erik/acdream
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions

docs(issues): #464 A/B round — no clip overflow, opaque piers, depth clear and alpha flush innocent, extra admissions innocent; the source is the cathedral's own content

Browse source 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-09-03 23:16:09 +02:00
parent 0b9e9b64ad
commit 720d04856c
1 changed files with 2 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

2
docs/ISSUES.md
View file

@ -121,6 +121,8 @@ then `/teleloc F4180113 39.502899 19.589222 181.119247 1 0 0 0`, zoom out, tilt
**SCRIPTED REPRODUCTION (lead, 2026-09-03 22:55, `logs/selfgate-20260903-225434-464-tilt3`, route `route-464-tilt3.txt`, the new `mouselook <dx> <dy>` automation verb landed at `ddae5704d` — one raw mouse-look sample per call, consumed while `CameraInstantMouseLook` is held; negative dy looks UP, about 0.0033 m of eye drop per unit):** at the owner's pose, sixteen samples of −20 walk the eye down from the default (39.50, 17.33, 183.18). **Frame `464-tilt3-01-neg160.png` (eye 182.61) SHOWS THE ARTIFACT** — the hall's structure and a hanging pier above and left of the arch, a pink doorway — while **frame `464-tilt3-02-neg240.png` (eye 182.32, the owner's height to a centimetre) is CLEAN and matches the retail screenshot.** Both frames: root `0xF4180114`, sweep uncontacted, no fallback. Their walk transcripts differ in ONE thing: the artifact frame's interior root has SEVEN outside views (`DC pv=0 ov=7`, 18 groups: `f3180105`, four `f4180104`, a lone `f4180109` first among the hall groups), the clean frame FIVE (`ov=5`, 14 groups, no `f3180105`, three `f4180104`, no lone `f4180109`). The owner's own frame and the retail capture at the same pose both have `ov=7` with retail drawing clean, so the two extra exit views are admitted correctly; what their content DOES on screen diverges. Prime suspect: the AD-17 clip-slot overflow — a view polygon with more than eight edges after screen clipping becomes a zero-plane slice and that view's punch fans draw UNCLIPPED (`ClipPlaneSet.From`: 'draw unclipped — still a superset of the polygon → over-include, safe' — NOT safe for a far-Z punch, which then punches depth to far outside the exit view). A replay dump of each outside view's edge count and overflow flag at both scripted eyes (client viewport 1760×990) is running.
**A/B round (lead, 2026-09-03 23:00–23:20, scratch worktree `464-ab` at `0b9e9b64a` with env-gated skips; route `route-464-tilt3.txt`, artifact frame `01-neg160`):** (1) the per-view clip dump (replay harness, viewport 1760×990) shows NO overflow at any of the three eyes — every outside view has 3–5 edges and exact planes — so AD-17's unclipped-fan fallback is NOT the mechanism. (2) Skipping EVERY delayed-alpha draw (`AB_SKIP_ALPHA`, both `DrawPreparedAlphaBatch` sites) leaves the artifact frame pixel-for-pixel the same: the hanging piers are OPAQUE geometry; the alpha-flush-versus-Z-clear hypothesis is dead. (3) Skipping the interior depth clear (`AB_SKIP_ZCLEAR`) also changes nothing: the depth-clear order is not it. (4) The S3 walk gate run with the retail `cathedral-stair-arch.walk` capture as a new row, at RETAIL'S exact eye (39.58, 16.99, 182.79, no pitch): our walk matches retail's 2,158-token sequence except one extra land cell (`LC/SC e2290001`) and the harness's known one-EC-per-cell limit (retail draws `f4180106` twice inside two groups; the live client's transcript does too) — no building/cell admission difference AT THAT EYE, and `f3180105` is absent on both sides there. But at the ARTIFACT eye (182.61, and the owner's 182.31 with a slight yaw) our live walk admits `DC pv=1 n=1: f3180105` (a cell of the building at `BLD f3180020` in the next landblock west), plus `BLD f3170034`/`f317003d` (two buildings two blocks west), plus a fourth `f4180104` group; the clean frame (182.32, no yaw) admits none of those. No retail capture exists at the artifact eye, so whether retail refuses them there is unproven; the targeted A/B decides whether their CONTENT is the artifact: skipping the draw of cell f3180105, of the two far-block building shells (f3170034, f317003d), or of cell f4180104 each leaves the artifact frame pixel-identical — NONE of the extra admissions is the artifact's source; the piers come from the cathedral's own content (the next split: the stairwell tower shell 0x01001FB7 versus the corridor cell 0x109 versus the whole hall group, running)
## #463 — Soak `turn` samples can report a 0.1 ms GPU p50 (the turn window catches non-world frames)
**Status:** OPEN — found 2026-09-03 by the Campaign OVERHAUL v2 S3 review's perf re-measure.