Skip to content

Fix systematically misaligned enchant/gem/consumable ID databases - #11

Merged
alexandermayes merged 1 commit into
mainfrom
growth/fix-game-data-ids
Sep 4, 2026
Merged

Fix systematically misaligned enchant/gem/consumable ID databases#11
alexandermayes merged 1 commit into
mainfrom
growth/fix-game-data-ids

Conversation

@alexandermayes

@alexandermayes alexandermayes commented Sep 4, 2026

Copy link
Copy Markdown
Owner

Context

A TBC Fresh user reported that enchant detection "seems to be wrong" and some consumables are "wrongly identified... maybe the ids are not of the right version of the game." They were right — and it was worse than it looked.

Root cause

The hand-maintained ID→name databases in lib/cla-constants.ts were systematically misaligned — whole sections shifted/rotated by one row, in every era section (Classic through MoP). Verified against actual game client data (wago.tools dumps of SpellItemEnchantment / SpellEffect / SpellName / ItemSparse / GemProperties), era-matched to the clients WCL logs come from:

Era sections Client build used
Classic + TBC 2.5.6.69546 (TBC Anniversary — the exact TBC Fresh client)
WotLK 3.4.5.63697
Cata 4.4.2.60895
MoP 5.5.4.69585

Findings

  • 88 of 106 verifiable enchant IDs mapped to the wrong name. e.g. glove enchant 684 displayed as "Major Agility (+15 Agi)" but is +15 Strength; every TBC head glyph, shoulder inscription, leg armor and spellthread was rotated (Glyph of Ferocity is 3003 not 2999, Nethercobra Leg Armor is 3012 not 2747, …). Same pattern in WotLK/Cata/MoP sections.
  • Gem databases equally broken. "Bold Living Ruby" keyed to Bold Crimson Spinel's item ID; several "gem" IDs were actually cloaks, profession formulas, or Monster - Dagger placeholder items. GEM_STAT_DB (drives "bad gem for your role" flags) had the same shifted IDs → false flags on correct gems.
  • Consumables:
    • 28093 ("Lightning Speed") / 28095 are Mongoose/enchant proc buffs, listed as weapon oils — on TBC Fresh anyone with the Mongoose enchant was credited a phantom "weapon enhancement".
    • Scroll buff IDs shifted (33077 is Agility, not Intellect); rank-V scrolls were flagged "suboptimal — use rank VIII" though VIII doesn't exist in TBC.
    • 57362/58067 counted as food buffs but are mage Refreshment.
    • Cata elixirs rotated (79474 is Naga, 79477 is Cobra, 79480 is Deep Earth — a guardian elixir, not battle).
    • 28519 is Flask of Mighty Restoration ("Versatility" doesn't exist); 53752 is Lesser Flask of Toughness, not Elixir of Mighty Defense; 54452 is a dead spell ID (removed).

Fix

  • Regenerated ENCHANT_NAME_DB, GEM_NAME_DB, GEM_STAT_DB mechanically from the client data above. Enchant names come from the applying spell (e.g. Enchant Gloves - Major Strength) plus the enchantment's actual stat text; gem stat types are derived from each gem's real enchantment text (kept as a trailing comment per entry for reviewability).
  • Surgical consumable fixes in CONSUMABLE_DB + the ID sets in lib/constants.ts (proc buffs removed from WEAPON_ENHANCEMENT_IDS, Refreshment removed from FOOD_BUFF_IDS, 53752 moved to FLASK_BUFF_IDS).
  • New regression suite lib/cla-constants.test.ts (10 tests) locking in known-good pairings.

Verification

  • npx tsc --noEmit ✓
  • npm run lint ✓ (no new findings)
  • npm test — 34/34 ✓ (10 new)

Notes / residual limitations

  • WotLK/Cata per-food friendly names (all generically named "Well Fed" in game data) can't be cross-checked by name; IDs left as-is. They only affect display labels, not detection.
  • A few engineering tinker enchant IDs (e.g. Nitro Boosts 3606) aren't representable in the dumped tables and kept their original names.
  • Deploy is manual (vercel deploy --prod) after merge — this PR does not self-deploy.

🤖 Generated with Claude Code

@alexandermayes @claude
Fix systematically misaligned enchant/gem/consumable ID databases
839c97c 
A TBC Fresh user reported wrong enchant and consumable identification.
Verification against actual game client data (wago.tools dumps of
SpellItemEnchantment/SpellEffect/SpellName/ItemSparse/GemProperties)
showed the hand-maintained ID->name maps in cla-constants.ts were
systematically shifted in every era section:

- 88 of 106 verifiable enchant IDs had the wrong name (e.g. glove
  enchant 684 shown as "Major Agility" but is +15 Strength; all TBC
  head glyphs / shoulder inscriptions / leg armors rotated by one)
- Gem maps pointed at wrong items (some "gems" were cloaks, formulas,
  or Monster placeholder items); GEM_STAT_DB's shifted IDs produced
  false "bad gem for your role" flags on correct gems
- 28093/28095 (Mongoose/enchant proc buffs) were listed as weapon
  oils, crediting enchant users with a phantom weapon enhancement
- Scroll buff IDs shifted; rank-V scrolls flagged "use VIII instead"
  though VIII doesn't exist in TBC; two "food" IDs were mage
  Refreshment; Cata elixirs rotated; a few flask names wrong

Fix: regenerated ENCHANT_NAME_DB, GEM_NAME_DB and GEM_STAT_DB from
era-matched client data (Classic/TBC = 2.5.6.69546 TBC Anniversary,
WotLK = 3.4.5.63697, Cata = 4.4.2.60895, MoP = 5.5.4.69585), applied
surgical consumable fixes, and added a regression test suite locking
in known-good pairings.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@vercel

vercel Bot commented Sep 4, 2026 •
edited
Loading

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
parseforge Ready Ready Preview Sep 4, 2026 5:44pm UTC

@alexandermayes alexandermayes mentioned this pull request Sep 4, 2026
Merged
@alexandermayes
alexandermayes merged commit 2888ca7 into main Sep 4, 2026
5 checks passed
@alexandermayes
alexandermayes deleted the growth/fix-game-data-ids branch September 4, 2026 18:20
alexandermayes added a commit that referenced this pull request Sep 22, 2026
@alexandermayes @claude
docs(03-07): complete production-deploy-and-honest-close-out plan
f66faa3 
- Add 03-07-SUMMARY.md: Phase 3 live in production (dpl_6Pj5Lz5Q1tSJYSCtUu3YTvtRx3mx),
  item-7 and share-rate figures counted, Part 5 closed LEFT OPEN — not signed
- REQUIREMENTS.md: OPS-01 Addendum 3 (Phase 3 gate re-run not met) and a new
  SHARE-01/02/03 addendum (capability proven live; live-traffic/human-review
  backstops still open) — no checkbox reverted, ledger made honest via addendum
- WINDOWS.md: two new open follow-ups (#10 item-7 re-measure, #11 share-rate re-run)

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
alexandermayes added a commit that referenced this pull request Sep 22, 2026
@alexandermayes @claude
docs(03): plan OPS-01 gap closure — 03-08 through 03-10
3df9f26 
Closes the one FAILED truth in 03-VERIFICATION.md (ROADMAP SC4 / OPS-01): all nine
docs/OPS-01-SHIP-GATE.md Part 5 rows are driven to a dated outcome.

- 03-08 (wave 6): the Search Console pass that was never run (rows 4-6) plus the three
  developer-only backstops — D-04 tone, real Discord unfurl, D-13 phone width (rows 7-9).
  Leads with a tracer that takes the homepage row end-to-end.
- 03-09 (wave 7): item-7 re-measure on a fully elapsed busier UTC window, the Vercel Web
  Analytics read for threshold 3, and the RESEARCH A3 consent-path disposition (rows 1, 3).
- 03-10 (wave 8): the date-gated share-rate row (no query before 2026-09-23T09:15Z), the
  Part 5 close-out and exactly one sign-off, REQUIREMENTS Addendum 4 plus a superseding
  SHARE addendum, and the WINDOWS #9/#10/#11 dispositions (row 2).

Every tool-dependent step names who holds the tool: the executor has no PostHog, Vercel,
Search Console or browser tool, so that evidence arrives as a blocking checkpoint and is
transcribed verbatim (the 03-07 handoff pattern). 03-REVIEW.md WR-01/WR-02/WR-03/IN-01 are
explicitly deferred to /gsd-code-review --fix.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
alexandermayes added a commit that referenced this pull request Sep 22, 2026
@alexandermayes @claude
docs(03-10): append Part 5 close-out and Phase 3 SIGNED sign-off
f737a64 
- Restates all nine closing-table rows with counted outcomes and
  evidence pointers (rows 1, 3-7, 9 counted PASS or delegated PASS;
  row 2 standing first reading; row 8 NOT OBTAINED)
- Records Task 1's decision verbatim: developer replied "sign-now"
  then "you call"; orchestrator resolved to sign-now-conditional
- Writes exactly one closing heading: Phase 3 Sign-off — SIGNED
  (2026-09-19), with the sign-now-conditional obligation that the
  D-14 re-read (on/after 2026-09-23T09:15Z) must append a dated
  confirmation or correction, and WINDOWS.md #11 stays open until then
- Signature explicitly excludes 03-REVIEW.md WR-01/02/03/IN-01,
  deferred to /gsd-code-review --fix

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
alexandermayes added a commit that referenced this pull request Sep 22, 2026
@alexandermayes @claude
docs(03-10): carry Phase 3 sign-off into REQUIREMENTS and WINDOWS led…
1d43ef7 
…gers

- Append OPS-01 Addendum 4 (supersedes Addendum 3): Phase 3 re-run met
  2026-09-19 against dpl_6Pj5Lz5Q1tSJYSCtUu3YTvtRx3mx, Part 5 SIGNED
  (conditionally on the D-14 re-read appending a dated confirmation
  or correction on/after 2026-09-23T09:15Z)
- Append SHARE-01/02/03 Addendum 2 (supersedes 2026-09-16 addendum):
  live-PostHog and D-13 backstops now closed; real Discord unfurl and
  first-hand D-04 verdict remain open
- Both additions are pure appends; every prior addendum, the
  Phase 1 gate closure text, and all checkbox/traceability lines
  are untouched
- Mark .planning/WINDOWS.md #10 fixed (item-7 re-measure closed);
  #9 and #11 stay open per the close-out's dated dispositions

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

This branch was successfully deployed

1 active deployment
Preview — 839c97c4 Deployed Sep 4, 2026 by vercel[bot]
View deployment
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@alexandermayes