[GH-ISSUE #495] [Feature] Decode Foundation keyed archives into object-reference graphs #265

New issue

Closed

opened 2026-10-15 18:23:12 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-15 18:23:12 +00:00

Owner

Originally created by @morluto on GitHub (Oct 2, 2026).
Original GitHub issue: https://github.com/morluto/rea/issues/495

Area

Evidence, artifacts, or reconstruction

Current limitation

REA can normalize property lists and decode compiled Interface Builder archives into UI graphs. General Foundation NSKeyedArchiver payloads remain a property-list-shaped object table: callers must manually resolve $top, $objects, CF$UID references, class descriptors, conditional links, and cycles to understand the saved model graph.

Use cases

Inspect application save files and persisted model state without launching the application.
Understand nested model ownership and identity links in local or cloud archive payloads.
Compare serialized object graphs across app versions while keeping build-specific schema claims explicit.

Desired outcome

Provide a read-only, bounded decoder for Foundation keyed archives that returns roots, typed object nodes, keyed/array edges, class descriptors, and UID/reference edges with cycle handling. Preserve original archive bytes and artifact provenance. Never instantiate application classes or run decoding code from the target. Unrecognized classes and unsupported value types should remain explicit unknown nodes rather than disappearing or being coerced to empty values.

Success criteria

Deterministic fixtures cover binary and XML property-list keyed archives, nested dictionaries/arrays, shared and cyclic references, conditional references, class descriptors, missing roots, and malformed UIDs.
Results distinguish archived null/nil, missing keys, unresolved references, unsupported types, and decoder failures.
Limits and truncation counts are exact; original artifact digest and source path/object identity are retained.
A selected root can be inspected without implicitly extracting or modifying files.
Tests verify the parser never loads or executes archived classes.

Relationship to existing capabilities

This complements inspect_plist and decode_interface_builder. Interface Builder decoding is a specialized archive interpretation; this capability should expose the reusable Foundation object/reference graph primitive and allow higher-level formats to add semantic projections later.

Originally created by @morluto on GitHub (Oct 2, 2026). Original GitHub issue: https://github.com/morluto/rea/issues/495 ## Area Evidence, artifacts, or reconstruction ## Current limitation REA can normalize property lists and decode compiled Interface Builder archives into UI graphs. General Foundation `NSKeyedArchiver` payloads remain a property-list-shaped object table: callers must manually resolve `$top`, `$objects`, `CF$UID` references, class descriptors, conditional links, and cycles to understand the saved model graph. ## Use cases - Inspect application save files and persisted model state without launching the application. - Understand nested model ownership and identity links in local or cloud archive payloads. - Compare serialized object graphs across app versions while keeping build-specific schema claims explicit. ## Desired outcome Provide a read-only, bounded decoder for Foundation keyed archives that returns roots, typed object nodes, keyed/array edges, class descriptors, and UID/reference edges with cycle handling. Preserve original archive bytes and artifact provenance. Never instantiate application classes or run decoding code from the target. Unrecognized classes and unsupported value types should remain explicit unknown nodes rather than disappearing or being coerced to empty values. ## Success criteria - Deterministic fixtures cover binary and XML property-list keyed archives, nested dictionaries/arrays, shared and cyclic references, conditional references, class descriptors, missing roots, and malformed UIDs. - Results distinguish archived null/nil, missing keys, unresolved references, unsupported types, and decoder failures. - Limits and truncation counts are exact; original artifact digest and source path/object identity are retained. - A selected root can be inspected without implicitly extracting or modifying files. - Tests verify the parser never loads or executes archived classes. ## Relationship to existing capabilities This complements `inspect_plist` and `decode_interface_builder`. Interface Builder decoding is a specialized archive interpretation; this capability should expose the reusable Foundation object/reference graph primitive and allow higher-level formats to add semantic projections later.

bsakaguchi

2026-10-15 18:23:12 +00:00

closed this issue
added the
triage

enhancement
labels

bsakaguchi commented

2026-10-15 18:23:13 +00:00

Author

Owner

@morluto commented on GitHub (Oct 8, 2026):

Verified fixed on main (7a7460f0). KeyedArchiveReader.ts + domain/apple/keyedArchive.ts ship object-reference graph decoding with keyedArchive.test.ts passing. Closing per audit.

@morluto commented on GitHub (Oct 8, 2026): Verified fixed on main (7a7460f0). KeyedArchiveReader.ts + domain/apple/keyedArchive.ts ship object-reference graph decoding with keyedArchive.test.ts passing. Closing per audit.

bsakaguchi referenced this issue

2026-10-16 21:12:29 +00:00

[PR #274] [MERGED] feat(mcp)!: harden contracts and Evidence references #1918

bsakaguchi added

enhancement

and removed

enhancement

labels

2026-10-31 00:28:57 +00:00