ItemDB-v1.2.0
What's new
Changelog
[v1.2.0] - 2026-10-04
New -- data for Wrath, Cataclysm and Retail
Requested by SmexyMats (inbox 926a233876bf), which generates its tooltip tables from ours and fell
back to a hand-made list on these three flavours. The operator chose "Ask ItemDB to build it".
- Cores. No Wrath or Cata client is installed, so neither has a walk; Retail has one since
2026-10-04 (below). Wrath 45,155 items, Cata 65,797, Retail 176,130.
- Wrath and Cata carry real stats from the new
tools/build-core-wago.py, which reads wagoItemSparse(StatModifier_bonusStat/bonusAmount,Resistances_0) and REFUSES to write unless it first reproduces the shipped walk-built stats of TBC (classic profile, applied to Wrath: 39,426 exact key matches, 0 unexplained) and Mists (modern profile, applied to Cata: 173,121, 0 unexplained). Every remaining mismatch is a listed reason (on-equip auras, whichbuild-equip-stats.pycovers; 11 named "Communal" items). Wrath 23,944 items with stats, Cata 30,403. Left out because no walk can confirm them: weapon DPS on both (ItemSparse damage reads about half the Mists walk's), all armor on Cata (the Mists walk disagrees with its own build'sResistances_0on 18,598 items), shield armor and school resistances on Wrath (the TBC walk never reports either).Tests/core_wago_check.py(45 checks, red-checked) pins the gate and the shipped blobs;build-core-wowsims.pynow refuses to overwrite a core this builder wrote. WoWSims wotlk/cata were the alternative and were not used: theirdb.jsonis a filtered subset and their stat enums would need a new key map and rating-divisor rework. - Retail is walk-built, with stats (
build-core.py Retail, from the operator's TOG Tools walk of the Retail client, 2026-10-04): 176,130 items -- 901 more thanItemSparse's 175,229, the items the client carries outside that table -- across 19 class files, names from the walk. Stats areGetItemStatson the linkGetItemInfo(id)returns for the bare id, so no bonus or upgrade ids: the base item (TOG ToolsModules/ItemDB/ItemDB.lua:266-275; Spirit Tender's Cuffs 175675: Intellect 4, Mastery 4, Stamina 6, Versatility 3, armor 3). It replaced the structure-only corebuild-core-wowsims.pywrote (ItemSparseholds onlyStatPercentEditorallocations and scaling curves, no flat amounts); that builder now refuses to overwrite it, by the existing walk-built header check. Reagents, Expansion, ItemLocations (121,347 items) and ItemPlaces were rebuilt on the new item set. No Retail scoring model yet. The walk also settles the "~38,000 missing items" question: TOG Tools walks to id 320,000 and keeps only ids the client has data for (GetItemInfoInstant), so 176,130 is every item the Retail client knows; the rest of the client's 213,415Itemrows have no data behind them.
- Wrath and Cata carry real stats from the new
- Pins.
FROZEN_BUILDgains Wrath3.4.3.58936(the last GLOBAL Wrath build; wago's 3.4.5 is the China Titan-Reforged line), Cata4.4.2.60895and Retail12.1.0.69933.PRODUCT_LABELmaps the productwowtoRetail, so the folder does not move when 12.x becomes 13.x. KNOWN COST: when the Anniversary client reaches 3.x the Wrath pin wins over its live build; move it then. - Retail slot tokens.
INVTYPEgains 27 and 29..34 (quiver, profession tool/gear, the four equipable-spell slots), read from the live client'sEnum.InventoryTypeand GlobalStrings. - Reagents for all three: Wrath 766 items, Cata 842, Retail 3,088 across 13 professions.
- Expansion for Wrath, Cata, Mists (new) and Retail. See Changed for the rule.
- ItemLocations (SmexyMats' follow-up, inbox
e9aefe20d418):- Wrath from CMaNGOS wotlk-db (23,256 items; the DB is now cached).
- Cata from SkyFire 5.4.8 (
mistsskyfire.sqlite, the post-Cataclysm world) through the full world-DB path: 31,975 items (gathered 826, drop 9,185, quest 8,476, vendor 9,986, pvp 271, plus crafted and reputation). Vendor stock is 5.4.8's, so a Cataclysm item whose vendor changed in Mists shows the Mists answer. - Retail from TrinityCore's world database, TDB 1210.26091 (client 12.1.0; GPL-2.0), fetched and
converted by the new
tools/fetch-trinitycore.py(desk toolitemdb-fetch), through the full world-DB path, then the AllTheThings passes below: 87,178 items in all. SkyFire stays a gathering-only supplement (GATHER_DB) for 170 old-world items TrinityCore links to no loot. TrinityCore's schema is detected by column throughout: lootItemType, per-difficultyLootID/SkinLootID/TypeFlagsincreature_template_difficulty,QuestSortID, gathering-node gameobject type 50, and per-expansion gathering LockTypes read by name ("Khaz Algar Mining 25"). The client-only mode this replaced was removed the same day. - Modern gathering from AllTheThings (MIT, pinned commit
c84276b52158, cached by the newtools/fetch-att.pythroughitemdb-fetch): TrinityCore has spawns and loot for few Dragonflight-onward nodes, so a further pass reads what ATT records under each gathering spell (Mining 2575, Herb Gathering 2366, the skinning spells, Fishing) and under prospecting and milling (salvagerecipeas well asheader), for items no world DB gave a gathered row. 642 items; per profession, Dragonflight 119, The War Within 84, Midnight 117, older 324 (a few items count under two). Never removes a row; prints any spell header under a gathering profession it does not read.Tests/att_gather_check.py(23 checks, red-checked). A rule from item category alone was measured first and rejected: only herbs matched the real source 98% of the time (metal and stone 76%, leather 72%, gems 78%). STILL UNCOVERED, Dragonflight / The War Within / Midnight reagents with no gathered or crafted row: gems 12 / 6 / 8, pigments 4 / 2 / 3, cooking meats 14 / 12 / 11 (mob drops TrinityCore lacks), cloth 6 / 15 / 8, metal and stone 4 / 2 / 2, plus 184 others. - Drops from AllTheThings (Retail): an item with no drop, gathered, vendor or quest row gets a
drop row from ATT's own records -- world drops (
03 - World Drops), zone drops inside a map header (02 - Outdoor Zones, named from the client's UiMap) and instance loot (01 - Dungeons Raids, through the same instance-name aliases). 10,443 items (Dragonflight 794, The War Within 328, Midnight 537, older 8,784); ATT names the creature for 2,746, the rest are ATT world or zone drops with no creature named. Vendor stock and an item's own contents are never read as drops, and an item the Encounter Journal already places by a boss is skipped (journal_boss_items, which ignores Cata's items-nilmerge call whose first table is the instance map).fetch-att.pynow stays on the recorded commit unless--update, and caches ATT'sMaps.luafor the map constants.Tests/att_drops_check.py(14 checks, red-checked). This gave 24 of the 37 uncovered Dragonflight-onward cooking meats and 15 of the 29 cloths a source. The later passes in the next bullet took the rest down to 12. - Every other origin AllTheThings and the client record (Retail; for SmexyMats' request
d95142e3, all additive, each record shape pinned inTests/att_drops_check.pyoratt_gather_check.pyand red-checked):- salvage as a gathered row: disenchanting -> Enchanting, crushing -> Jewelcrafting, scrapping -> Engineering;
- crafted: SpellEffect 294 ("craft loot") and 296 ("craft salvage item", names from TrinityCore's SpellEffects.cpp) besides 24 and 288 -- Chopped Mycobloom, Fresh Fillet and the Transmutagens are these -- plus ATT's crafted-items lists (2,893 items; the Wilderthread / Duskthread / Dawnthread / Weaverthread spools have no client recipe row at all, only ATT's);
- vendor stock, including profession supply vendors filed under a profession header in a zone (rule checked against one real file);
- quest REWARDS and reward headers -> the zone, else "Quest" ("Crafting Orders" for patron rewards). Quest-objective items, which ATT tags "(QI!)" in three shapes, never get a quest row: that removed 4,632 items' rows;
- treasures, objects and container contents (Airy Soul -> "Docile Airy Soul"), and instance loot ATT keeps outside the Encounter Journal;
- Midnight housing lumber as object drops per map (ATT names no profession for it).
- item constants: ATT writes some items by name,
i(ARTISANS_ACUITY);fetch-att.pynow also caches ATT'sConstants/Items.luaandatt_source_rows(..., items=)resolves them through it. The pass also reads21 - Holidays, and a holiday file's row with no map names the holiday ("Darkmoon Faire" for Artisan's Acuity, a Darkmoon profession-quest reward). - Fixed: a Companion Pet, Mount or Mount Equipment item (Miscellaneous subclass 2 / 5 / 6) is
no longer read as a container, so items nested under one in ATT no longer "drop from
Reins of the Quantum Courser" (
ATT_NOT_CONTAINER; 3 items' rows). - Blizzard's own collectable sources (
client_source_pass, before ATT's drop pass, so an item they place keeps ATT out): the client tablesCollectableSourceInfo+CollectableSourceVendorSparse/QuestSparse/EncounterSparserecord a vendor, quest or Encounter Journal boss for every collectable appearance (96,900 / 15,172 / 47,320 rows at 12.1.0.69933). Vendor rows for 12,629 items, quest rows for 7,189, boss-drop rows for 2,121; Retail locates 96,447 items, up from 88,842. A vendor or quest in an instance's map reads that instance's name -- which for an open-world map is its Journal world-boss tier, so a Valdrakken vendor reads "Dragon Isles" -- else "Vendor" / "Quest". Each item comes from its APPEARANCE (ItemModifiedAppearance.ItemID), never from the vendor table'sVendorItemID: on the Mining Pick's trade-supplier rows that column names Lasherweave Raiment, which Wowhead lists as not available to players.Tests/att_drops_check.pypins the join. - The same vendor rows feed Retail's "Where to get it" (
build-item-places.pyclient_vendor_spawns/add_client_vendors, Retail only): 15,960 items gained a vendor, and 1,281 vendors TrinityCore never spawns are placed at the client's own position (Provisioner Thom, Valdrakken's Blacksmithing Supplies, at 2112: 36.6,50.6). A creature TrinityCore spawns keeps its spawns; one with no name there is skipped.Tests/item_places_check.pypins the merge. - Quest givers in "Where to get it" (MINOR 40): a new place token
q<creature>:<questID>fromCollectableSourceQuestSparse;GetItemPlacesreturns it askind = "quest"withquestID(no chance), after objects, at most 4 per item (MAX_QUESTS). 7,453 Retail items gained a quest giver, 583 givers placed at the client's position alone. Checked: quest 58166 "Unknown Assailants" (Tirna Vaal) rewards 175675-175677, which carryq160439:58166, the giver on Ardenweald's map; Wowhead did not name the starter, so the giver is not verified.BuildWhereRowsshows it as a "Quest" row, applies the vendor faction filter to it, and no longer repeats the generic "Quest" location line when a giver row covers it. An older library skips the token. Specs:libitemdb_spec.lua,where_spec.lua,item_places_check.py.lib.PLACE_KIND(the token letters) is now public, read-only, so a consumer decoding tokens itself need not keep a copy that misses the next letter -- SmexyMats kept one (Peer Review, thread 8acd9f23). - The compile spec is one file per TOC (
Tests/data_compile_<flavour>_spec.lua, body inTests/data_compile.lua). Measured with the newspec-timessuite (run.lua --times, NOT A GATE): the singledata_compile_spec.luatook 69 s CPU and hit the harness's 60 s per-file limit part-way through Mists, so the TOCs after it never compiled -- a pass that depended on the machine's speed. Now the slowest, Mists (12 large Names files), takes 18-26 s.provenance_spec.luaread its-- sourceand-- client buildstamps by walking every line of every Data file; it now reads only the comment header they live in (22 s -> 11 s). - Every Retail item, measured (
report_unplaced, printed by every Retail build): of 175,229 shipped items, 135,314 (77%) now have a source -- ItemLocations or an Encounter Journal boss drop (19,049,Sources.lua, which an earlier count wrongly left out). Of the 39,915 without one, AllTheThings files 10,971 as never implemented and does not list 18,854 at all; 10,090 it lists and no pass reads yet (class-trial / boost / starting gear 1,844 is left out on purpose; the rest is listed by folder in the build output). - Folder sources (
att_folder_pass,ATT_FOLDER_SOURCES): an item no record-shape pass placed takes its source from the ATT folder it is filed in.08 - PvPwas never read at all -- 12,396 items, whole season files (Legion Season 3 Cruel.lua:1121, Cruel Gladiator's Ring -> "PvP") -- plus Trading Post 951, instance files (the file's own name) 928, holidays 784, Delves 470, Group Finder 439, In-Game Shop 258, Black Market 57. A file named for something that is not a place (anAchievements.luainside these folders) is skipped, so achievement rewards do not read "Achievements" as where they drop. Zones, events, expansion features and character files go by their PATH (att_path_row: the folder is the place, the file name below it the kind -- Quests, Vendors, Gear Conversion, Rares, Dungeons Raids; achievements, boosts and files naming no kind are left alone): events 2,335, zones 963, expansion features 548, Heritage 210. Onlyi(<id>)counts, never a{ "i", id, n }cost; quest objectives and never-implemented items are skipped. - Quest objectives (ATT's "(QI!)") with no other source now read their quest's zone
(
add_objective_rows), 4,064 items -- e.g. V-300 Serial Plate -> Tiragarde Sound. Earlier today they were given no row because "rewarded by a quest" is wrong for them; for an item nothing else places, "got during a quest in that zone" is the true answer, and an objective that also drops or is sold keeps only those rows. - Items missing from the database itself: the client's
Itemtable has 213,415 rows andItemSparse(names, details -- what Retail ships) 175,229, so 38,186 client items are not shipped.Tests/item_coverage_report.py(run_check) breaks them down: AllTheThings marks 17,262 never implemented and does not list 19,890; nine Wowhead samples were all 404s or placeholders ("[PH] Plate Chestguard of the Rising Dawn"). 1,034 it lists as real items (quest 377, armor 372, miscellaneous 149) -- not shipped yet, because the client carries no name or details for them. The Dragonflight-onward reagents with no row at all went 152 -> 11, each with a reason: Fang Adornments is never-implemented; two Shadowflame crests were a fragment exchange removed in 10.2.0; three Amirdrassil prototypes are quest objectives; Smudged Lens, Soul Sigil I and II and Spark of Tides are listed by ATT with no source and made by no client spell; Temporal Vestigial has only a description. The client's QuestPackageItem (quest reward packages) names none of the 11, and CurrencyContainer holds currencies, not items.itemdb-schemagained a row filter for that,Table:Column=v1|v2, inside its table field so its command and approval stay unchanged.
- Prospecting and milling are gathered rows now, like disenchanting ("Enchanting"):
prospecting_loot_template-> Jewelcrafting andmilling_loot_template-> Inscription, read wherever the world DB has the table. TBC 32 / 0, Wrath 51 / 15, Cata 58 / 15, Mists 71 / 17, Retail 166 / 57 (with ATT); Vanilla has neither table.fetch-trinitycore.pyconverts both now, and re-converts when its cache lacks a listed table: the skip used to compare only the release tag, so a cache built before a table was added counted as current. - Vanilla ItemLocations was rebuilt for the first time since v1.1.4 and gains that release's fixes, which had only reached TBC and Mists: reagents such as Arcane Powder, Ankh and Symbol of Kings read their PvP Barracks and ordinary vendors instead of "Alterac Valley", and the four battleground healing draughts read Vendor. An orphan row for item id 1 (in no Vanilla item file) dropped out.
- Sources (instance drops):
- Wrath from AtlasLoot (5,522 items).
- Cata: AtlasLoot up to Wrath (5,511 items) plus 20 instances from the client's Encounter
Journal through the new
tools/build-sources-journal.py: 16 new Cataclysm instances, and the 4 instances Cata revamped (Deadmines, Shadowfang Keep, Zul'Gurub, Zul'Aman) folded into the AtlasLoot entry instead of listed twice. ORDER MATTERS:build-sources.py Catarewrites both files with AtlasLoot only, sobuild-sources-journal.py Catamust run after it. - Retail, all from the Journal: 213 instances (142 dungeons, 70 raids), 1,138 encounters, 19,049 items. 14 bosses the Journal lists twice in one instance are merged into one; same-named instances get a suffix (Blackrock Depths, Magisters' Terrace).
- DropRates for Wrath (7,956 items) and Mists (7,561). Mists reads SkyFire, whose loot rules
differ from CMaNGOS: a reference row rolls its own chance and processes only its group, and rows
without lootmode bit 1 are hard-mode only (skipped). The builder picks those rules from the
table's columns, not the version name. Wrath also merges
DifficultyEntry1..3heroic loot, which the oldHeroicEntry-only read missed. - ItemPlaces ("where to get it") for Wrath (17,264 items, CMaNGOS) and Cata (21,078, SkyFire), with 8 locales of place names each, and Retail (23,802 items) from TrinityCore's world database with 8 locales of place names. SkyFire was ruled out for Retail first: its world is pre-Legion and Retail has rebuilt much of it (Undercity, Teldrassil, Silithus).
- RepLoot for Wrath (964).
- TOCs.
ItemDB_Wrath.tocandItemDB_Cata.tocnow load their data (they were deliberately empty placeholders). NewItemDB_Mainline.toc(Interface 120100) loads the whole addon -- Integrations, Tooltip, Price, Where and the bootstrap, with the same dependencies, OptionalDeps and SavedVariables as every other flavour -- and no Scoring file, since Retail has no stats. It shipped data-only at first; the operator asked for the full addon the same day ("we also need to create a retail .toc for you, so you work in retail"). Every module already feature-detects the modern client (data-driven tooltips viaC_TooltipInfo,C_AuctionHouse,C_MerchantFrame;HandleModifiedItemClickexists in the live UI source). Loaded on a Retail client with TOGProfessionMaster, no errors, after the Names.lua fix (operator, 2026-10-04: "togpm is running with no errors, and before it had a bunch"); the tooltip lines, price window and "Where to get it" window have not each been exercised there yet.Tests/toc_spec.luanow treats all seven manifests as full.wow-version-replication.ps1replicates to_retail_too.ItemDB_Mists.tocloads its newExpansion.lua.
New -- scoring and side tables for Mists, Wrath and Cata (parity with TBC)
The operator asked for parity so addons like Dibs work on every flavour ("lets get to parity").
- Scoring. New
Scoring/Mists.lua,Scoring/Wrath.lua,Scoring/Cata.lua, each loaded by its own TOC only (Tests/toc_spec.luanow expects them). Rating conversions at each level cap from WoWSims'base_stats_auto_gen.go(Wrath 80, Cata 85, Mists 90). Spell power is counted once (the data stores "+N spell power" as an equal damage and healing pair). Cata and Mists skip the stats patch 4.0.1 removed (defense, armor penetration, block value, ammo, feral AP); only Wrath registers the ammo scorer. Death Knight (6) and Monk (10) score through their weights. Specs:Tests/{wrath,cata,mists}_scoring_spec.lua(48 tests); 100% coverage on all three. - BiSWeights for Mists (45 specs, 11 classes incl. Monk), Wrath and Cata (41 specs, 10 classes incl. Death Knight), from the Wrath/Cata/Mists branches of HawsJon's Pawn scale. The builder now refuses any Pawn stat that is neither mapped nor explicitly ignored. Not priced: weapon Speed (Marksmanship's top Mists weight), resistances, school spell damage, Hp5, feral AP.
- Side tables. Mists: Factions, Mounts, Sets, ReqLevels, VendorPrices, SellPrices, Classes, EquipStats, UseEffects, BiS, ScoreModel, BiSWeights, DropRates. Wrath and Cata: Effects, Talents, TalentEffects, EquipStats, Sets, Factions, Classes, Races, ReqLevels, VendorPrices, SellPrices, UseEffects, ScoreModel, BiSWeights, Mounts, Hidden; Wrath also SpellSchools.
- AuraBuffs for Wrath (44 buffs), Cata (35) and Mists (31), each from that expansion's OWN raid
buffs: a curated spell-id list from WoWSims wotlk / cata / mop, every amount read from that
build's SpellEffect, and the build refuses when a curated id's name does not match. The first
cut emitted the Classic list (Wrath's Warchief's Blessing read as +500% damage, Trueshot Aura as
a flat 300 AP) and was not shipped. Cata's resistance buffs ship at 195 / 97 from the game's own
spell tooltips. A buff that fills TWO raid-buff categories (Mists Arcane Brilliance: spell power
AND crit) now competes in both, through the new
GetAuraSlotsand theslotsfield onGetAuraBuff/GetAuraBuffsranks (MINOR 39, additive): Arcane Brilliance with Leader of the Pack is 5% crit, not 10%.Tests/aura_buffs_check.pypins the values per version. - Gems, ConsumableGroups and BiS sets for Wrath and Cata; ConsumableGroups for Mists. BiS now covers all 11 Mists classes, and an earlier expansion's gear is no longer listed as BiS for a later one. ScoreModel gains Death Knight and Monk constants.
- Not shipped, on purpose: Mists Talents (tiers and specialisations, not trees).
- New desk tools (all repo-scoped):
itemdb-worlddb(tools/world-db-query.py, one read-only SELECT against a cached world DB, SQLitemode=ro),itemdb-tarball(tools/tarball-read.py, reads inside a cached WoWSims tarball without extracting),itemdb-fetch(tools/fetch-trinitycore.py) anditemdb-build-pinned(a builder at an explicit--wago-build).
Fixed -- found while building the new flavours
- Mists spell effects were wrong, some badly (shipped). Two defects of mine in
itemdb_common.spell_effect_base_points:- Every builder adds
+1to a SpellEffect's base (Classic,EffectDieSides1). Cata 4.4.2 and Mists 5.5.4 storeEffectDieSides0, so every value read one high (Flask of Endless Rage 181 on Cata against 180 on Wrath). - MoP stores many amounts as
Coefficienttimes the client's spell-scaling game table with a base of 0. Flask of Spring Blossoms shipped asITEM_MOD_AGILITY_SHORT=1; it is 1000. The multiplier comes fromgametables/spellscaling.txt(wago CASC file 1391660) at the version's cap, clamped to the spell'sMinScalingLevel..MaxScalingLevel(Spring Blossoms caps at 85: 1.9646 x 509 = 1000). Spells scaled by ITEM level (SpellMisc.Attributes_11 & 0x4, e.g. Relic of Xuen) keep their stored base.Tests/spell_scaling_check.pypins four flasks, Battle Shout and Relic of Xuen (red-checked twice). Mists Effects lost 41 items whose only "effect" was that bogus=1. Vanilla, TBC, Wrath and Forever resolve nothing and do not change.
- Every builder adds
- Wrath item procs never got their CMaNGOS rates.
build-proc-rates.pywroteproc_rates_wotlk.json;build-use-effects.pyreadproc_rates_wrath.json. Both names now come from one helper inbuild-proc-rates.py. Wrath scored procs 111 -> 313. - WoWSims proc rates were found only for Forever.
build-use-effects.pyandbuild-sets.pyderived the repo from the version name ("tbc", "mists"; real:tbc-new,mop); they now readWOWSIMS_REPO. No shipped count moved. - build-talents.py dropped Death Knights (a Vanilla class filter), read Cata's dead talent tabs
(negative
OrderIndex: "(4.0 OLD)") and wrote header-only files for a version with no trees. All three fixed; it now refuses rather than writing an empty file. - build-races.py read only the
AllowableRace_0/_1halves, so Wrath and Cata (one signed 32-bit column) read every item as unrestricted. - Retail loaded no item names at all.
Data/Retail/enUS/Names.luaput 175,229 names in one function, past Lua 5.1's 262,143 constants per function, and the client refused it ("constant table overflow", the operator's in-game report).itemdb_common.write_namesnow splits any table over 60,000 names into blocks, each in its own local function (each gets its own constant table; the[id]="name"lines are unchanged, so every reader still parses them). Files at or under that size are written exactly as before. NewTests/data_compile_spec.luacompiles every file each TOC loads under Lua 5.1, the client's own limit; it was red on exactly that file. - Retail counted no Dragonflight-onward recipe as crafting.
build-item-sources.pyknew only SpellEffect 24 (create item); modern recipes use effect 288, whose misc value is aCraftingDataid (2,450 rows at 12.1.0; its items come fromCraftedItemIDand oneCraftingDataItemQualityrow per quality tier). Effect 301 is enchanting (CraftingData Type 3, no item) and is not read. Retail ItemLocations 34,641 -> 38,048 items. - Weapon haste procs were never scored, on every flavour (found by Peer Review).
build-use-effects.pyread anItemSparsecolumn namedDelay, which no build has; the column isItemDelay. It exited 0 with no swing speed for any weapon. It now readsItemDelayand refuses to write when weapons carry damage but no speed. Scored procs: Vanilla 187 -> 190, TBC 235 -> 237, Wrath 313 -> 315; Cata, Mists and Forever rebuilt the same way. - Socket scoring was copied into four Scoring files and is now one shared scorer
(
lib.StatScorers.sockets), which also fixes gem prices going stale after new gem data loaded. The meta gem bonus (META_SOCKET_EFFECT) is now priced on TBC, Wrath, Cata and Mists. - Weapon speed is scored on Mists: Marksmanship values ranged speed, and Retribution and
Assassination value melee speed, now recorded as
MELEE_WEAPON_SPEEDbybuild-equip-stats.py(ranged keepsWEAPON_SPEED). A hunter's speed weight never counts a melee weapon and a melee spec's never counts a bow.Search { stat = "WEAPON_SPEED" }matches both keys, asstatalways matched a fragment (README). - build-hidden.py missed
QATest.../QAEnchant...names (its^QA\bpattern needed a word break after "QA"); about 100 test items per version were served. Vanilla hides 57 more. - "Where to get it" placed Dalaran's vendors in Crystalsong Forest on Wrath, Cata and Mists
(Dalaran is a dungeon-style map at 3.4.3); they now show on the Dalaran map.
build-item-places.pyalso reads the conditions table's negation flag now; no shipped "conditional" tag changed, because every affected drop already carried one. - Cata Sources:
build-sources.py Catanow runs the Encounter Journal merge itself after it writes, so rebuilding the AtlasLoot half can no longer wipe the Journal half (Tests/sources_order_check.py). Nine Cata bosses listed twice in a folded instance are merged, and Zul'Gurub and Zul'Aman are labelled dungeons. KNOWN COSTS: the Cata Journal loot of Venoxis and Mandokir sits under their Vanilla NPC ids (14507 / 11382); the Vanilla AtlasLoot bosses still list beside the Cata ones in the four folded instances; Retail's Jadefire Masters keeps one row for its two faction versions, since a boss row has no faction field.
Changed -- one profession resolver, which understands Retail's expansion child lines
Retail splits each profession into per-expansion skill lines ("Midnight Alchemy" 2906 is a child of
Alchemy 171) and hangs recipes off the child, and its CategoryID 11 also holds minigame lines
("Abominable Stitching", "Supply Shipments"). The first Retail run reported 29 "professions".
itemdb_common.profession_lines(build) now walks ParentSkillLineID to the root and, on a build
that gives any profession a SpellBookSpellID, keeps only roots that have one (Classic Era sets it
to 0 everywhere, so there the filter is off). build-reagents.py and build-item-sources.py
(craft_professions) both read it, so "used by" and "made by" cannot disagree; they were two
copies that had already diverged.
build-reagents.py also reads the ModifiedCrafting* tables on 9.x+ builds: modern recipes take
quality materials and optional reagents through crafting SLOTS, not SpellReagents. Before this,
Bismuth (210930) listed only Mining; it now lists Blacksmithing, Engineering, Jewelcrafting and
Cooking, all from the client's own slot data (Cooking recipes 445482 and 1245993-1245995 take the
"Common Ore" category; not confirmed in game).
build-expansions.py runs the first-ships-it chain through Vanilla, TBC, Wrath, Cata and Mists;
Retail continues with its own ItemSparse.ExpansionID (5 Warlords .. 11 Midnight; the numbering is
the live Encounter Journal's tier table, Blizzard_EncounterJournal.lua:106-122) and lists items
where that is unset too under -1. GetExpansion returns nil for -1 rather than a number that is
not an expansion; the tooltip's Expansion line now names Wrath through Shadowlands in SmexyMats'
colours (SmexyMats/Utilities/Utilities.lua:29-38). The default bucket is now the largest one,
which keeps Vanilla's and TBC's files unchanged in shape.
Instance names in Sources come from the client's Map table (via the block's InstanceID and the
same INSTANCE_NAME_ALIAS item-sources uses) when the key is not in SOURCE_INSTANCE_NAMES, a
map no other block shares, and the name does not collide with another instance.
Fixed
- Malformed instance names on TBC and Mists. Every AtlasLoot key outside the hand-kept Vanilla table was humanized from its CamelCase key, which shipped "Hallsof Lightning", "Trialofthe Champion", "Ahn Kahet", "Gruuls Lair", "Zul Aman" and "Magisters Terrace". They now carry the client's names ("Halls of Lightning", "Ahn'kahet: The Old Kingdom", "Gruul's Lair"). Consumer- visible: several TBC instances now read with the client's prefixes ("Coilfang: The Slave Pens", "Hellfire Citadel: Ramparts", "The Sunwell"). Vanilla is unchanged.
- Mists listed Pandaren cooking specialisations as professions. "Way of the Grill" and its five
siblings shipped as six separate professions in
Data/Mists/_core/Reagents.lua; they are now folded into Cooking (same 969 items, 11 professions instead of 17). - TBC Expansion.lua was stale against the current Vanilla and TBC cores (17,568 / 12,516); regenerated with unchanged logic: 17,490 Classic / 12,542 TBC.
- Expansion.lua stamped a build it never read. Every flavour's file carried the live client's
client build, though only Retail reads a DB2 (the rest derive from the shipped item sets). Once the Era client moved to 1.15.9.70003 that put a second build into Vanilla's core group andprovenance_specwent red. Non-Retail files now sayclient build: none readand carry asource tools/build-expansions.py sha256:line (the rule lives in the builder, as with ScoreModel.lua); Retail keeps its build stamp too. Vanilla'sReagents.lua, regenerated only to prove the shared profession resolver left it unchanged, went back to HEAD for the same stamp. Vanilla's count now reads 17,604 (was 17,682), matching the class cores today; which change moved it was not traced. - Three builders called a validating helper on a branch (
build-expansions.py'sshipped if ... else core_ids(v),build-item-sources.py's client-only path,build-sources.py's comprehension), whichitemdb_common_checkrefuses. Each call is now unconditional. build-rep-loot.pywrote an empty file when it had nothing (Cata); it now writes nothing and removes a stale one.
Tooling
tools/wago-schema.py+ theitemdb-schemadesk tool (approved 2026-10-04): print a wago DB2 table's columns, row count and first row at a version's pinned build. It replaces throwaway shell probes; the operator's directive the same day is that tool gaps are fixed, not shelled around.- The
itemdb-builddesk tool now lists the new versions in its help, through Writ'stools updateaction (shipped on request, inbox8aa52bc3e697), with a 600-second limit. - New
Tests/constant_headroom_check.py: fails when any shipped Data file's busiest function passes 80% of Lua 5.1's 262,143-constant limit -- the warning beforedata_compile_spec, which only fails once the game already refuses the file. Measured today: the largest unsplit files are Mists' Names.lua at 62% (they split on their next rebuild, being over 60,000 names), then Retail's Expansion.lua at 57%. Peer Review suggested one shared chunked writer for every builder; not done, because no other file is near the limit and the builders cannot all be re-run here to prove an unchanged output -- the check names the writer to split when one gets close. - Harness pin
b1d9bd5->407bd26: an error escaping the runner is now a named FAIL with a traceback instead of a bare(error object is not a string), which is whatdata_compile_spec.luahit. Adoption asks only for the pin; the suite is green on it. itemdb-schematakes its table list separated by commas OR spaces; a space-separated list used to reach the fetch as one table name and crash on the space in the URL.- CLAUDE.md: builders run through
itemdb-build, not the shell; that paragraph used to say the opposite.
[v1.1.4] - 2026-09-29
Fixed -- ordinary vendor goods read "PvP" on TBC and Mists
Refreshing Spring Water (159) showed only "Alterac Valley / PvP" as its source on TBC, with none
of its innkeepers, vendors or quests (SmexyMats peer review 60e304d70cab, found while generating
its tooltip data from ours). The cause was in tools/build-item-sources.py: the battleground
reward-vendor pass REPLACED an item's whole location set with the battleground whenever a
BG_VENDORS NPC sold it, exempting only ammo. TBC's Stormpike Quartermaster (12096) stocks common
food, water, reagents and flasks, so every one of those lost its real sources. The pass now skips
any item a vendor outside a battleground also sells (sold_outside_bg).
- TBC: about 60 items corrected (water, bread, flasks, reagents, arrows), e.g. 159 now reads Vendor and Quest. Item 1 (a placeholder id, "World Drop") also left the file on this rebuild.
- Mists: about 70 items corrected, the same pattern.
- Vanilla: unaffected -- the classic DB's quartermaster stocks none of them, and 159 already
read Vendor and Quest. Not rebuilt:
.build.infono longer lists the Era product, so the builder needs--wago-build, which the desk'sitemdb-buildtool cannot pass. - A trade-off, stated: the battleground-only healing draughts (17348, 17349, 17351, 17352) on TBC and Mists, and the Warsong tabards (19505, 19506) on Mists, now read Vendor instead of their battleground, because the CMaNGOS / SkyFire DB has a vendor outside a battleground selling them.
Changed -- the "Which click" chooser is LibAceGUIWidgets' capture widget
The /itemdb Tooltip tab's "Which click" control is now LAGW-ClickBinding (LibAceGUIWidgets
MINOR 38, built for this at TOGProfessionMaster's request, contract e0f3923e3548): the player
clicks it, then performs the combination. It has a Default button that restores Alt + Left Click,
read from the tooltip defaults through a new private lib:_GetTooltipDefault(key). It also allows
the middle and extra mouse buttons, which ParseClickBinding already accepted. On an older
LibAceGUIWidgets the dropdown of 14 combinations stays. Both store the same canonical string, so no
saved value changes. win.clickDropdown is renamed win.clickChooser.
Changed -- test harness pinned at b1d9bd5 (was 20453ac)
From 9c6c874 on, the harness's libs.load("LibAceGUIWidgets-1.0") includes the ClickBinding file,
so the spec no longer loads it itself. No other Adoption entry in that range needed a change here;
the suite is 777/777 at the new pin. The new pkgmeta.lua checker (a desk tool from 6eb30e0, asked
for on inbox d6eb270e8b4f) reports no FAIL on .pkgmeta. Its three WARNs were left as they are:
the "*.sh" and "*.bat" ignores match no tracked file today, but they are guards against a
future script shipping; and curseforge-project-id: is not needed in .pkgmeta because every TOC
carries X-Curse-Project-ID: 1558208.
Changed -- release docs
README: the source-locations pipeline paragraph states the battleground-vendor rule, and the
tooltip-settings row says whereClick is set on the LAGW-ClickBinding control (dropdown on an
older LibAceGUIWidgets). The CurseForge page gains the v1.1.4 notes. Checked against both files
for this release's changes: no item count, build number or feature row they quote moved.
Fixed -- a stale comment about VersionCheck-1.0
ItemDB.lua and CLAUDE.md said VersionCheck-1.0 is Classic-only. Its TOC declares Retail too
(120005, 120007, 120100); both comments now say so (SmexyMats peer review 34ee062fd221).
[v1.1.3] - 2026-09-28
Changed -- WoW Forever's items now come from the client's own item walk
Until now Forever's _core was synthesized by tools/build-core-wowsims.py: structure from wago,
stats from WoWSims' db.json, because no Forever client had been walked. The operator ran the
TOG Tools Item DB walk on Forever, and tools/build-core.py Forever now builds the core from it,
as every other flavour's is. What that changed, measured against the WoWSims build:
- 23,596 items, up from 19,171. The first walk stopped at item ID 240,000 while Forever's run
to 287,090; TOG Tools raised its ceiling (inbox
500e4ce2) and the re-walk carries them. - Weapon DPS on every weapon. The WoWSims build left most white weapons with no stats at all, so they scored 0 EP for every spec (Dibs, DIBSREQ-IDB-009: Copper Mace over Anvilmar Hammer showed "Dibs EP: 0").
- Spell power on 1,098 items that had none, and other stats WoWSims simply lacked -- e.g. item 20341 shipped with armor only and now carries its +39 frost damage and 28 Intellect.
- 249 school-tagged spell-damage items, up from 171.
- 131 items removed: placeholder IDs with no name (1-19 and similar) and two unused Vanilla
rings (18674, 18678) that the Forever client does not return. Their rows left the side tables
with them (
ReqLevels,SellPrices,ConsumableGroups,Sources).
Added -- build-core.py translates a Forever walk into the keys its scoring reads
Forever loads Scoring/Vanilla.lua, which prices crit and hit as flat percentages, mp5 as
ITEM_MOD_POWER_REGEN0_SHORT and damage-and-healing as ITEM_MOD_SPELL_POWER -- the Classic Era
walk's own spellings. The Forever client reports ratings and _SHORT keys instead, and shipped raw
every rating stat would have scored 0. WALK_VOCAB / FOREVER_KEYS / normalize_forever map each
walk key onto the Classic one. The mapping was derived from data, not names: every walk key was
cross-tabulated against the WoWSims-built core over the 19,161 items both carry (crit 14:1 onto
CRIT_PCT on 488 of 521 items, hit 10:1 on 172 of 183), and the divisors are read from wowsims'
own constants, which agree exactly. Per-school damage folds onto ITEM_MOD_SPELL_DAMAGE_DONE with
a SpellSchools.lua tag, as the Era walk ships it; the client's doubled resistance keys count once.
Hit and crit now score for casters on Forever. Forever merged them ("one stat now covers every
attack", outputlag.com), and item 21624 is +1% spell crit in Era's data and crit rating 14 in
Forever's. The WoWSims build mapped them to the melee key alone, so a caster got nothing for them;
each now lands on both HIT_PCT/SPELL_HIT_PCT and CRIT_PCT/SPELL_CRIT_PCT.
Weapon skill on Forever items is priced. The walk reports it as ITEM_MOD_DAGGERS_SHORT and
so on, in skill points -- the same unit as Classic: Death's Sting (21126) is dagger 3 and Hands of
Thero-shan (7951) dagger 4 in both clients, while items Forever retuned shrank (Huge Thorium
Battleaxe 10 to 2). The seven spellings the walk carries map onto WEAPON_SKILL_*, reaching nine
items; other spellings are unverified (the client's string table is not in the UI source) and
stay unmapped. Tests/forever_scoring_spec.lua's weapon-skill tripwire was re-decided to eleven.
rating_divisors, RATING_CONST, the school fold (fold_school_damage) and the SpellSchools
writer moved from build-core-wowsims.py to itemdb_common.py, so the two builders cannot
disagree. build-core-wowsims.py now refuses to overwrite a walk-built core.
Tests/forever_vocab_check.py (new, declared in .writ-suites.json) pins the translation.
Added -- drop sources for three of Forever's new dungeons
tools/build-sources-imenso.py had never shipped: the README said the new dungeons' item ids
were not in Forever's item table. The walk carries them, and it now writes
Data/Forever/_core/SourcesImenso.lua + SourceNamesImenso.lua (added to ItemDB_Camelot.toc):
33 items across The Hall of Thanes, City of Dalaran and Ruins of Lordaeron, 11 bosses, under the
source "imenso" with synthetic encounter ids. Excavation Site resolves too but none of its loot
is in the walk yet.
Two of the three only resolved after a second tie-breaker in resolve_area: a name shared by two
areas used to need the shipped Sources.lua to settle it, and a new dungeon is by definition not
there. instance_areas now reads the client's own Map table -- the area on a dungeon/raid map
(AreaTable.ContinentID whose Map.InstanceType is 1 or 2) beats its outdoor namesake (Ruins of
Lordaeron 16611 over 153), and two areas on one dungeon map are settled by Map.AreaTableID
(City of Dalaran 16544 over 16560). Still a refusal unless it narrows to exactly one; the shipped
file still wins when it has committed. Five cases added to Tests/imenso_reference_check.py.
Fixed -- core_ids let removed items back in through the side tables
itemdb_common.core_ids read every _core file except a hand-kept exclusion list, so side tables
(SellPrices, ReqLevels, ...) fed their own ids back into the item set every builder gates on.
After the walk rebuild it still returned the 131 removed ids, and a removed item could never
leave. It now reads only files that call lib:LoadCore.
Fixed -- two builder guards that misread the rebuild
build-locales.pyrefused Italian for "losing" one localized name. The name was English: the client's'Formula: Powerful Smelling Salts 'carries a trailing space and wago's does not. The localized-count comparison now ignores surrounding whitespace.build-sources-wowsims.py's expected-leak list gains seven TBC/Wrath instances (Shadow Labyrinth, Auchenai Crypts, The Botanica, The Mechanar, The Oculus, The Nexus, Ahn'kahet). Each was refused because Forever's AreaTable lacks it -- the same id-collision leak as the two before.
New -- the "Where to get it" click is the player's choice, Alt+Click by default (MINOR 38)
TOGProfessionMaster's contract 060eb26d, from a player on Discord: Alt+click "collides with
gargul". The operator: "we make it alt+click by default, but make it so the users can change it".
- New tooltip option
whereClick, default"ALT-LeftButton", in WoW's binding spelling.SetTooltipOptionvalidates it through the newParseClickBindingand stores it canonical; a binding with no modifier is refused, because an unmodified item click is the game's own. _WhereOnModifiedClicknow asks the newIsWhereClickinstead ofIsAltKeyDown. Modifiers match exactly, so a player's Alt binding does not fire on Shift+Alt. The button is compared whereGetMouseButtonClickedanswers (documented in the forever tree'sInputDocumentation.lua; listed in the classic and anniversary restricted environments), and the modifiers decide alone where it does not, sinceHandleModifiedItemClickhands its hook only the link. Not verified in game on any client.altClickstays the on/off switch, so a player who saved it off before this is still off.- The tooltip heading's hint names the player's click (
ClickBindingText): "Ctrl+Shift+Click: where to get it", not a fixed "Alt+Click". - The
/itemdbTooltip tab gains a "Which click" dropdown beside the switch, filling that row's empty half so the tab still fits. It offers every modifier combination with the left or right button. The contract asks for a click-capture widget from LibAceGUIWidgets (their contracte0f3923e3548); the dropdown stores the same value, so that widget can replace it later without a data change. - Specs: a non-default combination opens it and the old default no longer does, an extra
modifier does not, the button is honoured when known, an old
altClick = falsestays off, bad bindings are refused, an unreadable stored value never matches, the hint and the dropdown.
Fixed -- a full Auction House scan on Forever raised an error and stored no prices
The operator hit it on Forever: Frame:UnregisterEvent(): Attempt to unregister unknown event "AUCTION_ITEM_LIST_UPDATE", from fullUnregisterEvents inside fullStore. The modern
ReplicateItems scan never registers the legacy full-scan events, but fullStore unregistered
them unconditionally, and a client with the modern Auction House has no AUCTION_ITEM_LIST_UPDATE
to unregister. The error fired before a single price was written, so every full scan on that
client was discarded. That covers Forever; Mists takes the same modern path and is very likely
affected too (not verified in game). fullRegisterEvents now sets S.fullEventsRegistered and
fullUnregisterEvents undoes only what it registered. The spec's FrameUtil stub accepted any
unregister, which is why the suite never saw it; the modern full-scan spec now raises exactly as
the client does, and fails with the old line put back.
Fixed -- the Auction House scan button sat below the title bar on Forever and Mists
The operator's screenshot, Forever: the "ItemDB Scan" button hung half under the title bar, over
the Browse pane. showScanButton in Price/Scanner.lua anchored it at a fixed TOPLEFT 72,-15,
an offset laid out for Classic Era's legacy AuctionFrame. The modern AuctionHouseFrame (Mists
and Forever) is a PortraitFrameTemplate whose title bar is a 20-high TitleContainer at
TOPLEFT 58,-1 (SharedUIPanelTemplates.xml in the forever tree; PortraitFrame.lua's
SetTitleOffsets in the classic tree gives the same 58/-1). The button now centres on that bar,
4 px in from its left edge, whenever the frame has one; the legacy frame keeps its offset. Two
assertions in Tests/pricescan_spec.lua pin both placements. Not yet seen in game.
Fixed -- the dev replication watcher deleted live files from the other installs
wow-version-replication.ps1 (dev-only; never packaged) treated a save-by-replace as a delete. An
editor that saves by replacing the file raises Deleted and Created/Changed for one save, in
either order, and when Deleted landed last the file was removed from every replica install while
the source still had it. Sync-File now honours a delete only when the source file is really gone
and otherwise copies it.
Changed -- docs brought up to date for the walk-built Forever core
README.md and docs/Curseforge_Description.html state Forever's item count as 23,596 (was
19,171), say its stats come from the client walk rather than WoWSims, and list the three new
dungeons' drops in the per-version feature tables.
Tests
Tests/forever_weapon_dps_spec.lua(new): Copper Mace and Anvilmar Hammer carry the client's DPS and both score above 0 for a protection paladin on the shipped Forever data, the mace higher.Tests/forever_scoring_spec.lua: Forever's weapon-skill population is now two -- the rebuiltSets.lua(150 to 261 sets) adds Defias Leather's +1 dagger bonus. Re-decided and accepted, as that spec asks: same gated function, one attack shape.
[v1.1.2] - 2026-09-27
Fixed -- opening a vendor raised "attempt to call a nil value" on WoW Forever
The operator hit it on Forever: Price/Sources.lua:1076, inside GetMerchantItemInfo, fired
from the MERCHANT_SHOW merchant capture. Forever still defines the bare global, but calling it
fails inside the C function; the Forever client's own MerchantFrame.lua (Mainline) calls only
C_MerchantFrame.GetItemInfo, which MerchantFrameDocumentation.lua documents as returning a
MerchantItemInfo table (price, stackCount, numAvailable, ...). I had written the capture
against the bare global only, so every vendor visit on Forever errored and cached no prices.
captureMerchant now reads through a merchantItem helper that prefers the namespaced function
wherever it exists and falls back to the global on the flavours without it, and skips a slot when
neither exists. Two specs in Tests/price_spec.lua pin it: the namespaced call wins over a global
that raises, and a client with neither API skips the slot. .luacheckrc gains C_MerchantFrame.
Checked per flavour in F:\Blizzard API Docs: the classic_era, classic (Mists) and
classic_anniversary trees all declare the C_MerchantFrame namespace, but with
GetBuybackItemID as its only function -- no GetItemInfo -- so those clients still take the
bare-global branch exactly as before. Only Forever takes the new one.
Changed -- the feature tables split vendor prices into shipped and captured
README.md "Features by version" and the CurseForge "What works on which version" table had one
"Vendor buy and sell prices" row, which covers the shipped vendor price data only (Mists: No). The
merchant capture above is a second, separate source -- GetVendorBuyPrice answers
source = "merchant" from it on every flavour, Mists included -- and was listed nowhere. Both
tables now carry it as its own row; the README marks Forever's cell "from v1.1.2".
[v1.1.1] - 2026-09-27
Changed -- a per-version feature table in the README and on the CurseForge page
The operator asked for the docs to state feature parity across versions. README.md gains
"Features by version" and docs/Curseforge_Description.html "What works on which version": one row
per feature, one column per shipped flavour (Classic, TBC, Mists, Forever). Every cell was read
from which Data/<Version>/_core/*.lua files exist and which each TOC loads, not from the prose
that was already there. That makes explicit what the prose only implied: Mists ships no class,
faction or required-level data, no sets, buffs, talents, consumable groups, scoring, BiS, mounts
or vendor prices, and Forever has no gathering source and no "Where to get it" data.
Fixed -- every item link raised "attempt to call a nil value" on WoW Forever
The operator hit it on Forever, through Questbook's search: LibItemDB-1.0.lua:370, inside
qualityHex, called from buildLink. I captured GetItemQualityColor as the bare global only,
and Forever (1.60.x) ships just C_Item.GetItemQualityColor -- same four returns, the fourth the
hex string (wow-ui-source-forever ItemDocumentation.lua:920). The capture now prefers
C_Item and falls back to the global, like the class-info captures beside it. Every caller of
buildLink was affected on Forever: GetLink, Search rows, suffix links and scoring rows.
Tests/forever_api_spec.lua (new) loads the library into a Forever-shaped environment through
wow.loadAddonFile's env argument and builds a link; with the old line put back it fails.
New -- Search takes loose, a punctuation-insensitive name match (MINOR 37)
Asked for by TOGBankClassic (inbox e860cddb848a): its List Setup tab searches LibItemDB, and the
operator could not find the E'kos (12430-12436) by typing "eko", because Search did a plain
substring find on the lowered name. opts.loose = true now strips ' - : , and folds whitespace
runs on both the query and the name (looseForm in LibItemDB-1.0.lua). The loose names are
built once into lib._looseNames on the first loose search and dropped by LoadNames, so the
name-first early-out stays cheap on Mists' 88,259 items. Without the flag the result is unchanged.
Five examples in Tests/libitemdb_spec.lua (opts.loose) pin both halves and the index rebuild.
Changed -- six spec assertions that could not fail
Writ's new spec check (inbox 676ec4263aeb) flagged six is_nil assertions on a literal key, which
still pass when the literal is misspelled. Each now checks something that can fail:
itemdb_spec: the bareGetAddOnMetadatais a counting trap that must never be called.libitemdb_spec: every library key containing "sell" is listed and compared to the known set.libitemdb_spec: the whole stored tooltip table stays unchanged after a refused write.libitemdb_spec: a stand-inGameTooltipmust receive no lines.price_spec: the store's realm keys are exactly the player's own scope.
[v1.1.0] - 2026-09-24
New -- "Where to get it" on Mists, from SkyFire 5.4.8's world database; Mists stats re-walked
The operator, 2026-09-24, after I had called Mists blocked three times without searching: "i
think YOU need to do better web research, this data has to be out there". It was. The CMaNGOS
projects stop at Wrath, but SkyFire 5.4.8 publishes a full world database as a GitHub release
(ProjectSkyfire/SkyFire_548, tag sf_db_26, SFDB_full_548_26.002_2026_008_18_Release.zip,
21.6 MB, one 121 MB MySQL dump). Other 5.4.8 cores (Legends-of-Azeroth, PandariaCore) were
checked: none publishes a full dump; Legends-of-Azeroth builds on a local SkyFire copy.
tools/mysqldump-to-sqlite.py(new) turns the dump intotools/cmangos_cache/mistsskyfire.sqlitewithout a MySQL server: 224,734 creature spawns, 368,727 creature loot rows, 47,208 vendor rows, 70,934 object spawns, 9,453 conditions. Its value reader is pinned initem_places_check.py.build-item-places.py Mists. SkyFire's schema differs from CMaNGOS in four places, each detected from the tables rather than keyed on the version: faction isfaction_A/faction_H(creature_sides); loot conditions are TrinityCore-style rows keyed by loot entry and item, ANDed within anElseGroupand ORed across them, with "quest not started" as type 14 where CMaNGOS uses 22 (trinity_loot_tags; the other type numbers checked against the DB's ownCommenttext);npc_vendorcarries currencies (type2, skipped) and negative items that reference another vendor's list (expanded; TrinityCore-lineage behaviour, not verified in SkyFire's source); and there is no vendor-template table (cmangos_vendor_rowsnow checks). Open-world maps now come from the client'sMap.InstanceType(0) where the map has zone rectangles, instead of the four pre-Cataclysm continents alone, so Pandaria, Deepholm and the other later maps place their spawns. Classic and Burning Crusade rebuild to the same counts.- SkyFire's Pandaria gathering nodes have no loot link (
gameobject_template.data1 = 0): 36 Mining/Herbalism nodes, among them Ghost Iron Deposit's 265 spawns and Green Tea Leaf's 287, and no table names Ghost Iron Ore as a node's yield. Searched for a fix upstream and in Legends-of-Azeroth's patches; none. A lootless gathering node now yields the item named after it (node_yield_names: a herb node is its herb, a deposit yields<name> Ore, Kyparite and Trillium listed), with no chance rather than an invented one; 23 of the 36 resolve, and the 13 that do not are quest or event herbs. An object token with no chance ships as a bareg<id>. - Built:
Data/Mists/_core/ItemPlaces.lua, 25,270 items and 10,215 creatures and objects (drops 42,634, of which 843 conditional; vendors 28,102; objects 11,470), plus eight MistsPlaceNames.luafiles borrowing the TBC translations through the rename guard. All nine are inItemDB_Mists.toc. Spot-checked: Ghost Iron Ore (72092) lists its Ghost Iron veins besides two creature drops and two vendors, and the veins sit in The Jade Forest (155 spawns), Kun-Lai Summit, Dread Wastes, Krasarang Wilds and Townlong Steppes with map points. The first Mists build placed them as@Pandaria:263, the whole continent named like an instance: the open-world set read the client'sMaptable a second time fromwago_csv, which returns a one-passcsv.DictReaderalready consumed by the line above, so only the four old continents counted. Caught by reading the built file; the rows are now read once into a list. Not tried in game. - Mists gets
ItemLocations.luatoo (the tooltip's Source line andGetSources' quest / vendor / crafted / drop / gathered rows):build-item-sources.py Mists, 37,228 items (crafted 5,898, drop 11,046, gathered 864, quest 8,632, reputation 1,579, vendor 12,136, pvp 348), inItemDB_Mists.toc. Its SkyFire differences, again detected from the tables: quests hang offcreature_queststarter/creature_questenderwithReward*ItemIdcolumns keyedId; there is nospell_template(a 5.x server reads the client's DBCs), so crafted items come from the client's ownSpellEffecton wago (create_item_spells); skinning isskinloot/type_flags; and the lootless Pandaria nodes use the samenode_yield_namesas the places builder. The open-world test and the node rule moved intoitemdb_common(open_world_maps,node_yield_names) so the two builders cannot disagree; Classic and TBCItemLocations.luarebuild byte-for-byte the same size andItemPlaces.luato the same counts. Spot-checked: Ghost Iron Ore reads Mining, Ghost Iron Bar Mining (smelting), Green Tea Leaf Herbalism, Mist-Touched Leather Skinning and Leatherworking. Known SkyFire gap: a few creatures that are mined or herb-gathered are not flagged so in its data (Vengeful Hui, Quilen Statue, Doom Bloom), so Ghost Iron Ore and Green Tea Leaf also read Skinning. Windwool Cloth reads only Vendor, because the drop fill only places items nothing else did -- the same rule every version uses. Forever still has no gathering source: it is Blizzard's own game, releasing November 2026, with no emulator database, and Wowhead's Forever data is web pages, not a download. - Mists item stats re-walked on the operator's fresh 5.5.4 walk: 88,265 items (6 new), about
67 Armor and 16 Weapon rows changed. Four new items wago's
Itemtable does not know yet went to anUnknown.luano TOC loads;build-core.pynow files such an item under the class the walk itself recorded (parse_walk(..., item_class)), and that file is gone.provenance_spec's list of unstamped files losesData/Mists/enUS/Names.lua, which the rebuild stamped. The other Mists languages do not have the 6 new names yet.
Changed -- harness pin moved to 20453ac; the changelog's v1.0.0 section archived
Harness. WoWAPITesting delivered the FontString fix this workspace asked for (thread
0633f2ae): regions now have SetScript, and anything that is not a Button refuses OnClick with
the client's message, so the crash the operator hit in the search strip would now go red offline.
The pin moved from 5a75f26 to 20453ac. The adoption entries in between ask only for stubs to
be deleted for APIs ItemDB does not stub. 755 examples pass on it, with LibItemDB-1.0.lua,
Where.lua, Tooltip.lua and Price/Window.lua at 100% line coverage. where_spec's ban on
RowList:SetColumns stays: the window uses two lists by design, not as a stand-in.
Changelog. The live file reached 119,581 characters, 419 under the 120,000 working ceiling.
The released v1.0.0 section (1,214 lines) moved byte-for-byte to the top of
CHANGELOG_ARCHIVE.md, directly above v1.0.0's already-archived earlier entries, so the two halves
now sit together. Both joins were read back and both files lint clean.
Fixed -- the /itemdb Tooltip tab ran past the bottom of the window
Seen in game by the operator, 2026-09-24: "the tooltip tab extendes beyond the window
borders". Nine full-width checkboxes, the icon-size slider and both colour pickers were stacked
in the TabGroup's List layout, taller than the 400-px window, so the two pickers were drawn over
the status bar. buildTooltipTab now uses two columns: _PaintPriceWindow sets the TabGroup to
Flow on the Tooltip tab and back to List on the others. The checkboxes are
SetRelativeWidth(0.5), and the slider (0.4) shares a row with the two pickers (0.3 each). A new
pricewindow_spec example asserts every control's bottom edge sits inside the tab's content
frame. It was red-checked by putting the one-column layout back, which fails that check.
Price/Window.lua stays at 100% line coverage. Verified in game by the operator, 2026-09-24:
"tooltip is fine now", and a picked colour survives /reload: "the color stays".
Fixed -- a drop the server only gives under a condition read as a plain chance (MINOR 36)
Peer Review, thread f73d04403a2c finding 7: build-item-places.py shipped creature_loot_template
rows that carry a condition_id (2,314 Vanilla, 3,886 TBC, measured) as ordinary chances, so a
Horde-only quest drop or a Blacksmithing plan read as something anyone gets. Measured what the
conditions are before labelling them, by type and by each condition row's own comments text:
team (Horde 1,242 / Alliance 147 rows on Vanilla), a profession skill (435 are Blacksmithing), quest
rewarded / taken / not started, a game event or holiday, and a long tail (items carried, auras,
Argent Dawn commissions, AND/OR combinations).
- Builder:
condition_codeturns a condition into a short tag --A/H,s<skill line>,e,qd,qt,qn, orc-- and an AND/OR takes its halves' tag when they agree. A drop with a tag ships it as a third field,d<creature>:<chance>:<tag>(drop_token). No creature carries the same item on both a conditional and a plain row (measured, 0 pairs on both DBs), so the winning row's tag is the drop's. Pinned initem_places_check.py(46 cases). - Built: 407 of Classic's 11,552 shipped drops carry a tag, and 709 of TBC's 13,948. Item and place counts are unchanged; the data-shrink check passes.
- Library:
GetItemPlacesrows gaincondition("Alliance","Horde","skill"withconditionSkill,"event","questDone","questTaken","questNotStarted","other"). - Window: the Chance column reads e.g.
3.0% (Horde),<0.1% (Blacksmithing),25.0% (after a quest),5.0% (event),1.0% (conditional). The profession is named byC_TradeSkillUI.GetTradeSkillDisplayName, which every flavour's API docs list, so it reads in the client's language; the factions useFACTION_ALLIANCE/FACTION_HORDE. The other labels are English, like the existing(quest). A quest-taken condition on a quest-only drop still reads(quest)once. The window's "i" explains each label. - Faction-only drops follow the faction switch. With "My faction's vendors" on, a drop tagged
for the other faction is left out of
BuildWhereRowsalong with that faction's vendors. Peer Review asked for this over shipping it as a known gap (thread c66dfa1ad14c); pinned inwhere_spec. The switch's label still names vendors only.
New -- translated creature, vendor and node names in "Where to get it" (MINOR 36)
Peer Review, thread f73d04403a2c finding 6: the window showed "Rock Elemental" in "Ödland" on a
German client, because build-item-places.py read only creature_template.Name and
gameobject_template.name. It now also reads the CMaNGOS locales_creature and
locales_gameobject tables and writes Data/<V>/<locale>/PlaceNames.lua for deDE, esES, esMX,
frFR, koKR, ruRU, zhCN and zhTW on Vanilla and TBC (16 files, each guarded on GetLocale(),
listed in ItemDB.toc and ItemDB_TBC.toc). A file holds only names that differ from English.
lib:LoadPlaceNames fills lib.placeName, and GetItemPlaces rows now carry name in the
client's language with English as the fallback, plus englishName, always English. The
boss-deduplication in BuildWhereRows compares englishName, because the drop graph's boss names
are English. Coverage on Vanilla, measured from the build: 4,710 of 4,723 entities in Russian.
Both versions take their translations from the TBC database: classicmangos.sqlite's locale
tables are empty (0 rows, measured), and entries are the same numbers in both. A creature or
object TBC renamed would have shown its TBC-era translation on a Classic client. Measured at
Peer Review's request (thread c66dfa1ad14c): 16 shipped Classic entities, among them Horde Grunt
(TBC: Warsong Grunt), Ghoul (Rotten Ghoul), Draka (Drakan) and six "Alliance Chest"s (Tattered
Chest). place_name_rows now drops a translation whenever the translating database's own English
name differs from this version's, so those 16 stay English; each Classic language file lost
exactly 16 rows (ruRU 4,710 -> 4,694). Pinned in item_places_check.py.
These 16 files, and the 10 new Mists files (ItemLocations.lua, ItemPlaces.lua and eight
PlaceNames.lua), are NOT yet in Tests/data_baseline.json; that file's floors are set with
--record, which is the operator's.
New -- Italian item names on Forever (12 languages)
tools/build-locales.py Forever --locales itIT wrote Data/Forever/itIT/Names.lua: 17,779 of
19,171 names in Italian and 1,392 (7.3%) English fallback. The guard reported the fallback rate
and wrote the file anyway, as designed, because nothing was on disk to lose. The run exits 1
because the RandomProps step got a 404 from wago for ItemRandomProperties at 1.60.1.69913.
That's expected: Forever has no random-suffix table, and no Forever language ships a
RandomProps file. The file is in ItemDB_Camelot.toc, and the data-shrink floor now records it
along with the two new ItemPlaces.lua files. Those three floor entries were added by hand
rather than with --record, which would have re-set every existing floor too.
This waited a day on a blocker I invented: "a builder with arguments needs an approved writ tool". No law says that. The builder runs in the shell like any build tool.
New -- the "Where to get it" window, opened by Alt+click or a key binding (MINOR 36)
The operator's design, 2026-09-24: "use libaceguiwidgets for the UI and lets make it like FGI, a strip on the top and zebra stripped rows to show the info with columns".
Where.lua(new, in every TOC afterPrice\Window.lua, and intoc_spec'sCORE_ORDER). It uses LibAceGUIWidgets' ClearFrame for the chrome. A strip across the top of the content holds the item's icon and link, a place count, and a text button that switches between "My faction's vendors" and "Every vendor". It's built the way FGI'sGUI/Tabs/Scan.luabuilds its strip.W.RowListsits below on its own area frame, as FGI'srowsAreadoes. Columns: Source, Name, Zone, Chance, Spawns. Hovering a row lists its map coordinates. Clicking one callsWhereTrack.- No timer for the first paint. My first draft carried the price window's
C_Timer.After(0)refresh, and writ refused it. RowList already hooks its parent'sOnSizeChangedand refreshes when the client first sizes the new frame, so the timer was never needed here. The spec asserts that the rows draw on the first open. BuildWhereRows(id, faction)is frame-free and mergesGetItemPlaceswithGetSources. Boss rows with their rate come first, then place rows, then quest / crafted / reputation / PvP. A generic "World Drop" or "Vendor" location row, or a gathering profession an object row already names, is not repeated. Forever, which has no places data, still shows them.WhereTrack(row)converts the chosen point withC_Map.GetWorldPosFromMapPosand hands{ name, kind, world = { continent, x, y }, location = { zoneName } }to Questbook'sTrackPlacewhen that method exists. That's the target shape Questbook's privateTracker:Trackalready takes (Modules/Tracker.lua:115-127). Requested as contract LIBREQ-QB-TRACKPLACE; Questbook has since shipped it (Questbook/Core.lua:333) with exactly that signature and a boolean return, so no change was needed here. On a Questbook without it, clicking a row does nothing and the hover shows no "click to be guided" line. Verified in game by the operator, 2026-09-24: "validated tracking is working".- Stop icon (
WhereStopTracking,_WhereShowStop): Questbook's own stop, beside the window's "i", dressed withW:DressBottomRowthe way Questbook'sChrome:Dressdresses its row, with Questbook's texture and its 0.4 idle alpha. Lit while Questbook is guiding, dimmed otherwise, and the tooltip reads the state at hover. It is dressed only when Questbook offers a publicStopTrackingwhen the window is built. Requested as LIBREQ-QB-STOPTRACKING (thread 96906d235c7b) withIsTracking; Questbook shipped both under those names (Questbook/Core.lua:354and:362), uncommitted there as of this entry. Verified in game by the operator, 2026-09-24: "the stop button is there and working". The alpha does not follow a stop made inside Questbook while this window is open (there is no listener); the tooltip is still right. - Search strip (operator, 2026-09-24: "add a strip across the top to let you search ItemDB
for an item too in the popup"). LibAceGUIWidgets' own
CreateSearchBox(magnifier, placeholder, clear-X) spans a strip across the very top, above the item strip. My first cut hand-built anInputBoxTemplatebox instead, and the operator pointed at the library's. Typing three letters, or pressing Enter, callsWhereSearch. That runslib:Searchon the name (any case, capped at 100, "100+ matches" when cut) and shows the results in a second RowList with Item / Type / Subtype / Level columns, on its own area frame. The two lists' area frames are shown and hidden in turn. My first cut switched one list withRowList:SetColumns, and it raised on the operator's first search:FontString:SetScript(): Doesn't have a "OnClick" scriptatLibAceGUIWidgets-RowList.lua:1368. That's a library defect, reported to LibAceGUIWidgets (thread 91e50ebd7ba7). The harness lets the call through, so 745 green tests missed it; reported to WoWAPITesting (thread 0633f2ae1eea).where_specnow fails if either list'sSetColumnsis ever called. Clicking a result opens its places. Hovering a result shows the game's own item tooltip. Clearing the box returns to the item. lib:Searchtests the name before unpacking the row. It used to unpack every item's core row first. Measured offline on the real Mists data (88,259 items, plain Lua 5.1): a search that misses took ~270 ms and now ~35 ms, and one that fills the 100-row cap went from ~21 ms to ~3 ms. The search strip pays this on every keystroke. Same results for every query. Also noticed and not changed: when a query hits the cap, which 100 are kept followspairsorder, and only those are then sorted by name.- A boss is not listed twice. The drop graph's Boss row and the places data's Drop row for the same creature collapse to the Boss row. They match on the creature id, which is what the drop graph's encounter key is, and fall back to the name (Peer Review, thread f73d04403a2c).
- Tooltip hint: the "ItemDB" heading on a material's tooltip now reads
ItemDB (Alt+Click: where to get it), the hint in grey, while the Alt+click setting is on./itemdb where [id, link or name]opens the window from chat, empty when given nothing.BuildWhereSearchRowsis the frame-free half. The search runs on each typed letter from the third on, with no delay, because writ refuses a timer; its cost on Mists is measured in thelib:Searchbullet above. - Alt+click: one
hooksecurefunc("HandleModifiedItemClick", ...)for the session, guarded so a library upgrade reloading the file doesn't hook twice. On Era that function acts on the CHATLINK and DRESSUP modifiers only (ItemButtonTemplate.lua:137-159), so the default UI gives Alt+click on an item no job. It's a new tooltip setting,altClick, on by default, with a checkbox on the/itemdbTooltip tab. Whether any other installed addon uses Alt+click is not verified. - Key binding: a new
Bindings.xmldeclaresLIBITEMDB_WHERE, unbound, under AddOns > LibItemDB. It opens the window on whatever itemGameTooltipis showing (WhereHovered). - The window's position and the faction switch live under a
wherekey inLibItemDB_PriceDB, the library's only SavedVariables table. - Tests:
Tests/where_spec.lua, 32 examples (the suite went from 704 to 736), covering the row builder, the point picker, the Questbook hand-off, the window's geometry and data, the Alt+click gate, the key binding and the one-hook guard.Where.lua,Tooltip.lua,Price/Window.luaandLibItemDB-1.0.luaare all at 100% line coverage.verify-manifestis now a declared suite in.writ-suites.json, because the desk'srun_checkwon't pass it its.argument.
New -- GetItemPlaces: where to get an item, with zones, chances and map points (MINOR 36)
The first half of the "Where to get it" window. GetSources says what KIND of source an item has
(Drop, Vendor, Mining). GetItemPlaces says where to go: the creatures that drop it and their
drop chance, the vendors that sell it, and the objects (ore veins, herbs, fishing pools, chests)
that yield it. Each one lists the zones it stands in, with up to three map points per zone.
tools/build-item-places.pyreads both CMaNGOS databases. Drops use the loot table's own rows withChanceOrQuestChance. An equal-chance group row (chance 0 inside a group) gets its share of what the group's explicit rows leave. A negative chance ships as negative and the library reports it asquestOnly. Reference-loot rows are not followed for creatures, because those are the shared world-drop pools every mob of a level range rolls. Objects do follow them, because that is where a node's gems are.- Vendors include
npc_vendor_template. It is keyed bycreature_template.VendorTemplateIdand holds 5,167 TBC rows and 242 Vanilla ones.build-item-sources.pynever read it, which is my defect. Measured: 74 items on Vanilla and 1,721 on TBC are sold only through a template, and 124 of TBC's shipped location rows (22 on Vanilla) lack the Vendor entry they should have. Both builders now read vendors through one shareditemdb_common.cmangos_vendor_rows, pinned initem_places_check.py.ItemLocations.luais rebuilt for Classic and TBC. Reading the templates put "Alterac Valley" on common food such as Dalaran Sharp, because a template vendor stands inside the battleground. So a battleground vendor place is now dropped whenever ordinary vendors also sell the item. That also removed such labels that direct stock had already added: Classic vendor rows went from 1,445 to 1,306. The data-shrink check passes, so no item lost its last source. - Built:
Data/Vanilla/_core/ItemPlaces.luahas 6,476 items and 4,723 creatures and objects (539 KB).Data/TBC/_core/ItemPlaces.luahas 11,157 items and 6,509 creatures and objects (775 KB). Both are listed in their TOCs next toItemLocations.lua. - Two defects in my first build, caught by reading the file rather than trusting the green
suite. First, the separator was written as the Python literal
"\031", which is an octal escape and so byte 25, not the library's byte 31. Every entity would have decoded as one long name. Second, the side came fromEnemyGroupalone, which marked monsters with an emptyEnemyGroup(Kobold Vermin, Timber Wolf) as friendly to both factions;FactionGroupbit 8 (monster) now means no side. Both are now small functions (pack_entity,lua_literal,side_of) pinned initem_places_check.py. - Zones come from the client's own
UiMapAssignment, becausecreature_zoneis empty in both databases. Neighbouring zones' rectangles overlap at their borders. A spawn in two of them goes to the zone that creature's unambiguous spawns use, otherwise to the smaller rectangle (a city inside its zone). Coordinates are that zone's 0-100 map coordinates. Measured: Tharynn Bouden (creature 66) comes out at Elwynn Forest 41.8, 67.2, her known Goldshire spot. A consumer gets world yards fromC_Map.GetWorldPosFromMapPos, so the handover's open question about whether CMaNGOS and Questbook use the same axes no longer arises. The client does the conversion. - Faction side comes from
FactionTemplate.EnemyGroupbits (1 players, 2 Alliance, 4 Horde). Per-factionEnemies_nlists are not read, so it can only err toward listing a vendor a side can't use. CONTINENTSand the instance-name aliases moved frombuild-item-sources.pyintoitemdb_commonasCONTINENTS/INSTANCE_NAME_ALIAS, because both builders must agree on what an instance is and how it's spelled.- Tests: eight new specs in
libitemdb_spec.lua.Tests/item_places_check.pypins the conversion on the real Elwynn rectangle and Tharynn Bouden's real spawn. Red-checked by swapping the axes, which only that case catches. It also pins the border tie-break, the point thinning and the chance format. It's declared in.writ-suites.jsonand CLAUDE.md.
Fixed -- the reagent tooltip never appeared on Classic Era
The operator hovered Runecloth on Classic Era and saw SmexyMats' lines but none of ours. v1.0.0 shipped the "Used by" line to a client that never drew it, and there were two defects.
It hooked the wrong mechanism. EnableReagentTooltip registered a
TooltipDataProcessor.AddTooltipPostCall because the global exists. When the registration
succeeded, it returned without hooking OnTooltipSetItem. Era's UI source does define it
(TooltipDataHandler.lua:199). But post-calls run only from
TooltipDataHandlerMixin:ProcessInfo (:298), which builds a tooltip from a C_TooltipInfo
getter (:251), and Era's C_TooltipInfo has no functions. Era draws bag, link and vendor
tooltips natively, so the registered function was never called. I reasoned from "the function is
defined" to "the behaviour happens", wrote that into the spec as fact, and the spec asserted
exactly one post-call, which pinned the bug as passing. The data-driven path is now taken only
when C_TooltipInfo.GetBagItem exists. Otherwise the per-frame OnTooltipSetItem hook on
GameTooltip / ItemRefTooltip is used, the same hook SmexyMats uses. On the data-driven path,
the post-call now passes the item id from tooltipData.id instead of resolving a link.
The same item never got its line twice. The per-tooltip memory that prevents duplicate lines
was never reset, so hovering the same item again (the tooltip rebuilt from nothing) was refused as
"already added". It is now cleared on OnTooltipCleared, hooked once per frame.
The spec could not have seen either one. libitemdb_spec.lua runs without the widget layer,
so there was no GameTooltip when Tooltip.lua loaded. The classic hook had nothing
to attach to, and nothing noticed. The file now stages a GameTooltip for the load and restores it
afterwards. A file-scope frames.reset() was tried first and reverted, because the widget layer
leaked into pricescan_spec.lua and broke its AH-button example. The load is now pinned for both
client shapes: Era-shaped (classic hook, no post-call) and data-driven (a single Item post-call
carrying the id). Re-showing the same item and hooking OnTooltipCleared once are pinned too.
Harness pin moved to 5a75f26. The harness had the same wrong belief in env/wow.lua's comment.
I reported it (inbox 495fd48f), and it corrected the comment and added two specs proving that
SetHyperlink fires no post-call on the Era shape. The harness's behaviour did not change, so there
was nothing to adopt in code. The whole suite is green against it, with every shipped Lua file at
100%.
New -- the reagent tooltip says where a material comes from, and the data knows about gathering
This finishes the SmexyMats fold-in. v1.0.0 shipped its "used by" half. Its "sourced from" half
was still missing, even though the operator's 2026-09-22 directive asked for full parity. Checking
the gap turned up a data hole underneath it: ItemLocations had no gathering source at all.
Light Leather read only Crafted, Copper Ore only World Drop, and Peacebloom only Vendor.
Mining, Herbalism, Skinning and Fishing were invisible to GetSources.
Data. tools/build-item-sources.py gains gathered_items(), which reads the same CMaNGOS DB it
already reads:
- Gathering nodes are
gameobject_templatechests (type 3). The node's lock (data0) is resolved through the client's wagoLocktable: a slot withType = 2(a skill key) names a LockType in its_Index, and LockType 2 is Herbalism and 3 is Mining on both builds read. Lock ids are never hard-coded. It is a whitelist of those two LockTypes, because chest locks (Open / Treasure) are skill-typed too and would otherwise read as gathering. - Fishing holes (type 25),
fishing_loot_templateandskinning_loot_templateare read as well. - Reference rows (
mincountOrRef < 0) are expanded recursively, with a cycle guard. That is where the gems off a mining node live.
The rows are additive, written after the drop fill, so nothing already placed changes. This was
verified against HEAD by parsing both files: 0 non-gathered rows changed, and 0 new items without a
gathered row. Vanilla goes from 8,594 to 8,742 located items (257 gathered). TBC goes from 13,226
to 13,458 (427 gathered). One known imprecision is stated in the builder: the server also uses
skinning_loot_template for the few creatures gathered with Herbalism, Mining or Engineering (TBC
motes), so those read Skinning.
Display. AddReagentLines adds Source: Mining, Skinning, Drop, Vendor under the Used-by line.
Gathering professions come first, and a boss-graph row reads Drop. A kind with no display word is
skipped rather than printed raw. The line appears only on reagents, as SmexyMats did. There is no
new API and no MINOR bump. GetSources gains a gathered kind, which the README documents for
consumers that switch on source.
The Source line names the profession that MAKES an item, and disenchanting counts. SmexyMats'
"Source(s)" line named the producing profession, and ours said "Crafted".
build-item-sources.py now resolves each create-item spell through the client's
SkillLineAbility to the profession that teaches it, using the same SkillLine.CategoryID 9/11
rule and [DNT] exclusion that build-reagents.py uses. The crafted row's location becomes
Tailoring, Blacksmithing, or Mining for smelting. A spell no profession teaches still reads
Crafted, but only when no named profession also makes the item. Copper Bar read
"Mining, Crafted" until that rule was added. disenchant_loot_template adds Enchanting as a
gathered source for dusts, essences and shards. Vanilla now has 538 of 1,668 crafted rows still
generic, and TBC 573 of 2,534. Measured samples: Bolt of Linen Tailoring, Copper Bar Mining,
Strange Dust Enchanting, Light Leather Leatherworking + Skinning, Runecloth World Drop
(which matches SmexyMats' single Drop icon for it). Consumer-visible: a crafted row's
instance changes from "Crafted" to a profession name. Dibs matches crafted gear by kind
(IsCrafted), not by that string, so it keeps working and has been told.
Profession icons, on by default, as SmexyMats drew them (MINOR 35). Both lines draw professions
and source kinds as game icons at size 20, using SmexyMats' own defaults (IconsEnabled = true,
TooltipIconSize = 20). Icons are referenced by file data id, never by path. Every id was read
from the client's ManifestInterfaceData on Vanilla 1.15.9 and TBC 2.5.6 under Interface\ICONS\,
and the professions use the textures SmexyMats named. A word with no verified icon (Inscription,
Archaeology, the bare "Crafted") stays a word. Quest, PvP and Reputation have no SmexyMats icon, so
those three are our choice of existing icons. achievement_pvp_a_01 was rejected because TBC's
manifest lacks it. New GetTooltipOption(key) / SetTooltipOption(key, value) hold enabled,
icons and iconSize (clamped 8..64). Is/SetReagentTooltipEnabled now delegate to enabled.
Caught by the new spec before it shipped: the first GetTooltipOption read the saved settings
through an a and b chain. With no saved settings, that chain yields false, not nil, so every
setting would have read as off on a fresh install, including enabled.
An "Expansion:" line leads the tooltip, as SmexyMats' did (MINOR 35). New GetExpansion(itemID)
and Data/<Version>/_core/Expansion.lua, built by the new tools/build-expansions.py. The client's
own ItemSparse.ExpansionID was measured and rejected: it reads 254 ("unset") for 17,602 of 17,682
Vanilla items and 29,931 of 30,084 TBC items. An item's expansion is instead the first game whose
data ships it. Everything in Data/Vanilla/_core is Classic (17,682). A TBC item that Vanilla does
not ship is The Burning Crusade (12,516 of TBC's 30,084). Mists is refused by the builder rather
than guessed, because with no Wrath or Cata item set, everything past TBC is indistinguishable
there. The line uses SmexyMats' colours (Classic E6CC80, TBC 1EFF00) and can be turned off
through the expansion setting. The tooltip now follows SmexyMats' order: Expansion, Source,
Used by. Moving Source above Used by is a visible change. The file uses packed strings rather than
[id]= rows, and core_ids() lists it as a non-item file, so it can never be counted as items.
An "Item ID:" line on every item, last, as SmexyMats' ItemIDs option drew it (default on,
like SmexyMats). It is the one line not gated on the item being a material, so AddReagentLines
now always records the tooltip in its de-duplication memory. It returns true when either the
material lines or the id line was added. Known overlap, not a defect: AllTheThings also prints
an Item ID line, so a player running both sees two until they switch ours off with itemID.
A Tooltip tab in /itemdb, with SmexyMats' /sm switches. The window gains a third tab. It
has one checkbox per line (all lines, Expansion, Source, Used by, Item ID), one for icons versus
words, and an icon-size slider (8-64) that is disabled while icons are off. Each control writes
through SetTooltipOption. The tab is rebuilt on every visit, so it always shows what is stored. A
price event arriving while the tab is open no longer repaints the hidden sources list, and the
deferred first-open refresh now fires only on the Sources tab (it used to fire on anything that was
not Scan). New source and usedBy settings let each of those lines be switched off on its own.
Tests/env_price.lua now loads Tooltip.lua in TOC order, so the window spec passes when run
alone. Not carried over from SmexyMats, on purpose: the [SM] label prefix (its own branding),
its load message and chat error reporting (both about SmexyMats' own internals) and its Reset
button. Colour-blind mode is carried over. New colorblind setting (default off) and color1 /
color2 settings (six hex digits, stored lower-case; anything else is refused). When the mode is
on, every label takes colour #1 and every value colour #2, overriding the expansion's own colour as
SmexyMats did. The Tooltip tab has the switch and two colour pickers, and the pickers stay disabled
until the mode is on. A picker's 0..1 channels are rounded, not truncated, to hex, so a colour is
stored as exactly the colour it will be read back as. All four tooltip lines now go through one
helper that picks their colours, so a line added later cannot miss the mode.
Our lines open with an "ItemDB" heading, because the operator could not find them. On Deeprock
Salt they were drawn correctly but sat directly under TSM's block with nothing marking them as
ours: "i don't see the item DB lines in here". The suite's other addons open their block with a
blank line and their name (TOGPM, TOGBankClassic), and SmexyMats prefixed every line with [SM]
(its SMText option, on by default). The heading is drawn once, only when at least one line
follows, only on a frame that has AddLine, and it can be switched off through the new header
setting or the Tooltip tab.
Found in this session's self-audit and fixed. PROFESSION_CATEGORIES was defined in both
build-reagents.py and build-item-sources.py, and it now lives once in itemdb_common.py. If the
two copies had drifted, the tooltip's "Used by" and "Source" lines would disagree about what counts
as a profession. The colour pickers saved only on OnValueConfirmed, which AceGUI fires only from
the picker's opacity callback (AceGUIWidget-ColorPicker.lua:39-40). These pickers have no opacity,
so a picked colour could be shown and never kept. They now save on OnValueChanged too. New
GetTooltipSwitches() lets the window spec require a checkbox for every on/off setting, so an option
added to the defaults cannot silently never appear in the window.
The labels are localised, as SmexyMats' were. Expansion, Source, Used by and Item ID use
SmexyMats' own translations (its Locales/*.lua: deDE, frFR, esES, esMX, ruRU, koKR), and every
other language falls back to English. The expansion NAME is the client's own EXPANSION_NAME<n>
string when it has one, read at call time, with English as the fallback. Not verified: that the
Classic Era client defines EXPANSION_NAME0/1. They were read from the GlobalStrings of the
reference tree, which is not the Era tree, and the English fallback covers their absence. The
profession and source WORDS (words mode, and professions with no icon) are translated too, again
from SmexyMats' tables. Words SmexyMats never translated (First Aid in most languages, Quest, PvP,
Reputation, Crafted) stay English. A spec reloads Tooltip.lua under GetLocale() = "deDE" and
pins the German lines. The /itemdb window's own text stays English, as SmexyMats' options page
did (its Data/Options/Options.lua names are hard-coded English).
SmexyMats' layout, because it is easier to read. The operator compared the two tooltips side by
side: "it's easier to people read". Every line is now left-aligned. With icons on, Source and
Used by each put the icon row on a line of its own under the label (SmexyCore.lua:233, :246).
Expansion and Item ID, and every line when icons are off, keep the value on the label's line in
its own colour. A frame with no AddLine falls back to the two-column AddDoubleLine form.
A corpse gathered with Herbalism or Mining no longer reads "Skinning". The server keeps that
loot in skinning_loot_template too, and creature_template.CreatureTypeFlags says which
profession gathers it. This was verified from the data: on TBC, the 33 creatures flagged 0x100
(Bog Giant, Underbog Lord) carry Felweed and the other herbs, and the 36 flagged 0x200
(Shattered Rumbler, Tavarok) carry ore and gems. This replaces the
imprecision stated above. TBC's gathered rows go from 457 to 432, as items that read both Mining and
a wrong Skinning now read Mining only. Classic's database flags no creature this way, so Vanilla is
unchanged. Deeprock Salt (8150) correctly reads Drop only on Classic, where SmexyMats shows a
Mining pick. In the Classic CMaNGOS database it comes from creature drops (Rock Elemental, Stone
Fury, Stone Golem, and others, none of them gather-flagged) and from Knot Thimblejack's Cache
(gameobject 179501, a keyed Dire Maul cache). The operator confirmed it in game on 2026-09-24:
"you can't mine them in classic, so you're data is right then for classic". SmexyMats' Mining
comes from its Wowhead-parsed tables (SmexyCore.lua:2), which carry later expansions'
mineable-corpse sources. This is a case where parity would have meant copying a wrong answer.
Scope: gathering ships for Vanilla and TBC from their CMaNGOS DBs, and for Mists from
SkyFire's (see the Mists section at the top of this release). Forever's sources come from
build-sources-wowsims.py (crafted plus chest/object/trash drops), and WoWSims has no gathering
data, so gathering is not shown there.
Older releases (v1.0.0 and earlier) are in CHANGELOG_ARCHIVE.md. Nothing was edited or dropped, only moved.
This mod has no additional files