This repository has been archived on 2026-09-16. You can view files and clone it. You cannot open issues or pull requests or push a commit.

Sylpheed RE agent 3ecb3432f1 docs: the freeze was caught and the stuck-wait probe reports nothing new ...

The fourth run froze 9 seconds into the watcher's window, in flight, and the
probe built for that moment showed the healthy-run baseline and nothing else: one
pair, the same poller on the same object VA as every healthy run, only the thread
handle differing. No new (thread, object) pair appeared.

So the hypothesis the probe was built to catch is refuted - the freeze is not a
guest thread looping on KeWaitForSingleObject timeouts against ONE object - while
the CPU signature is unchanged from the gdb run: 1255 ticks over 10 s, 401 in the
TimerQueue thread and 292/280 in two guest threads.

What survives is stated as two specific blind spots of the instrument rather than
a shrug: the waits may cycle over DIFFERENT objects, which resets the streak and
makes them invisible to a same-object counter; or they may SUCCEED rather than
time out, which leaves a timeout counter nothing to count and would fit the
kernel-log evidence of a self-suspending worker cycling thousands of times
successfully.

Next is a v2 that counts calls per thread per second regardless of object or
result. The freeze lottery paid out on the first attempt this time.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PMRJjbxLqZtsb5Vb7KunPE

2026-08-24 07:40:09 +00:00

..

hud-runtime

re: in-flight HUD — independent confirmation of the ammo mapping

2026-07-28 18:50:46 +00:00

ui-layout

formats: a third measured paint order — the main menu, with two primitives

2026-08-19 08:56:21 +00:00

weapon-datasheet

re: runtime DATA SHEET reads the live weapon record (Route B, first capture)

2026-07-28 18:01:25 +00:00

achievements-xach.txt

re(challenge): the gate's bit space is the game's 24 ACHIEVEMENTS

2026-08-13 19:32:43 +00:00

arsenal-all-owned-br-cannon.png

re: the develop blob's index space is SOLVED -- it follows strings.tbl, cut items and all

2026-08-11 19:01:14 +00:00

arsenal-categories-all.png

re: the Arsenal develop economy, and the save blob's index space (first 32 confirmed)

2026-08-11 17:38:45 +00:00

arsenal-develop-conditions.csv

re: the Arsenal develop economy, and the save blob's index space (first 32 confirmed)

2026-08-11 17:38:45 +00:00

arsenal-gun-below-fold.png

re: the Arsenal develop economy, and the save blob's index space (first 32 confirmed)

2026-08-11 17:38:45 +00:00

arsenal-gun-list-predevelop.png

re: developing one Arsenal weapon separates Points from its twin and names the save blob

2026-08-11 17:16:57 +00:00

arsenal-layout-validation.png

re: validate the screen layout against the running game -- eight rows, +/-2 px

2026-08-11 22:25:18 +00:00

arsenal-main-screen-layout.txt

re: placement X/Y are SIGNED, the tail word is time, and the resting pose is max-dwell

2026-08-11 22:14:36 +00:00

arsenal-mg1-developed.png

re: developing one Arsenal weapon separates Points from its twin and names the save blob

2026-08-11 17:16:57 +00:00

arsenal-probe-blob39-tomahawk.png

re: the develop blob's index space is SOLVED -- it follows strings.tbl, cut items and all

2026-08-11 19:01:14 +00:00

arsenal-probe-blob48-jamming.png

re: the develop blob's index space is SOLVED -- it follows strings.tbl, cut items and all

2026-08-11 19:01:14 +00:00

arsenal-probe-save-result.png

re: saves can be written back, which settles what the develop blob's 4s mean and refutes the tail

2026-08-11 18:10:28 +00:00

autopilot-300s-undamaged-stage02.png

re: the autopilot now survives, and kills

2026-07-30 05:38:00 +00:00

autopilot-escort-warning-stage02.png

re: surviving is not winning -- Stage 02 is an escort

2026-07-30 05:43:42 +00:00

autopilot-first-kill-stage02.png

re: the autopilot now survives, and kills

2026-07-30 05:38:00 +00:00

axis-probe-rows-pinned.csv

re(flight): rows pinned by world-Y, and nothing yaws -- the yaw gap is answered

2026-08-13 22:35:36 +00:00

axis-probe-stage02.csv

re(flight): axis probe -- lx is roll, ly drives one clean axis, rest not trustworthy

2026-08-13 22:10:23 +00:00

boot-signed-in-no-swallow.log

tools: fix four boot-harness bugs; find the swallowed-keystroke cause and refute it

2026-08-19 08:21:08 +00:00

cache-flush-throw-cold-cache.log

docs/re: the cache-flush throw — a 100-second trigger, and two withdrawals

2026-08-18 21:44:22 +00:00

challenge-extras-after-poke.png

re(challenge): the cleared-stage mask is CONFIRMED on the running game

2026-08-13 20:28:19 +00:00

challenge-map.txt

re(challenge): the EX missions are a MODE, not a stage number

2026-08-13 18:59:52 +00:00

challenge-screen-config.txt

re(challenge): the unlock is a bit test, and the mission table is on disc

2026-08-13 19:14:48 +00:00

ctrl-probe-stage02.csv

re: the autopilot now survives, and kills

2026-07-30 05:38:00 +00:00

difficulty-screen.png

docs/re: the cache-flush crash also fires at SELECT DATA, blocking every mission

2026-08-19 04:00:53 +00:00

disc-unit-signatures.txt

re(route-b): 13 defaulted Size_Y values read out of the running game

2026-08-12 20:28:55 +00:00

dpad-tactical-map.png

docs/re: in-flight control mapping, and why lethality is an aim-dwell problem

2026-07-30 18:49:16 +00:00

e106-cover-slack-before-after.png

fix(mesh): require exact pool coverage -- found by rendering, not by a metric

2026-08-12 06:56:49 +00:00

e106-static-assembly-fixed.png

fix(mesh): anchor XBG7 resources by neighbourhood -- inconsistency 125 -> 63, ships render right

2026-08-12 00:31:20 +00:00

e106-static-assembly-volume-bug.png

re: the format layer is NOT exonerated -- assemble_ship draws a non-renderable volume

2026-08-11 23:15:22 +00:00

escort-decay-ab.csv

re: making turrets targets does not save the escort — the effect is inside the noise

2026-08-11 05:29:52 +00:00

escort-stage02-hud.png

docs/re: the hull anchor is class-wide — the escort objective is scoreable live

2026-07-30 17:21:05 +00:00

f101-index-shift-before-after.png

re(xbg7): render the before/after — the index shift shredded capital-ship hulls

2026-08-13 07:28:40 +00:00

filepad-flight-stage01.png

re(flight): START skips the briefing, and the player-entity lock is stage-specific

2026-08-13 21:45:54 +00:00

filepad-mainmenu.png

re(challenge): the cleared-stage mask is CONFIRMED on the running game

2026-08-13 20:28:19 +00:00

fire-probe-ammo-counters.png

pilot: RB fires, Y is the main mount, and guided missiles produce the first kills

2026-07-30 18:40:39 +00:00

first-missile-kills.png

pilot: RB fires, Y is the main mount, and guided missiles produce the first kills

2026-07-30 18:40:39 +00:00

hangar-loadout-screen.png

re: capture the Hangar screen — it is a loadout picker, not a craft data sheet

2026-08-11 17:19:21 +00:00

kill-counters-all-runs.png

pilot: real ballistics from the solved Shell records — and the metric that says it did not help

2026-07-30 18:16:26 +00:00

kills-target-commitment.png

pilot: target commitment takes kills from 2 to 9; and the target pointer does not exist

2026-07-30 19:19:23 +00:00

load-game-oracle.png

formats: two more measured paint orders, and the first independent confirmation

2026-08-19 11:46:09 +00:00

loadgame-slot01-headerless-install.png

docs: the dynamic-RE state was gone - rebuild it, measured rather than remembered

2026-08-23 17:15:53 +00:00

main-menu-draw-capture.png

docs/re: a second screen captured, and the two title states pinned down

2026-08-18 23:10:04 +00:00

main-menu-draw-order.log

docs/re: a second screen captured, and the two title states pinned down

2026-08-18 23:10:04 +00:00

main-menu-oracle.png

formats: a third measured paint order — the main menu, with two primitives

2026-08-19 08:56:21 +00:00

main-menu-paint-order.txt

docs/re: a second screen captured, and the two title states pinned down

2026-08-18 23:10:04 +00:00

main-menu-reached.png

docs/re: the main menu HAS been reached — the Ⓐ blocker is withdrawn

2026-08-18 22:18:51 +00:00

mission01-t485.png

re: Stage 02 is lost at ~11 minutes, and the pilot's survival rule is what guarantees it

2026-08-10 20:57:43 +00:00

mission03-turretrule-t430.png

re: making turrets targets does not save the escort — the effect is inside the noise

2026-08-11 05:29:52 +00:00

mission-select-all-story-unlocked.png

re(challenge): poking unlocks the whole story campaign; MISSION SELECT is story-only

2026-08-13 20:54:57 +00:00

mission-select-derived-order.png

formats: derive the paint order from the sprite layer key

2026-08-19 05:28:05 +00:00

mission-select-ends-at-stage16.png

re(challenge): poking unlocks the whole story campaign; MISSION SELECT is story-only

2026-08-13 20:54:57 +00:00

mission-select-stage01-only.png

re(challenge): MISSION SELECT renders the cleared-stage mask, and a control run

2026-08-13 20:44:27 +00:00

mission-select-stage01-record.png

re: SHAB is the per-stage record table -- record 0 is Stage 01, and c is its best time

2026-08-11 20:50:06 +00:00

mission-state-stage02-escort-weighted.jsonl

pilot: escort-weighted targeting (DEFEND), plus the capital-ship keep-out it needed

2026-07-30 17:49:38 +00:00

mission-state-stage02.jsonl

docs/re: the hull anchor is class-wide — the escort objective is scoreable live

2026-07-30 17:21:05 +00:00

movie-subtitle-voice-map.csv

re: the movie/subtitle/voice link is solved statically -- INDEX called it a dynamic unknown

2026-08-11 21:10:08 +00:00

ob-flag-refuted-stage02.json

docs+tools: the per-entity OB flag is refuted, and the counter decrements when the player kills

2026-08-23 23:21:08 +00:00

ob-hunt-stage02-run4.json

docs: REMAINING OB is at 0xbdb59668 - confirmed to the corpus's own standard

2026-08-23 18:39:22 +00:00

ob-hunt-stage02-run4.log

docs: REMAINING OB is at 0xbdb59668 - confirmed to the corpus's own standard

2026-08-23 18:39:22 +00:00

options-key-config-actions.png

docs/re: the game's own in-flight action list, off the disc

2026-07-30 18:54:16 +00:00

pause-tutorial-screen-layout.txt

re: placement X/Y are SIGNED, the tail word is time, and the resting pose is max-dwell

2026-08-11 22:14:36 +00:00

pitch-burst-final.csv

re(flight): both axes at settled speeds -- shape confirmed, absolute scale is not

2026-08-13 23:42:28 +00:00

pitch-burst-settled-speeds.csv

re(flight): short bursts confirm the rate-vs-speed SHAPE, without needing the clock

2026-08-13 23:29:23 +00:00

pitch-gametime.csv

re(flight): the clock factor is universal — angular matches too once corrected in-run

2026-08-13 17:56:16 +00:00

pitch-rate-curve-clean.csv

re(flight): clean pitch sweep -- magnitudes agree, the interpolation law does not

2026-08-13 23:12:50 +00:00

pitch-rate-speed-bleed.csv

re(flight): the rate probe measures a MOVING speed -- pitching bleeds it hard

2026-08-13 22:48:29 +00:00

rate-curve-aliased-BAD.csv

re(flight): polling faster than the guest updates manufactures a clean curve

2026-08-13 23:00:59 +00:00

rate-curve-windowed.csv

re(flight): polling faster than the guest updates manufactures a clean curve

2026-08-13 23:00:59 +00:00

ready-room-extra-mode.png

re(units): S10 closes the story campaign for real -- 69 units, 7 204 values

2026-08-13 21:14:21 +00:00

readyroom-take-off-armed.png

tools: launch_mission reaches Stage 02 flight unattended again - wait for screens, not seconds

2026-08-23 18:02:12 +00:00

resume-fix-reliability.csv

docs+tools: the threading fix measured 5/5 against 1/5, and a driver lock

2026-08-19 11:12:46 +00:00

roll-about-forward-axis.csv

re(flight): roll DOES depend on speed -- the withdrawal is reversed

2026-08-13 21:57:10 +00:00

roll-burst-final.csv

re(flight): both axes at settled speeds -- shape confirmed, absolute scale is not

2026-08-13 23:42:28 +00:00

roll-gametime.csv

re(flight): withdraw the roll result — the probe was measuring pitch as much as roll

2026-08-13 18:12:20 +00:00

roll-ramp-inconclusive.csv

re(flight): the ramp test is inconclusive, and names the tool the residual needs

2026-08-13 23:53:23 +00:00

savedata-game02-samestate.bin

re: a second save names Points, flight time and clear ratio — and shows the payload is pure state

2026-08-11 06:02:21 +00:00

savedata-game03-developed-mg1.bin

re: developing one Arsenal weapon separates Points from its twin and names the save blob

2026-08-11 17:16:57 +00:00

savedata-stage02-5pct.bin

re: the save file is a GDHA/zlib chunk stream — layout read off the title's own serializer

2026-08-11 05:40:26 +00:00

savegame-details-slot02.png

re: a second save names Points, flight time and clear ratio — and shows the payload is pure state

2026-08-11 06:02:21 +00:00

savegame-details-slot03-developed.png

re: developing one Arsenal weapon separates Points from its twin and names the save blob

2026-08-11 17:16:57 +00:00

savegame-stage05-probe-details.png

re(savegame): GHAD +52 is the stage number — a hand-edited save plays Stage 05

2026-08-13 09:46:08 +00:00

savegame-stage05-probe-inflight.png

re(savegame): GHAD +52 is the stage number — a hand-edited save plays Stage 05

2026-08-13 09:46:08 +00:00

select-data-crash.png

docs/re: the cache-flush crash also fires at SELECT DATA, blocking every mission

2026-08-19 04:00:53 +00:00

select-data-reached-no-crash.png

docs/re: the mission path measured end to end — the cache crash is the blocker

2026-08-19 05:08:46 +00:00

ship-capture-no-geometry.log

tools+docs: a crash-free mission run, a capture that caught no geometry, and a correction

2026-08-19 12:56:35 +00:00

shipcap-close-f105-2994.png

re: control the range, segment the frames — f105/e105/e106 all match static assembly

2026-08-10 18:19:26 +00:00

shipcap-stage02-launch.png

re: ship-placement generalisation — static audit + first Stage-02 F10 capture (WIP)

2026-07-31 05:16:20 +00:00

speed-law-throttle-phases.csv

re(flight): the throttle is a target-speed selector, measured against the definition

2026-08-13 15:08:11 +00:00

stage02-briefing-crash.png

tools: repair launch_mission.sh; Stage 02 now reaches the briefing and stops there

2026-08-19 13:27:46 +00:00

stage02-defaulted-field-runtime-values.txt

re(route-b): map generalises to units -- 406/0 across 11 objects, 121 defaulted fields

2026-08-13 00:12:45 +00:00

stage02-flight-live-dump.png

re: live unit definitions from a running Stage 02, and the recipe that gets there

2026-08-12 20:18:32 +00:00

stage02-freeze-at-time-0102.png

docs: the in-mission freeze ends in 1171 refused resumes of one thread

2026-08-23 19:36:03 +00:00

stage02-freeze-at-time-0337.png

docs: the resume-spin lead is refuted by its own control, and the freeze is a guest-side spin

2026-08-23 20:03:38 +00:00

stage02-freeze-cpu-while-frozen.txt

docs: the resume-spin lead is refuted by its own control, and the freeze is a guest-side spin

2026-08-23 20:03:38 +00:00

stage02-freeze-gdb-backtraces.txt

tools+docs: the freeze from inside - all 79 threads are in a wait, and two of them burn CPU cycling

2026-08-24 01:13:01 +00:00

stage02-freeze-kernel-tail.log

docs: the resume-spin lead is refuted by its own control, and the freeze is a guest-side spin

2026-08-23 20:03:38 +00:00

stage02-freeze-resume-spin.log

docs: the in-mission freeze ends in 1171 refused resumes of one thread

2026-08-23 19:36:03 +00:00

stage02-freeze-stuck-wait-probe.txt

docs: the freeze was caught and the stuck-wait probe reports nothing new

2026-08-24 07:40:09 +00:00

stage02-freeze-thread-cpu.txt

docs: the in-mission freeze ends in 1171 refused resumes of one thread

2026-08-23 19:36:03 +00:00

stage02-inflight-restored-state.png

docs: the dynamic-RE state was gone - rebuild it, measured rather than remembered

2026-08-23 17:15:53 +00:00

stage02-inflight-run2-ob-address-zero.png

docs: the REMAINING OB address does not survive a run - refuted twice over

2026-08-23 17:16:07 +00:00

stage02-inflight-unattended-launch.png

tools: launch_mission reaches Stage 02 flight unattended again - wait for screens, not seconds

2026-08-23 18:02:12 +00:00

stage02-inflight-upstream-baseline.png

docs: a stock-upstream baseline runs Stage 02 crash-free

2026-08-19 18:44:23 +00:00

stage02-inmission-freeze.png

docs+tools: a mission freeze that made an experiment lie, and the counter is not a class head-count

2026-08-23 19:10:41 +00:00

stage02-live-unit-definitions-deep.txt

re: a deeper window does not help -- the 0x820af844 object is not the definition record

2026-08-12 20:49:19 +00:00

stage02-live-unit-definitions.txt

re: live unit definitions from a running Stage 02, and the recipe that gets there

2026-08-12 20:18:32 +00:00

stage02-pilot-firing-after-sign-fix.png

tools+docs: the pitch stick sign was inverted - the pilot fires now

2026-08-23 22:39:33 +00:00

stage02-you-killed-warplanes-0003.png

docs+tools: the per-entity OB flag is refuted, and the counter decrements when the player kills

2026-08-23 23:21:08 +00:00

stage05-mission-approach.png

re(xbg7): a Stage-05 capture validates the pad fix out of sample, and shows how containers load

2026-08-13 10:20:00 +00:00

stage-s01-capture-truth-offsets.txt

re: the proven offsets ARE candidates -- these mis-anchors are selection failures

2026-08-12 02:11:41 +00:00

stage-s02-capture-truth-offsets.txt

fix(mesh): winding consistency replaces the connectivity heuristic

2026-08-12 05:47:01 +00:00

stage-s02-index-buffer-truth.txt

re(xbg7): the [index][vertex] layout is runtime-verified, and the indices= mystery was batching

2026-08-13 05:57:44 +00:00

throttle-curve-hud.csv

re(flight): the linear discrepancy is a world-unit vs displayed-speed difference

2026-08-13 17:06:35 +00:00

throttle-curve-lt.csv

re(flight): LT mirrors RT, and roll does not depend on speed

2026-08-13 16:30:19 +00:00

throttle-curve-rt.csv

re(flight): the throttle is analogue — target speed interpolates cruise -> maximum

2026-08-13 16:14:47 +00:00

title-build4-decl-words.txt

docs/re: the title paint order is build 4's, carries no depth, and is not in the file

2026-08-18 20:16:05 +00:00

title-composited-measured-order.png

formats: composite in the MEASURED paint order, and drop the ghost instances

2026-08-19 02:32:14 +00:00

title-draw-order-with-alpha.log

docs/re: alpha does not break the tie — but it validates the fade decode

2026-08-19 01:09:25 +00:00

title-input-delivered-not-acted-on.log

docs: a fourth paint order attempted; the title is intermittent, not input-bound

2026-08-19 09:56:32 +00:00

title-loader-thread-never-runs.log

tools+docs: the title's loader thread is created and never runs

2026-08-19 10:19:06 +00:00

title-loader-thread-stalls.log

docs: the title does act on (A) - the loader thread it spawns is what stalls

2026-08-19 10:03:41 +00:00

title-screen-composite-2026-08-18.png

docs/re: the title-screen oracle — what it settled, and what it refuted

2026-08-18 16:32:52 +00:00

title-screen-draw-order.log

docs/re: the title paint order is build 4's, carries no depth, and is not in the file

2026-08-18 20:16:05 +00:00

title-screen-oracle.png

docs/re: the title-screen oracle — what it settled, and what it refuted

2026-08-18 16:32:52 +00:00

title-screen-paint-order.txt

docs/re: the title paint order is build 4's, carries no depth, and is not in the file

2026-08-18 20:16:05 +00:00

title-signin-dialog.png

docs/re: why a scripted run cannot get past the title — three traps and a blocker

2026-08-18 20:16:27 +00:00

title-unresponsive-attract.png

docs/re: the scripted route to the menu does not work — measured, not assumed

2026-08-18 22:45:08 +00:00

title-with-emulator-menu-open.png

docs/re: why a scripted run cannot get past the title — three traps and a blocker

2026-08-18 20:16:27 +00:00

title-with-press-a-live.png

tools: two dead flags that made every emulator boot script fail silently

2026-08-19 07:48:16 +00:00

turn-law-pitch-phases.csv

re(flight): turn rates confirm AV_* are rate caps and _Min/_Max mean at min/max speed

2026-08-13 15:30:00 +00:00

turn-law-roll.csv

re(flight): LT mirrors RT, and roll does not depend on speed

2026-08-13 16:30:19 +00:00

turn-law-settled.csv

re(flight): settled turn rates match the definition exactly — withdraw the time-base claim

2026-08-13 16:46:33 +00:00

tutorial-advanced-controls-captions.png

re-capture: read the ADVANCED CONTROLS tutorial; its moves regress the pilot, so they ship off

2026-07-30 19:50:25 +00:00

tutorial-hud-target-select.png

pilot: target select is A pressed TWICE — from the HUD tutorial, not from probing

2026-07-30 20:47:54 +00:00

tutorial-lesson-descriptions.png

docs: map where the autopilot's objective and action knowledge lives on the disc

2026-08-19 13:13:33 +00:00

tutorial-lesson-list.png

docs: map where the autopilot's objective and action knowledge lives on the disc

2026-08-19 13:13:33 +00:00

tutorial-menu-labels.png

docs/re: the game's own in-flight action list, off the disc

2026-07-30 18:54:16 +00:00

tutorial-mission-clean-run.png

tools+docs: a crash-free mission run, a capture that caught no geometry, and a correction

2026-08-19 12:56:35 +00:00

tutorial-mission-reached-then-crash.png

docs: the mission path is reachable now - it freezes there instead of dying earlier

2026-08-19 12:05:23 +00:00

unit-runtime-fields.csv

re(units): S10 closes the story campaign for real -- 69 units, 7 204 values

2026-08-13 21:14:21 +00:00

weapon-runtime-fields.csv

re: read the defaulted weapon stats out of live guest memory

2026-07-29 05:29:34 +00:00

xenia-profile-panel-signed-in.png

docs/re: why a scripted run cannot get past the title — three traps and a blocker

2026-08-18 20:16:27 +00:00

xpr2-colour-check-deltasaber.png

re: XPR2 channel order confirmed against the running game -- the second "dynamic unknown" closed

2026-08-11 21:21:55 +00:00