# Timeline of the optimization effort

Condensed from `.pi/journal.md` (the authoritative running log — read it for detail).
Commit hashes are on `main`.

## Background (pre-effort)

- 2026-09: pack curation era — FTB Chunks → Open Parties & Claims + Xaero's swap,
  FTB Quests/Teams removed, map migration tooling (`analysis/FTB-XAERO-MAP-MIGRATION.md`).
- 2026-10-06: 60-mod expansion to 209/212 mods + full static audit of all jars
  (`analysis/audit/`), then the **Paxi `onion-fixes` datapack** (`fc649b4`) generated by
  `analysis/audit/gen_fixes.py` to resolve cross-mod recipe collisions. Rules: edit the
  generator, not the JSON.

## Client-performance initiative

| Date | Phase | What happened |
| --- | --- | --- |
| 2026-10-14 | **Phase 1 — static baseline** | `analysis/client-perf/` created: jarscan (951 per-frame mixin classes; Handcrafted+Chipped+Rechiseled = 58% of block atlas), 9 research briefs, **LEVERS.md** (ranked levers), observed-default configs. Nothing changed in the pack. |
| 2026-10-14 | **Phase 1b — delivery + bench design** | `proposed-configs/` generator (`build.py`; tiers T-default/deck/mid/high + 32 bench variants C0…spr-strict), **BENCH-MATRIX.md** (run sheet), bench datapack `bench/onion-bench/`, PLAYER-TUNING.md. Verified packwiz `preserve` semantics and Default Options 21.1.8 behavior. Puzzles Lib 21.1.64 patch prepared. |
| 2026-10-14 | **Owner decisions applied, shipped** | Owner settled the pending list (see `05-decisions-and-rules.md`). Commits: `a01bc3b` (Default Options + 8 provisional mid-tier files + Sodium Extra, Reese's Sodium Options, BadOptimizations, Particle Core + Puzzles Lib 64 → **217 mods**), `d80f85b` (spawn tuning: alexsmobs/fowlplay/critters partial files), `5919955` (archive cleanup), docs commit (client-perf, audit, phase0, journal). |
| 2026-10-14/15 | **Phase 2 begins — boot + harness** | Machine moved to Manjaro (client-capable). Client boot check **essentially PASS** (main menu clean, 29.3 s, atlas 4096×4096×2, Default Options seeded). The one real blocker — **Critters & Companions camelCase spawn keys** NPE — fixed by owner as `8e8483d`. A 20:34 FML crash report is that same CAC bug (not WATUT). Harness built (`harness/run_scene.py`, `rcon.py`), user-local MangoHud 0.8.4, bench world generated via temp server `~/onion-bench/server-bench` (S2 Create floor + S3 mob pen placed and verified; **S4 airships not built**). One end-to-end harness run (C1/S1) — **vsync-locked, non-decision-grade**; `vblank_mode=0` added to the wrapper afterwards, never re-verified. Dedicated server booted clean from the repo at 21:01 (idle, RCON-keepalive). Committed as `e5e369a`. |
| 2026-10-14 late | **Session interrupted** | After `e5e369a` (23:16): server kept running in tmux `onion-mc`, mods re-saved two config files (benign), RCON poller pinging every 30 s until ~23:58. **No matrix block was ever run.** State captured into `analysis/state/` (this directory). |

## HEAD ancestry (optimization-relevant commits)

```
e5e369a chore(bench): add client bench harness and Phase 2 handoff   ← HEAD, pushed
8e8483d fix(config): correct Critters and Companions spawn keys      ← owner, unblocks boot
611a1a5 docs: client-perf analysis, bench harness and work journal
d80f85b feat(config): moderate spawn tuning (Alex's Mobs, Fowl Play, Critters)
5919955 chore(analysis): archive finished reviews
a01bc3b feat(perf): Default Options + 4 perf mods + Puzzles Lib 64   ← 217 mods
aa1d273 ... (Phase 1 baseline was taken here)
fc649b4 feat(pack): Paxi onion-fixes datapack (recipe collisions)
```

## What has NOT happened yet

- Matrix blocks **B0–B14: zero runs**.
- Any config value shipped in `a01bc3b` is still **unmeasured** (provisional mid-tier).
- WATUT world-load crash unresolved.
- S4 bench scene unbuilt; S5 tour only partially exercised.
- Nothing has been reported back to the owner from Phase 2 data (B7 shader data, B12b
  trial verdicts are owed).