docs(phase-34): task 2 — E1 done: scratch mIu4d at 100%, preset request decompme/decomp.me#2106, manual search closes ledger row 14; preset creation pending

This commit is contained in:
Drew T
2026-09-08 12:07:15 -06:00
parent c9a3d7b891
commit 9e8a055e6c
5 changed files with 46 additions and 5 deletions
+1
View File
@@ -0,0 +1 @@
{"created_at":"2026-09-08T18:05:08Z","html_url":"https://github.com/decompme/decomp.me/issues/2106","labels":[],"number":2106,"state":"open","title":"[PRESET] Create Compiler Preset - Brave Fencer Musashi (SLUS-00726)","user":"Druthulu"}
@@ -0,0 +1,17 @@
---
name: Create or update a compiler preset
about: Request a new compiler preset or a change to an existing one
title: "[PRESET] Create/Update Compiler Preset XXXXXXX"
labels: ''
assignees: ''
---
I would like to create/update a compiler preset on decomp.me.
- **Preset Name:** e.g. `Superfrog 64`
- **Preset ID (when updating):** e.g. `13` (see [Compiler Presets API](https://decomp.me/api/preset))
- **Platform:** e.g. `n64`
- **Compiler ID:** e.g. `ido5.3` (see [Compilers API](https://decomp.me/api/compiler))
- **Compiler Flags:** e.g. `-G0 -O1`
- **Libraries (if applicable):** e.g. `directx5.0`
+5
View File
@@ -0,0 +1,5 @@
scratch page -> 403
scratch api -> 403
bug_report.md
create-or-update-a-compiler-preset.md
feature_request.md
+10 -3
View File
@@ -529,6 +529,11 @@ Modern cpp preprocesses → **vintage cc1** compiles to asm → **maspsx** emula
- **Do NOT use the SOTN preset** (`Castlevania: Symphony of the Night` / `gcc 2.6.3-psx` / `psyq_263_221`) — wrong era, guaranteed near-miss diffs.
- decomp.me's API is Cloudflare-challenged (403 to scripts) — scratch searches/uploads needing the API must be done manually in a browser
(the one-time manual BFM search is ledger row 14, closed by the E1 browser session).
- **Outcome (P34 task 2, 2026-09-08 — Drew's browser session):** the proving scratch <https://decomp.me/scratch/mIu4d> (`func_80018F20`,
target = the `--gas` paste, **100% / score 0** on the first compile after the flags were entered); the preset request
**<https://github.com/decompme/decomp.me/issues/2106>** ("[PRESET] Create Compiler Preset - Brave Fencer Musashi (SLUS-00726)", opened
2026-09-08 18:05Z, state open); the manual search (row 14) found no other BFM scratch. **Pending on decomp.me's maintainers:** the
preset's creation — record its id / URL here when it appears (`https://decomp.me/api/preset`).
### §6.6 Matching a function (INCLUDE_ASM → C; the NON_MATCHING guard) — As-built Phase 6
@@ -1003,7 +1008,7 @@ A Track-1 match proves the dump is the canonical redump dump, which transitively
| 11 | Canonical git remote URL (off-box push/pull backup) | **CLOSED (P33 C9, 2026-09-07):** `https://github.com/Druthulu/BFM-decomp.git`, public from C10; the pre-rewrite history in the private archive `Druthulu/BFM-decomp-archive` (§4.3) |
| 12 | Per-libnum stamp detail (raw-track scan reported 16 hits vs 12 genuine in extracted EXE — extracted-EXE scan is ground truth, see §5.1) | **RESOLVED 2026-06-13** — DetectPsyQ at headless import recorded `PsyQ Version = 4.0.0` (§2.5 step 3) |
| 13 | Overlay load addresses (resident 0x800CDF58 / location 0x80128508, EXE ptr table ~0x62620) | **JP-only — re-derive for US** (owned by docs/memory-map.md) |
| 14 | Greenfield claim: decomp.me scratch search is script-blocked (Cloudflare) | **Open until the E1 browser session (Drew, after the flip; `docs/decompme-preset.md` §5 step 4):** the one-time manual search for BFM scratches — record the result here. The preset itself is already PROVEN locally (2026-09-07, `tools/decompme_replica.sh` PASS through decomp.me's 0.13 + `86ccd7d8` toolchain). |
| 14 | Greenfield claim: decomp.me scratch search is script-blocked (Cloudflare) | **CLOSED 2026-09-08 (P34 task 2, Drew's E1 browser session):** manual search on decomp.me for `Musashi` → no results; `Brave Fencer` → no results; `SLUS-00726` → no results; `SLUS_007` → many scratches, all other titles sharing the SLUS_007xx prefix under other compilers — none is this game. The only BFM scratch is our own probe, <https://decomp.me/scratch/mIu4d> (100% / score 0). The greenfield claim's last check is done. The preset was proven locally first (2026-09-07, `tools/decompme_replica.sh` PASS through decomp.me's 0.13 + `86ccd7d8` toolchain; 2026-09-08 step D: the paste itself round-trips). |
### `tools/gap_triage.py` — harvest pre-filter (added P31 S55)
@@ -1200,8 +1205,10 @@ fills fast). Nothing is leaking — but the host does not get the memory back on
an empty-body TU → the CONTROL fails first (rc 3, R56). `--upstream` 2026-09-07: `pins UNCHANGED`. Instrument lesson (R40):
the first comparison read the padded `.text` and called the identical object DIFFERS.
- `.gitignore`: `.run/P33/e1/` allowlisted for the verdict log only (the regenerated target `.s` is game disassembly — H1).
- Pending (Drew, after the flip — `docs/decompme-preset.md` §5): create the scratch, see 100%, file the preset-request issue
(text in §5), run the manual search, record the issue/preset URL + search result here and in ledger row 14.
- **Done 2026-09-08 (P34 task 2, Drew):** scratch <https://decomp.me/scratch/mIu4d> at 100% / score 0; the preset-request issue
<https://github.com/decompme/decomp.me/issues/2106>; the manual search recorded in ledger row 14 (closed). The first attempt failed on
the target's form — the §5 doc and the tools were fixed and proven before the retry (`verbatim_target_s.py --gas`, replica step D).
**Still pending:** the maintainers creating the preset (its id goes into §6.5 when it exists).
### P33 G1 (S88, 2026-09-07) — `docs/gen3-handoff.md`
- The census was DERIVED from the tree at writing (R33/R41; the commands are in the doc): 143 raw address casts, 61,898
+13 -2
View File
@@ -42,7 +42,7 @@ bundle and dictionary are deleted at C11) and a public fresh-clone re-proof (B3'
- [x] **0a** Open the phase: this file, R84–R95 into DIGEST §3 (+ one §0 line), the `.run/P34/` evidence allowlist, TaskCreate ×10, `doc_links --strict` 0/0, commit — xHigh (Claude)
- [x] **0** The probe once more: `tools/public_rewrite/probe_github.sh | tee .run/P34/probe/pre_flip.log` → PASS, control OK, no WARNING — Low (Claude)
- [x] **1** C10 — the flip (Drew, Max) + post-flip checks, metadata, ruleset, Actions green on the public repo (Claude, Low)
- [ ] **2** E1 — decomp.me: scratch → 100% → preset-request issue → manual search (Drew, xHigh); recorded in SETUP §6.5 + ledger row 14 (Claude)
- [x] **2** E1 — decomp.me: scratch → 100% → preset-request issue → manual search (Drew, xHigh); recorded in SETUP §6.5 + ledger row 14 (Claude)
- [ ] **3** E2 — the Archipelago issue (Drew, xHigh); URL recorded; §5 reply handling pending unless a reply arrives (Claude)
- [ ] **4** D3 — decomp.dev registration + frogress slug/key/push (Drew, xHigh); verified by curl (Claude) → **P6 rules check after this task**
- [ ] **5** F3 — the wiki: dry run (Claude), first page + `wiki_sync.sh --push` (Drew), pages verified — Low
@@ -121,7 +121,18 @@ output) · `.run/P34/c11/` (the prune listing with sizes, `git worktree list`, t
- The bundle's `1_target_asm.s` is now the gas form (42 lines); 2–4 unchanged in content; `5_issue_body.md` untouched (Drew's text).
- Drew retries the scratch with the new file → 100% / score 0 expected → the issue → the search → part 2 records the URLs.
## 🛑 SESSION CHECKPOINT — PHASE 34 OPEN (S93, 2026-09-08): tasks 0a, 0, 1 ✓; task 2 part 1 ✓ (the paste form proven); NEXT = Drew's decomp.me retry, then task 2 part 2 (record the URLs)
### 2026-09-08 — Task 2 (part 2) — E1: the browser session done, recorded (xHigh; S93)
- **Drew (browser):** the retry with the `--gas` paste compiled to **100% / score 0** — scratch **<https://decomp.me/scratch/mIu4d>**
(the page answers 403 to scripts, the Cloudflare block SETUP §6.5 records; Drew's report is the observation). The preset-request issue
**<https://github.com/decompme/decomp.me/issues/2106>** — `gh api` (`.run/P34/outward/e1_issue_2106.txt`): title "[PRESET] Create
Compiler Preset - Brave Fencer Musashi (SLUS-00726)", opened 2026-09-08T18:05:08Z by Druthulu, state open, no labels; decomp.me's
template `create-or-update-a-compiler-preset.md` fetched and compared — our body carries every field it asks for
(`e1_preset_issue_template.txt`). The manual search: `Musashi` / `Brave Fencer` / `SLUS-00726` → nothing; `SLUS_007` → many scratches
of other SLUS_007xx titles under other compilers, none this game → **ledger row 14 CLOSED** (the greenfield claim's last check).
- Recorded in `docs/SETUP.md`: row 14 closed; §6.5 outcome bullet; the P33 E1 bullet → done. **Pending by name (P9):** the preset's
creation by decomp.me's maintainers (its id → §6.5 when it appears).
## 🛑 SESSION CHECKPOINT — PHASE 34 OPEN (S93, 2026-09-08): tasks 0a, 0, 1, 2 ✓; NEXT = task 3 (E2 the Archipelago issue — DREW), then 4 (D3 boards — DREW) → P6 rules check
### 0. How to use this block
A fresh session reads CLAUDE.md's load order, replays THIS block verbatim, and resumes at the first unchecked task above. Everything