Refresh the roadmap for the landed console palette resolution #262

Merged

jqueuniet merged 0 commits from refs/pull/262/head into main

2026-07-17 14:46:45 +02:00

jqueuniet commented

2026-07-17 14:36:05 +02:00

(Migrated from codeberg.org)

Docs-only follow-up to #259 / #261: the roadmap still described console viewport colour as unsolved in five places.

Status table — Phase 1 no longer defers console runtime colour; Phase 7 no longer lists console colour among the deferred items, and credits the validated model (#214) and its wiring (#260).
Per-port matrix — the "Image fragments — PNG colour" row goes from 🚧 2 PNGs / port (LUT pairing engine-state-driven) to ✅ 4,338 PNGs (Sega CD) and ✅ 4,758 (Amiga EU).
Amiga per-port note — "Image-to-scene pairing is engine-context-driven and the unsolved piece" is no longer true; restated as the resolved conv(bank[map_graphics_style][light stripe]) model reached through LightingContext.
Phase 1 palette bullet — describes the actual resolution chain (find_image_palette_ctx, Archive::lighting_stripe, LightingContext), not only the record decoders.
Phase 5 extract bullet — documents --style / --light and that PC-family archives resolve through the master and ignore them.

The Phase 7 colour item is rewritten rather than patched. It had accreted into a single ~70-line chronological narrative of every investigation session, and still asserted the jump-table claim the live capture session disproved ("Sega CD: runtime slot == SKUL-local index"), self-correcting forty lines later. It is now three labelled paragraphs — validated model, landed wiring, remaining residue — and is ticked.

Per-object cls3=0x0E LUT application stays deferred, but is folded into the residue paragraph rather than kept as its own open item: the selector is pinned (entry 492, table[(byte4 >> 4) - 1] behind the entry-496 gate) and nothing in the fixtures needs it to render correctly.

No code changes.

Docs-only follow-up to #259 / #261: the roadmap still described console viewport colour as unsolved in five places. - **Status table** — Phase 1 no longer defers console runtime colour; Phase 7 no longer lists console colour among the deferred items, and credits the validated model (#214) and its wiring (#260). - **Per-port matrix** — the "Image fragments — PNG colour" row goes from `🚧 2 PNGs / port (LUT pairing engine-state-driven)` to `✅ 4,338 PNGs` (Sega CD) and `✅ 4,758` (Amiga EU). - **Amiga per-port note** — "Image-to-scene pairing is engine-context-driven and the unsolved piece" is no longer true; restated as the resolved `conv(bank[map_graphics_style][light stripe])` model reached through `LightingContext`. - **Phase 1 palette bullet** — describes the actual resolution chain (`find_image_palette_ctx`, `Archive::lighting_stripe`, `LightingContext`), not only the record decoders. - **Phase 5 extract bullet** — documents `--style` / `--light` and that PC-family archives resolve through the master and ignore them. The Phase 7 colour item is rewritten rather than patched. It had accreted into a single ~70-line chronological narrative of every investigation session, and still asserted the jump-table claim the live capture session disproved ("Sega CD: runtime slot == SKUL-local index"), self-correcting forty lines later. It is now three labelled paragraphs — validated model, landed wiring, remaining residue — and is ticked. Per-object `cls3=0x0E` LUT application stays deferred, but is folded into the residue paragraph rather than kept as its own open item: the selector is pinned (entry 492, `table[(byte4 >> 4) - 1]` behind the entry-496 gate) and nothing in the fixtures needs it to render correctly. No code changes.

No reviewers

No labels

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

jqueuniet/dm2-tools!262

No description provided.

Rows
Columns