Original Penguin Command matched firing test, 2026-10-24

Historical source last updated, 2026-10-24. Runtime and implementation claims below describe that dated source; see current app notes and source for subsequent changes.

Blocked: A was consumed and dispatched, but this matched pair does not establish visible firing causality. Exactly two guest runs were performed; no third attempt, rebuild, dependency recovery, or fabricated state.

The control first passed the required full cannon/ammunition, non-menu baseline at batch 39999. Its image SHA256 is 95112041c84196db00873e676117ac51d58562a21cabe06f5498d1f4b0504972, also the accepted prior ordinary-play batch-40000 image. The treatment's 39999 image is byte-identical. All three cannons, ammunition groups, six cities, right border and full bottom edge are visible within the 840x660 desktop capture. Crosshair is at (260,220). This is an authentic ordinary game reached through its menu, not a fixture or replacement asset.

Matched commands and actual delivery

The adjacent evidence directory retains both executed argv, input recipes, full production dispatch logs, command and terminal receipts, images, remote image hashes, comparison code and resource/source receipts. Recipes use their original absolute paths and must not overwrite existing run outputs.

Both commands used the standing READONLY runner/runtime/game under penguin-integrated-clock-firing-oct24 on bx_fkg5uw8h. Only new recipes and evidence used penguin-postrelease-matched-firing-oct24. The only input difference was balanced A-down40000 / A-up40200. Both used normal mouse move33000, menu down34000/up34500 and aim34800; no gameplay mouse button was added. Identical captures:39999,40020,40100,40201,40500,41000. No final PNG was requested; all twelve remote images remain retained and charged.

Event Actual production observation
Menu down, both runs Dispatch34004; guestMs6800800; lastGuestTick6800803
Menu release, both Dispatch37629; guestMs7525800; lastGuestTick7525801
Aim, both Dispatch37629; guestMs7525800; lastGuestTick7525803
A-down, treatment Dequeue6 and DispatchMessageA40032; WM_KEYDOWN0x100, wParam0x41, hwnd0x10001; guestMs8006400; lastGuestTick8006402 at dequeue,8006404 at dispatch
A-up, treatment Dequeue7 and DispatchMessageA40282; WM_KEYUP0x101, wParam0x41, hwnd0x10001; guestMs8056400; lastGuestTick8056402 at dequeue,8056404 at dispatch

The actual consumption times differ from queue schedules. In particular, 40020 precedes A-down consumption, and40201 precedes A-up consumption. The trace explicitly leaves renderer enqueue IDs and schedule-to-dispatch correlation unavailable. The production calls and matching key fields are retained; scheduled events alone were not used as delivery proof.

Pixel evidence and conclusion

All comparisons decode the existing PNGs with the standing validated pngjs; no image is altered. pixel-comparison.json records every matched pair:

Batch Different pixels Different pixels at y>=550 Interpretation
39999 0 0 Matched full ordinary-play baseline
40020 0 0 Before actual A-down dispatch
40100 0 0 First captured frame after A-down; no treatment-only visible response
40201 0 0 Still identical; A-up not yet consumed
40500 1 0 Only pixel(82,536); no ammunition-area change
41000 1010 4 Sparse later differences across bbox(29,48)-(578,599); insufficient to identify a projectile, blast or ammo decrement

The baseline,40100,40500 and41000 collected frames were personally viewed at their full captured extent. At40100 the images show descending enemy paths and city destruction in both runs. Those common effects cannot establish firing. The single40500 pixel changes from RGBA[45,40,7,255] to[55,9,9,255]. Such a difference is not a projectile trajectory, blast or ammunition decrement. By40500 the left ammunition group has disappeared in both runs alongside ordinary enemy impacts; that common depletion is not attributed to A. Later small differences are not sufficient to distinguish a firing response from changes in ordinary animation/timing. Score and wave transitions are not used as causal evidence.

The precise remaining blocker is no distinguishable treatment-only immediate projectile/blast/ammunition response in the sampled temporal envelope despite proven A dispatch. This does not prove the authentic A binding is broken. The original README's A/S/D binding evidence is reused. No alternative input, new capture window or further guest run was attempted after this ambiguity.