLink(ctx, entry)
+			if err != nil {
+				return Destination{}, fmt.Errorf("read committed symlink %q: %w", partial, err)
+			}
+			target := linkText
+			targetDirectory := directory
+			if path.IsAbs(target) {
+				relative, inside, err := resolver.relativeAbsoluteDestination(target)
+				if err != nil {
+					return Destination{}, err
+				}
+				if !inside {
+					return Destination{Kind: DestinationOutsideCheckout, Path: unresolvedPath(target, remaining)}, nil
+				}
+				targetDirectory, target = ".", relative
+			}
+			active[partial] = true
+			destination, err := resolver.resolveDestination(ctx, targetDirectory, strings.Split(target, "/"), active)
+			delete(active, partial)
+			if err != nil || len(remaining) == 0 {
+				return destination, err
+			}
+			switch destination.Kind {
+			case DestinationDirectory:
+				return resolver.resolveDestination(ctx, destination.Path, remaining, active)
+			case DestinationEntry:
+				destination.Kind = DestinationMissing
+			case DestinationSymlinkCycle:
+				return destination, nil
+			}
+			destination.Path = unresolvedPath(destination.Path, remaining)
+			return destination, nil
+		}
+		if len(remaining) != 0 {
+			return Destination{Kind: DestinationMissing, Path: unresolvedPath(partial, remaining)}, nil
+		}
+		return Destination{Kind: DestinationEntry, Path: partial}, nil
+	}
+	return Destination{Kind: DestinationDirectory, Path: directory}, nil
+}
+
+func (resolver PathResolver) relativeAbsoluteDestination(destination string) (string, bool, error) {
+	root, err := filepath.Abs(resolver.RepositoryDir)
+	if err != nil {
+		return "", false, fmt.Errorf("resolve candidate repository directory: %w", err)
+	}
+	root = filepath.ToSlash(root)
+	if destination == root {
+		return ".", true, nil
+	}
+	// Preserve the suffix for component traversal, including symlinks before '..'.
+	relative, inside := strings.CutPrefix(destination, strings.TrimSuffix(root, "/")+"/")
+	return relative, inside, nil
+}
+
+// unresolvedPath preserves components that could not be traversed. Cleaning
+// them would hide missing entries, non-directories, or checkout boundaries.
+func unresolvedPath(base string, remaining []string) string {
+	if len(remaining) == 0 {
+		return base
+	}
+	return base + "/" + strings.Join(remaining, "/")
+}

internal/git/source.go

@@ -123,6 +123,46 @@ func ReadCommitFile(ctx context.Context, repositoryDir, commit, name string) ([]
 	return []byte(content), nil
 }
 
+// CommittedPaths resolves individual paths from captured objects. It requests
+// only metadata and link contents needed for traversal.
+func CommittedPaths(repositoryDir, commit string) PathResolver {
+	metadata := make(map[string]TreeEntry)
+	return PathResolver{
+		RepositoryDir: repositoryDir,
+		Lookup: func(ctx context.Context, name string) (TreeEntry, bool, error) {
+			if err := ctx.Err(); err != nil {
+				return TreeEntry{}, false, err
+			}
+			if entry, seen := metadata[name]; seen {
+				return entry, entry.ObjectID != "", nil
+			}
+			if !validObjectID(commit) || name == "" || path.IsAbs(name) || path.Clean(name) != name || name == ".." || strings.HasPrefix(name, "../") {
+				return TreeEntry{}, false, fmt.Errorf("invalid committed path identity %q at %q", name, commit)
+			}
+			output, err := readLocalObjects(ctx, repositoryDir, nil, "--literal-pathspecs", "ls-tree", "-z", "--full-tree", commit, "--", name)
+			if err != nil {
+				return TreeEntry{}, false, fmt.Errorf("inspect committed path %q: %w", name, err)
+			}
+			entries, err := parseTreeEntries(output)
+			if err != nil {
+				return TreeEntry{}, false, err
+			}
+			for _, entry := range entries {
+				if entry.Path == name {
+					metadata[name] = entry
+					return entry, true, nil
+				}
+			}
+			metadata[name] = TreeEntry{}
+			return TreeEntry{}, false, nil
+		},
+		ReadLink: func(ctx context.Context, entry TreeEntry) (string, error) {
+			contents, err := ReadBlob(ctx, repositoryDir, entry.ObjectID)
+			return string(contents), err
+		},
+	}
+}
+
 // ReadTree enumerates all entries in a commit and records blob sizes. Paths
 // are NUL-delimited by Git so whitespace in a filename is preserved.
 func ReadTree(ctx context.Context, repositoryDir, commit string) ([]TreeEntry, error) {