Alex Vanderveen 03e4d6be56 cd_diff: anchor on the palette file's Setlocs, not the Getstat heartbeat ...

The first live run aligned on the game's periodic Getstat poll -- every
30 frames, identical everywhere -- and reported a fork between native
mid-scene and the oracle still at boot (its tail showed lbas 304, 2864,
12422: boot loads; the scene had not been played there yet).

Anchor on Setlocs into the palette file's LBA range instead, compare the
two LBA request sequences positionally, and print each side's full
command window around the file with the first diverged request marked.
A side with no palette-file Setlocs in its history is told to play the
scene there rather than diffed against unrelated traffic.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-08-24 12:55:58 -04:00

..

test_cd_diff.py

cd_diff: anchor on the palette file's Setlocs, not the Getstat heartbeat

2026-08-24 12:55:58 -04:00

test_cd_fork.py

cd_fork: check the disc's own sectors against what landed in RAM

2026-08-24 12:55:58 -04:00

test_cd_verify.py

cd_verify: reconstruct the game's CD command conversation

2026-08-24 12:55:58 -04:00

test_class_census.py

census: report WHERE a missing class is built, not just how many

2026-08-24 12:55:58 -04:00

test_colour_inputs.py

colour_inputs: ask ONE emulator whether the region is phase-driven

2026-08-24 12:55:58 -04:00

test_colour_source.py

colour_source: read the oracle's source pointer via pc_break

2026-08-24 12:55:58 -04:00

test_disasm.py

disasm: name the unaligned and cop2 forms the colour path is built from

2026-08-24 12:55:57 -04:00

test_display_list.py

dlist: compare colour STRUCTURE, which phase cannot fake

2026-08-24 12:55:57 -04:00

test_duckstation_oracle.py

oracle: share the game's memory card, so it can reach the scene

2026-08-24 12:55:58 -04:00

test_effect_palette.py

effect_palette: read 256 KB per sample, not 2 MB

2026-08-24 12:55:58 -04:00

test_fingerprint_diff.py

oracle: port frame_fingerprint, add first-divergence diff

2026-08-24 12:55:57 -04:00

test_gpu_frame.py

gpu_parity: align the images by VRAM origin, and stop faking frame alignment

2026-08-24 12:55:58 -04:00

test_lockstep.py

lockstep: say what a skip reason MEANS, not just how many

2026-08-24 12:55:57 -04:00

test_packet_writers_dbuf.py

packet_writers: trace both halves of the double buffer

2026-08-24 12:55:58 -04:00

test_packet_writers.py

gte: point the instruments at the colour bug; they clear the GTE

2026-08-24 12:55:57 -04:00

test_park_guard.py

packet_writers: hand the emulator back on EVERY path

2026-08-24 12:55:58 -04:00

test_phase_parity.py

effect_palette: read both prim shapes; verify breakpoint removal

2026-08-24 12:55:58 -04:00

test_probe_regs.py

probe: sweep for the block leader instead of guessing one window

2026-08-24 12:55:58 -04:00

test_ram_parity.py

debug: restore pause/step, repaint hold, and the GPU analysis tools

2026-08-24 12:55:57 -04:00

test_range_writers_watch.py

range_writers: --watch for data written once at scene load

2026-08-24 12:55:58 -04:00

test_range_writers.py

range_writers: attribute writes over an arbitrary span, as a Studio button

2026-08-24 12:55:58 -04:00

test_scale_trace.py

scale_trace: always say HOW the numbers were taken; trust adjacent frames

2026-08-24 12:55:58 -04:00

test_scale_within_frame.py

scale_within_frame: refuse the constant-within-frame test for scratch regs

2026-08-24 12:55:58 -04:00

test_slice_check.py

slice_check: count colour-word writes per frame

2026-08-24 12:55:58 -04:00

test_stall_report.py

tools: harden stall report evidence semantics

2026-08-11 09:37:32 +03:00

test_table_watch.py

table_watch: catch the table changing, whatever changes it

2026-08-24 12:55:58 -04:00

test_tool_imports.py

census: the images show missing OBJECTS, not a mis-rendered glow

2026-08-24 12:55:58 -04:00

test_vertex_taps.py

vertex_taps: read what the colour routine loads, per vertex, one frame

2026-08-24 12:55:58 -04:00

test_wait_for_class.py

wait_for_class: make the wait affordable, and un-park before polling

2026-08-24 12:55:58 -04:00

test_walk_coherence.py

walk_side: a scoped read is the picture; stop re-reading to verify it

2026-08-24 12:55:58 -04:00

test_wedge_scan.py

wedge_scan: locate the effect by additive count; scan the whole ring

2026-08-24 12:55:58 -04:00