Files
goldeneye-pc-port/docs/dev
James Dansereau a24495672e docs: scrub stale status claims (D77 music, D227 sky, D217 green textures) (#55)
* docs: scrub stale status claims (D77 music, D227 sky, D217 green textures)

The M-104 roadmap refresh flagged four live inaccuracies; this fixes all
four plus files the missing music-quality finding:

- README: in-level music no longer described as silent/absent (D77 fixed,
  PR #42); sky line no longer 'still being fixed' (D227 merged, PR #43) --
  D229 green water is now the named sky-path residual.
- GRAPHICS-BACKLOG.md: D77 row updated PARTIAL -> RESOLVED with root cause.
- findings.md: D217 status cell led 'OPEN -- root cause characterised
  (M-86)' although the green defect was fixed and merged (PR #47/#49, two
  human playtests); now leads FIXED with the D228 split-out noted.
- New D230: in-level music quality residual (wrong instruments on some
  tracks -- elevator named specifically -- plus occasional garbling).
  Scope pointed at the seq/instrument-bank decode path, not the proven
  mixer/trigger layers.

findings-index.csv regenerated; gen_findings_index.py --check passes.

* docs(D190): root-caused and closed — level 45 is MP-only, no solo setup in ROM

M-48's 'Linux-only crash' was a misframing. Full chain (reproduced on
Windows this session, 0xc0000005 in modelLoad via domakedefaultobj with
obj=0x784f):

- setup_text_pointers[45] = "UsetupimpZ" but that name is absent from the
  ROM file table; only the MP variant "Ump_setupimpZ" exists. BASEMENT is
  a 4P MP level (front.c multi_stage_setups[]) — its solo name was never
  shipped, like the other MP-only levels (STACK/TEMPLE/COMPLEX/CAVES/
  LIBRARY).
- A solo -level_45 boot requests the nonexistent file. On PC the inert
  rmon stubs leave fileGetIndex's dynamic-path response uninitialized ->
  size-0 table entry -> 0-byte load into a stale pool buffer -> the
  propDef walk runs over stale memory (M-48 hit add_ptr_to_objective;
  same garbage walk, different contents).
- N64 parity: file genuinely absent there too; the boot is broken in the
  original as well. No fix taken (a port guard cannot synthesize a valid
  setup; fabricating ROM content has no ground truth).
2026-09-12 09:19:04 -04:00
..
2026-09-01 19:28:59 -04:00