4.3 KiB
Closed Saved Scenario spacing gate — c4a2cf6
This archive preserves the Windows listing and failed focused gate for c4a2cf63f11818cf7fc766acf3a2779768a2a658, followed by a separately classified full-inventory diagnostic of the same already-built executable. It is failure evidence, not spacing acceptance. Parent ran the stages; this archive task ran no Cargo, compiler, tests or native input.
| Stage | Actual result |
|---|---|
| List | Cargo 0; 1,082 unique tests listed; exact focused membership 288/288. |
| Focused | 286 passed, 2 failed, 0 ignored, 794 filtered; 117.64 test seconds; Cargo 101. |
| Full-inventory diagnostic | 1,076 passed, the same 2 failed, 4 ignored, 0 filtered; 765.34 test seconds; executable exit 101, runner exit 1. No compiler execution. |
| Accepted full / strict / GUI build | None. The failed focused gate was not waived; strict and GUI build were not run. |
List ran 2026-09-13 09:00:52–09:44:11 UTC; focused 09:44:53–09:47:03 UTC; diagnostic 09:57:24–10:10:10 UTC. The focused receipt's selectionValid=false records failed outcomes: independent archive verification confirms all selected 288 tests executed exactly once. This is not a missing-test or selection-path failure. The diagnostic was copied only after parent confirmed terminal closure and its final receipt was verified.
Recorded failures and separate correction
Both focused and diagnostic logs record:
- behavior::scenario::presentation_tests::complete_fallback_scene_selects_but_cannot_capture_or_finish_coordinate_gestures — presentation_tests.rs:202:5, assertion input.press.is_some().
- behavior::scenario::workspace::spacing_tests::saved_message_same_batch_pointer_events_commit_once_and_one_undo_redo_restores_exact_source — workspace/spacing_tests.rs:195:5, full-Project assertion. In the focused log, the sole Project difference is Saved A's updated_report send_y: actual 271.99994 versus expected 272.0; replacing that one value makes the complete debug strings identical.
0409a481c7571e887f8912f0a8cdb93cfdcc01a7 is a separate, unexecuted test-correction candidate at archive preparation. It does not change the recorded c4 source or convert these outcomes into passes. Corrected source requires its own acceptance gates.
Binding and scope
All five gate captures and three diagnostic captures contain the same 610 path/size/SHA-256 source records and nine external-input records. Raw list membership equals the expected 1,082 names; focused membership equals the selected 288; diagnostic membership equals the complete list, with exactly the same two failures. Archive verification checks stage-bound script/input hashes, raw-log and capture hashes, the focused-to-list receipt link, and diagnostic provenance links to the original gate.
The same desktop test executable is bound after list, before/after focused and before/after diagnostic: 150,361,088 bytes, SHA-256 cf31c674a7ef6c19dbe72f65c660030b5af6127d6450e35117ffc201ef813249. Diagnostic runtime libunwind.dll is 190,464 bytes, SHA-256 aa282fa47bb2c4e210cea5e470cdbdca58568f0be0bc73e404d86ab9cd95ef6a, identical before/after. Binaries are not copied here. The closed diagnostic receipt SHA-256 is c63e0ae818bd11a03ce68471f5cebe7d2ecce9a97dfbe73d7f006acede56b2da.
Server provenance retains the verified 57 server tests from 37be410 because core/server/manifests were unchanged. This attempt did not repeat server execution or establish strict acceptance. Preparation scripts and their machine paths describe the recorded environment; this archive is not a self-contained toolchain installation.
All 34 original gate files and 11 completed diagnostic files are preserved byte-for-byte, including preparation language and raw failure logs. manifest.json records sizes, SHA-256 and Git blob identities, plus the independent archive checks. Only archive README, manifest and binary-preserving .gitattributes are additional. No product source, original evidence, DECISIONS or older archive is changed.