fcdf14c0dc8cab514b953046cf8a47b994b34d53
648
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  gromandClaude Opus 5
|
a9bb9624e9 |
app+ui: nampower's press-interception queue — the half that changes how keys feel
The half of upstream's queue that `queue.rs` has said for three rounds it deliberately left out. A press the client would refuse for a **timing** reason is now held in the same slot an explicit `QueueSpellByName` uses, and the drain sends it the moment the guard clears: no packet, no global cooldown, no error line while it waits. One queue, because upstream has one — a second set of slots beside the first would be a different feature wearing the same name. **The guard reopens early, and that is the headline.** `PendingCast` kept one timestamp and now keeps two. The guard reopens at the cast's own already-known length plus `NP_MinBufferTimeMs`; the record still lives for the cast plus two seconds, so a resolution arriving a round trip later is still recognised as the answer to that cast. That split is not tidiness — the `modalNextSpell` chain that starts a hunter's Auto Shot reads the record, and a record that died with the guard would break shooting. The buffer is not a latency figure, which is why upstream's default is 55 ms and not a ping: our packet reaches the server half a trip after the send, the server finishes its cast half a trip after our own cast time elapses, so a next cast sent at exactly `cast_time` arrives just as the server frees up and the latency cancels out of the arithmetic. The buffer is jitter margin on top. `NP_NonGcdBufferTimeMs` is wider because a non-GCD spell has nothing *but* the guard between two presses. **`NP_RetryServerRejectedSpells` is the mitigation for that**, and that is why upstream defaults it on. If the margin is ever wrong the server refuses the next cast and the press is simply gone — the dropped cast the early guard buys its speed with. It goes back once, held a buffer's worth, on exactly two codes: 0x3c not-ready and 0x61 another-action. Every other refusal is the server telling the player something true, and re-sending one would be a packet a second that never succeeds. **What a window measures, and how that was decided.** Upstream publishes four and no sentence saying. Two readings fit: how long an entry survives, or how close to the guard's end a press must land. This takes the first, because `queue.rs` already reads `NP_SpellQueueWindowMs` that way and one number meaning two things in one client is a trap nobody finds twice; because `Replacing queued nonGcd spell X with Y` means the last press wins, so holding every press costs nothing in staleness; and because the other reading would make the 250 ms cooldown window refuse almost every press made during a global cooldown, which is the complaint nampower exists to answer. Two rows do read a window as remaining time, and both say so themselves. `NP_InterruptChannelsOutsideQueueWindow`'s name states one. And `NP_CooldownQueueWindowMs` has to be one: a cast ends in a second, a Hearthstone cooldown is half an hour, and holding that press means either firing it half an hour later or swallowing it with no error line. A quarter of a second is an allowance for pressing a touch early, not a queue. **Where the rung sits.** Five rungs stay above it — the profession opener, the auto-repeat re-press, the dead caster, the targeting cursor's clear and the on-swing re-press — because each is what the press *means* whatever is in flight. Everything below is a real refusal a player should read: no reagent, out of range, no power, stunned, moving, wrong form. Upstream hooks the cast entry and holds those too, so its error arrives a second later about something the player stopped doing. Refusing them now is the deviation and the better half of the trade. Fourteen CVar rows went live. What stays stored is named with its cost: the dynamic buffer optimizer and its ceiling (off in upstream's own defaults), `NP_OnSwingBufferCooldownMs` (nothing to rate-limit — a re-press of the queued strike already takes the reference's silent bail), the channel latency trim, and the three quickcast rows, all off by default and all needing the world point under the mouse in the cast path. `CastSpellNoQueue`/`CastSpellByNameNoQueue` became meaningful: they take `CastLadder::send_now`, which skips the rung. The queue's own pop takes it too — a pop that could be re-held is a loop that never casts anything. The test worlds now seed the **registered** defaults rather than `Default`'s all-zeroes, which is a configuration nobody runs: three ladder tests were passing only because the queue was off in them. Each now asserts the queue behaviour and keeps the reference's 0x61 in a world where a player unticked the queue. And one test drives the whole cycle through the real drain — press held, nothing on the wire, guard cleared, one packet out — because the failure this guards against is a keypress that produced no packet, no error and no cast, which is the one a player cannot describe. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
68f397af72 |
app+protocol: nampower's combat-log family — 35 events off one decode
Every packet this family reports was already decoded: `combat_log/` builds its chat lines from exactly these structs. `ui_np_events.rs` is a second consumer of that decode, not a second decoder, which is also why its arguments cannot drift from the chat log's. Built: `SPELL_DAMAGE_EVENT_*`, `AUTO_ATTACK_*`, the three `SPELL_HEAL_*`, the three `SPELL_ENERGIZE_*`, `SPELL_MISS_*`, `DAMAGE_SHIELD_*`, `ENVIRONMENTAL_DMG_*`, `SPELL_DISPEL_BY_*`, `UNIT_DIED`, the `*_UPDATE_DURATION_SELF` pair, the thirteen `UNIT_*_GUID`/ `PLAYER_GUILD_UPDATE_GUID` rows and `UNIT_COMBAT_GUID`. **A registration is the switch.** Upstream's README calls every `NP_Enable*Events` CVar deprecated — "registering the event is preferred" — and three of the eleven families have no CVar at all, so a CVar-only reading would leave `SPELL_DAMAGE_EVENT_*` unreachable for every addon and `SPELL_ENERGIZE_*` unreachable for all of them, since nothing installed writes its CVar. `refresh_listeners` reads the VM's own registration table (the same one `GetFramesRegisteredForEvent` reads), so a family is wanted when its CVar is set OR something is listening. That is also what keeps the cost honest: with nobody registered a damage packet queues nothing, rather than building eight arguments for a fire that reaches no frame. Nine `NP_*` CVars move from stored-only to LIVE, and `NP_EnableSpellEnergizeEvents` is registered for the first time — the first pass took the names off `NampowerSettings`' writes rather than off the DLL's own CVar table. The two reasons written at the auto-attack and heal rows were both about the *cast* path, and both were answered by the combat-log decode that was already there. The descriptor indices the GUID-keyed family watches are written out and then cross-checked against the two tables the tree already keeps — `unit_rec::FIELDS` and the `FIELD_UNIT_*` constants. A hand-written descriptor index is exactly the number that is plausibly wrong and silently so: the event simply never fires, which no amount of play distinguishes from "the server never sent one". `AttackerState` gains `sub_damage_count`, which is `AUTO_ATTACK_*`'s arg6; vmangos writes one block, but a server splitting a swing across schools sends more. Two test notes worth keeping: the edge tests run through the schedule, because `run_system_once` builds a fresh system each call and the `MessageReader` cursor restarts — a test written that way passes whatever the system does with the second edge. And the data-gated suite held at exactly 269, after the event gate correctly named all 35 new names. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
702eb475a4 |
app+ui: the queue, the disenchant sequencer — 62/62
QueueSpellByName, QueueScript, ChannelStopCastingNextTick, DisenchantAll. Every one of nampower's 62 registered Lua functions is now defined; the `comm -23` in the doc prints nothing. **Upstream's queue has two halves and one is deliberately not built.** The Lua surface — remember this spell, send it when the client will accept it — is `spell/queue.rs`. The other half intercepts *every ordinary press*, so a keypress during a cast is held rather than refused, and it is governed by NP_QueueCastTimeSpells, NP_QueueInstantSpells, the buffer rows and the quickcast pair. That half is not a binding: it changes how every key in the game feels, its bugs are double-casts and dropped casts, and it is not something to switch on as a side effect of porting three function names. Those twenty CVars stay stored-only, each row saying so; the machinery they would sit on is now here, and turning it on is its own piece of work with its own live test. Two of the twenty-two went live, because they are the explicit queue's own knobs: NP_SpellQueueWindowMs and NP_QueueSpellsOnCooldown. Three slots, because upstream has three and its own event says so: SPELL_QUEUE_EVENT's code is 0/1 on-swing, 2/3 normal, 4/5 non-GCD, so the queue is keyed by what kind of cast a spell is — and the kind comes off its own Spell.dbc row through the same three tests the send path makes. Each pops on its own condition: non-GCD and on-swing as soon as no cast guards (neither uses the global cooldown), normal when the GCD has run out too. Three decisions at the queue, each pinned: - A replacement raises no popped event. A second queue of one kind replaces the first — upstream's own log line is `Replacing queued nonGcd spell X with Y` — and the replaced entry was never sent, so a consumer counting pops against queues would see one too many. - An expiry raises nothing either. `*_POPPED` means "this was sent". - The expiry is measured from the end of the cast in flight, not from the moment of queueing, so queueing early during a long cast is not punished. A lapsed deadline cannot shorten the window. **`DisenchantAll` is specified completely by the DLL's own strings**, which is why it is built this way rather than designed. `0x93087` onward carries the argument note, the precondition, the three quality keywords, every stop reason and the three chat lines — one of them as a Lua chunk that builds the item link itself. "move during cast to cancel" is the abort, and it needed nothing new: moving already cancels a cast here, the cast then fails, and a failed cast stops the run. The one piece of machinery a sequencer usually needs is the client's own, and the player's way out is the one upstream documents. How a step's outcome is read: the item is gone, or it is not. There is no "disenchant succeeded" packet — the item leaves the bag. Each step remembers the guid it cast at and waits for it to vanish with a ten-second deadline. An observable test with no new wire plumbing, right for the same reason the cast bar's own edges are. The queue's drain and the sequencer both live inside `drain_spell_casts`, which that file's own note explains: only one system may hold `CastLadder`, because a second would share its resources with every other cast-send system in an order the ambiguity gate cannot know is safe. `spell/queue.rs` names no VM — the first draft mentioned `UiScript` in one comment and `ui_script_wall.rs` counted it as gameplay code naming the VM, which is the gate working as intended: that module really is gameplay state and not a feed. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
b652c90a4b |
app+ui+protocol: the four measured-out names, and the unit descriptor by name
Each of these had a reason for being out. Each reason turned out to be
answerable, and two of them were simply wrong.
`GetSpellPower(school)` is **delegated to `GetSpellBonusDamage`**, which
this client already answered — one number under two mods' names, as
SuperCleveRoidMacros says itself ("Reads the same client field
nampower's GetSpellPower does"). Delegating is the only way to guarantee
they cannot drift, and the test asserts equality across all seven
schools and the off-range values rather than a literal. The earlier
"needs a descriptor band fed" was wrong: it was already fed.
`GetRaidTargets()` is the 8-slot **board**, now pushed beside the party
roster. The earlier reason was right about the scan it rejected and
wrong about the alternative: `GroupState::raid_targets` is the
reference's own table and copying it costs eight words a push. A mark
with nothing on it is absent, because a zero GUID is a string that
passes `if t[4]`.
`LearnTalentRank` is one `LearnTalent` press per missing rank, which is
what a player clicking the button does. The earlier reason — the burst
cannot be checked before it is sent — is true and is not a reason to
withhold it: the check that matters is the server's and the client has
never had another. What can be checked locally is, so a rank already
held queues nothing and one past the maximum is clamped.
`GetUnitData` is the whole descriptor, from a new raw block on
`UnitState`. The earlier reason — "a decision about which fields" — is
now a table of seventy columns under the DLL's own names, read off its
string pool in UnitFields order; `GetUnitField`'s scalar half reads the
same block, so the two cannot answer differently for one name, and a
test walks the table asserting exactly that.
**The indices are derived and the derivation is checked.** Only thirteen
`FIELD_UNIT_*` constants existed here; the rest come from 1.12's block
sizes. Those thirteen span index 22 to 147 and every one lands where the
layout puts it — thirteen independent confirmations, run as a test
rather than asserted in a comment, with two more that no columns overlap
and that nothing falls inside the aura blocks' 47..=124.
Three decisions at that mapping:
- An unstreamed column is absent, not zero. `ObjectFields::at` is
`get_raw`'s strictness rather than `get_u32`'s absent-is-zero
convenience: `health = 0` for a unit whose health has not arrived is a
number a caller divides by.
- `aura`/`auraApplications` come from the filtered per-unit aura list,
not the raw slots — the same place `GetUnitField` has always answered
them from. An unfiltered second answer would be a different unit by
the same name.
- `status` and `zoneId` are left out of nampower's seven computed
columns: nothing says what `status` is a status *of*, and a zone is
only knowable for the player. The other five are filled by the caller,
which knows them.
`SpellRecValue` is now `RecValue`: it serves both record readers, and a
second copy of the same four shapes would drift.
58 of 62 registered names. The four left are `DisenchantAll` and the
queue trio.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
e716709bca |
app+ui: the nine file and credential verbs, fenced and keyed
Built on grom's explicit say-so, after a round where they were withheld and the reason recorded. Both halves of that decision stay in the doc, because the reasoning is what makes the second half a decision rather than a drift. ReadCustomFile/ImportFile (one operation, two upstream names), WriteCustomFile/ExportFile, CustomFileExists, ExecuteCustomLuaFile, StartItemExport, EncryptPassword, EncryptedServerLogin. **The fence is app-side and it is not a `..` scan.** `local_state::custom_file_path` admits one relative name whose every component is an ordinary name, inside `benilla-config/CustomFiles/`, and refuses a leaf that is already a symlink. Each escape a string check misses has its own rule and its own test row: an absolute path (`join` *replaces* the base rather than appending), a Windows prefix, a `..` nested inside a longer path, and a planted symlink — the one no amount of sanitising the string will catch. A subdirectory rather than the config root, because a verb that could create `config.toml` or `autologin.toml` by name would let an addon rewrite the client's own settings; that is a different power from "keep a file of your own". The VM side holds three closures and no paths, so no spelling of a `filename` reaches outside that directory. **The credential pair's deviation is forced, not chosen.** Both are registered by nampower under "Registering *Glue* Lua functions" and this client has no glue Lua VM at all — login, character-select and character-create are Rust. `EncryptPassword` is pure and works the same anywhere. `EncryptedServerLogin` cannot mean "submit the login form" in the only VM there is, so it stores the credential for the next login: `enabled` on, the plaintext field cleared, the blob kept, and `AutoLogin::effective_password` reads it back at the refill. What that protects is an `autologin.toml` or a SavedVariables file read by someone else or pasted into a forum: the blob is authenticated and keyed on 32 random bytes in a different file, mode 0600. What it does not protect is anything running as this user, which can read the key as easily as the blob — there is no user secret in the chain and there cannot be, because the point is logging in without typing one. Both halves are at the module. Encrypt-then-MAC with HMAC-SHA1 as the PRF for both halves, from the two primitives the tree already had (sha1 moves from a dev-dependency to a real one). The HMAC is tested against RFC 2202's own vectors, because a hand-written one still round-trips with itself when it is wrong — and a first draft of the key reader did exactly that kind of silent wrong: it returned raw bytes on creation and hex text on read, and the caller guessed between them with "32 bytes and not all hex digits". It guessed wrong the moment a random key came out all-hex. Returning one thing is the fix, and the note is at the function. **The existing plaintext path is untouched.** Storing the password in plaintext was grom's own explicit choice when the alternatives were a keyring or nothing, so this is additive and nothing migrates by itself. The machinery to make the Remember Password tick-box write a blob instead is now all here — one line — and that is his call. Base64 was missing from the whole tree; it is in `benilla-ui`'s `script/encoding.rs` with RFC 4648's own vectors, and `C_EncodingUtil.EncodeBase64` will want the same one. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
f87ef148b8 |
app+ui: the trinket cooldown — and a stock verb that always said "cold"
Built: GetTrinketCooldown, UseTrinket, GetSpellRangeData, SetLocalRaidTargetIndex, GetQuestLogQuestIds, GetQuestDialogQuestId, CombatLogFlush. The find on the way through is not a nampower verb. **`GetInventoryItemCooldown` is a stock 1.12 verb and it answered `(0, 0, 1)` unconditionally** — right shape, always "cold", so nothing errored and no worn trinket ever drew its cooldown swirl on the character pane. Stock PaperDollFrame.lua:691 calls it on every slot update. Feeding an equipped slot's cooldown for GetTrinketCooldown fixes that one off the same column, and the test runs `CooldownFrame_SetTimer`'s own arithmetic against it. The feed keys its query on the (use spell, item entry) pair exactly as the bag feed does, so one trinket reads the same cooldown from the doll and from a bag. **The risk was the gate, and my first test for it was vacuous.** `feed_char` sits behind a change gate and no descriptor moves when a cooldown is armed, so `Cooldowns::feed_epoch` had to become a gate input. The first test passed with that input deleted: `run_system_once` builds a fresh system each call, so the gate's `Local` memo resets and `vm_reset` opens the gate every time. A test for a gate input has to run the system the way the schedule does. Rewritten with `add_systems` + `update`, it fails with the input removed — verified by probe, and the reason is written at the test. `feed_char` also hit Bevy's sixteen-parameter ceiling, which was a fair signal its parameters wanted grouping by what they are for: `ItemCooldownInputs` and `RepairInputs` replace six. Two decisions recorded at the code: - `GetSpellRangeData(rangeIndex)` answers three values, not a table. Every table-returning verb in this DLL has its key names in the string pool and this row's are not there. `flags` bit 0 is the melee family, where the engine takes the combat-reach sum instead of the authored pair — a caller ignoring the bit reads 0..5 for every melee ability. - `CombatLogFlush()` is a deliberate no-op. Upstream buffers its combat-log writes and this empties the buffer, so what an addon asks for is "my lines are on disk now" — and `ui_chat/logging.rs` writes each line to an unbuffered File as the window shows it, so that holds at all times. Defining it is the honest answer: the capability is present, and a missing global makes `HasFeature` report it absent. Also fed, both one wire column each: `QuestState::quest_id` (the wire names the quest on all three giver panels and 1.12's Lua names it nowhere — a title is not a key) and `SpellDisplay::stack_amount`'s neighbour work from the previous commit. Five names stay out, each now with the field list that makes it a decision rather than a shrug — `GetUnitData` (80 raw descriptor columns, read off the pool, zero call sites, and a partial answer is worse than none), `GetRaidTargets` (the reference keeps a board; this client feeds the mark per unit, and a scan over named tokens would silently omit a marked unit), `DisenchantAll` (a server-paced sequencer, which is the queue), `GetSpellPower`, `StartItemExport` (writes a file — it belongs with the withheld seven). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
f06caa9968 |
ui+app+formats: the Spell.dbc row, and seven columns that answered nil
`GetSpellRecField` knew three names. Counting what the corpus actually asks it for: `name` (4 sites), `school` (3), `stackAmount` (2), `spellIconID` (2), then one each of `rank`, `rangeIndex`, `mechanic`, `manaCost`, `effectMechanic`, `effectApplyAuraName`. Seven of those ten answered nil, and each was a feature quietly off rather than an error — DoiteAuras caps its stack display on `stackAmount` (DoitePlayerAuras.lua:539) and SuperCleveRoidMacros classifies a crowd-control effect off the two arrays (Utility.lua:7776-7794). The eighty field names are the DLL's own, read off its pool in Spell.dbc column order. The mapping to `SpellDisplay` lives in `ui_script/spell_rec.rs`, beside the struct it reads — a copy on the VM side would be free to drift — and `GetSpellRec(spellId)` is that same field list walked through that same per-field reader, so the two verbs cannot answer differently for one name. A test asserts exactly that, key by key, over one row. The array columns are a 1-based table of three, which is measured, not guessed: SuperCleveRoidMacros walks `effectMechanic` as `for i = 1, 3 do local em = effectMechanics[i]`. `totem`, `reagent` and `reagentCount` take the same shape, two and eight long. One new DBC column — `StackAmount`, 39, bracketed by two constants this tree already had (38 ModalNextSpell, 40 Totem[0]). It is the only one added: fifteen further names upstream answers are left nil rather than putting fifteen unused u32s on a struct the spell catalog holds one of per row. They are listed at `FIELDS` with that reason, so the gap is a decision rather than an omission. `GetSpellRecField`'s three string columns keep the spellbook fallback the verb had before the probe, so a VM with a book and no catalog — a capture, or the boot VM before the catalog loads — still answers a learnt spell's name and rank. Every other column is a DBC row and has nowhere else to come from, which is why the fallback is exactly three names wide and says so. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
cf0e157f34 |
ui+app: the bag walk, the modifier fold, and two shapes the DLL's pool corrected
Built: `GetBagItems`, `GetSpellModifiers`, `GetSpellIdCooldown` — plus
fixes to two verbs that already existed and had the wrong *shape*,
found the same way `GetItemIconTexture`'s wrong argument was: by
comparing what we answer against nampower's own string pool.
`GetBagItems()` hands back every occupied bag slot as
`{ [bagIndex] = { [slot] = row } }`. DoiteAuras asks seven times over
and walks both levels with `pairs`, reading `itemId` and `stackCount` —
it is how that addon answers "do I have this item" without 100
`GetContainerItemInfo` calls per condition evaluation. Empty slots and
empty bags are absent, because a nil value in a Lua table *is* absence
and `pairs` is how both levels are read.
`GetSpellModifiers(spellId, modifierType)` answers
`flat, percent, hasModifier`, where percent is the final multiplier (90
means 90% of the original) — the consumer says so in its own words and
multiplies by `percent / 100`. The 26 modifier types come from the
DLL's usage string. The fold is `spell/mods.rs`' existing reader,
reached through a new `spell_modifier_probe` the lifecycle reinstalls
whenever the wire moves the tables: a VM-side reimplementation of a
four-gate 64-bit fold would be free to drift from the mana cost the
action bar greys a button on. The third return is the point — the
consumer tests it before touching either number, and the duration path
multiplies, so `0, 0` for an unmodified spell would zero every debuff
timer that asks.
`GetSpellIdCooldown(spellId)` answers the twenty-key table, from a new
`Cooldowns::legs`: `info`'s three legs kept apart, same walk and same
arms, so the folded number a swirl draws and the legs an addon reads
cannot describe different timers.
**`GetBagItem`'s row had the wrong key names** — `bag`, `slot`,
`count`, `quality`, `link`, `locked`, this client's own
`GetContainerItemInfo` spellings. nampower's row is contiguous at
`0xa8a68` and is the block `GetEquippedItem` already used, so a caller
reading `.stackCount` got nil and no error. One `bag_slot_table` now
serves both, with three columns absent rather than zeroed: this client
resolves a bag slot's enchants to names at the app boundary and keeps
no random-properties id for a container slot, and a zeroed enchant id
would be a claim.
**`GetItemIdCooldown` returned the wrong kind of value** — the stock
`start, duration, enable` triple. At the version we report it is a
table, and the addon layer settles it: `API.GetItemCooldownInfo` and
`API.GetSpellCooldownInfo` return these calls' results *directly* and
their callers index them. `cd.gcdCategoryRemainingMs` is the primary
path of `API.GetGCDRemainingMs`, which the `[gcd]`/`[nogcd]` macro
conditionals divide by 1000 — against a triple that is
`(0.0).gcdCategoryRemainingMs`, which raises. Latent rather than live
(nothing calls those wrappers today), still wrong.
Two things the schedule gates caught and improved:
- The probe refresher started as an exclusive system, which is
ambiguous against everything in Update it is not ordered against:
+274 pairs over the ceiling. `NonSendMut` is all it needs.
- Its memo started as a bare `Local`, which `session.rs`' own gate
refuses — a memo about what this VM has been told must die with the
VM. As a `VmMemo` it now reinstalls on the first frame after a
`/reload`, which is better than the reasoning I had written for the
`Local`: a fresh VM has no closure at all until something installs
one.
One count the earlier sweep got wrong, corrected in the doc:
`GetSpellPower`'s "7 call sites" are BetterCharacterStats' own method
`BCS:GetSpellPower(school)`, not nampower's global. Nothing in the
corpus calls the global at all. A bare name match is not a call site.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
ecee994ab6 |
app+ui: the player's own cast state, and SPELL_CAST_EVENT off the commit
`GetCurrentCastingInfo` and `GetCastInfo` are the two most-called
nampower functions in grom's corpus — 25 and 20 call sites outside
`NampowerAPI.lua` — and one of them calls with no existence guard at
all: `local _,_,_,_,_,_,autoattack = GetCurrentCastingInfo()`
(SuperCleveRoidMacros/Core.lua:94), in the frame that keeps clearing the
target while an autoattack is queued. On this client that line raised.
The earlier doc row said the client had "a spell id, a deadline and a
guard bit — no start instant, no channel state, no GCD end". That was
true of `PendingCast` and wrong about the tree: `ui_cast_events.rs`
already folds a cast window per caster off the same packets, with a
start, an end, a channel bit and the accumulated pushback, behind
`C_Spell.UnitCastingInfo`. What was actually missing was the caster's
own machinery — the reference's cast globals, which have no descriptor
field anywhere — and the GCD, which `spell/cooldowns.rs` has run
correctly all along and never published.
So `GetCastInfo` folds two halves: its timing columns come from the
player's own window, its identity and GCD columns from a new per-tick
`SelfCast` snapshot. The feed declares itself after the window drain, so
the two halves can never describe different casts. Added:
- `SelfCast` (benilla-ui) — the inflight spell `0xceca88`, the bar's
spell `0xceac48`, the auto-repeat key `0xceac30`, casting/channeling/
onswing/autoattack, the cast counter, the cast type, the target, and
the GCD's start and end on the GetTime clock.
- `ui_self_cast.rs` (benilla-app) — the push, plus `SPELL_CAST_EVENT`.
- `Cooldowns::gcd` — the running global cooldown as one timer. `info`
still answers the per-category question every cooldown swirl asks;
this answers "when does the GCD end", which has no spell to key on
because its consumers divide it (`API.GetGCDRemainingMs`) or compare
it against seconds (`[gcd:<1]`).
- `PendingCast`'s cast counter, in the one place every class of commit
already passes through.
Three decisions where the tempting answer was the wrong one:
- `GetCastInfo` is **nil** when nothing is casting, and an instant's GCD
does not count. Answering a table whenever the GCD is up would make
the consumer's fast path always hit — and `API.IsCasting()` *is*
`GetCastInfo() ~= nil`, so `[casting]` would read true for a second
and a half after every instant. The wrapper's own comment says what to
do instead ("GetSpellIdCooldown … works for instant casts").
- `castId` is two different numbers under one upstream name: a counter
in `GetCastInfo`'s table, beside a separate `spellId`; a **spell id**
in `GetCurrentCastingInfo`'s returns, which is how both of its readers
treat it (Init.lua:89 stores it as castingSpellId, Core.lua:5427 hands
it to C_Spell.GetSpellName). Each is implemented as what its own
consumer reads, and both are spelled out at the binding.
- `SPELL_CAST_EVENT` fires for commits, not refusals. Upstream's
`success` can be 0; this client's local refusals funnel through
`CastErrors`, which carries a spell id and a reason byte and neither a
cast type nor a target — a row from there would invent two of five
fields. Nothing in the corpus reads the failing branch
(Core.lua:5177 is one `if success == 1 then` with no else). Written
down at the type rather than left to be found.
Tests run the consumers' own bodies transcribed, as `cast_info.rs` next
door does: `API.GetCastInfo` with its `GetCurrentCastingInfo` fallback
arm, `API.IsCasting`, `API.GetGCDRemainingMs`, `API.GetCastRemainingMs`
and `CleveRoids.UpdateCastingState`. A positional seven-return shape and
a ten-key table are only correct relative to the code that reads them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
aab2130c62 |
ui+docs: nampower's own registration table, and the icon id it takes
`scripts/dll-surface.py` answers "what might this DLL publish". For nampower.dll there is a better question with an exact answer: the names it registers sit contiguously in the string pool, in registration order, bracketed by the DLL's own log lines — `Registering Custom Lua functions` immediately before the first, `Registering Glue Lua functions` before the last two. That is 60 functions plus 2 glue ones, with no inference at all, where the pointer filter gives a candidate list a reader still has to judge. The event table (59 names) and the CVar names sit the same way, just before it. A compiler emits one string literal per `lua_register(L, "<name>", fn)` and pools them in source order, so one registration function's names land in one run, and the log line that opens it is the literal printed immediately before the first call. Look for the log line, then read forward until the strings stop looking like identifiers. Two things fell out of the same pool that corrected code already written: - The 44 `Usage:` strings are the argument contract, printed verbatim when a call's arguments are wrong. `GetSpellRangeData(rangeIndex)` takes a SpellRange.dbc row index, not a spell id, which this doc had wrong; `GetSpellModifiers(spellId, modifierType)` enumerates all 26 modifier constants. - `GetItemIconTexture(displayInfoId)` takes an ItemDisplayInfo row id, the exact sibling of `GetSpellIconTexture(spellIconId)`. Ours took an *item* id and resolved the display id itself, which is a different function: upstream's addon layer keeps that hop in a wrapper of its own (`NampowerAPI.lua:2648` is the pass-through, `:2665` is `GetItemIconTextureByItemId`). So a caller handing us a real display id got "Item not found in DBCache", and one handing an item id got an icon the real DLL would not have answered. Fixed, and the test now pins that an item id answers nothing — it populates the item cache on purpose so the wrong reading would be visible if it came back. With the table read, the surface is no longer a count of what a heuristic flagged: 62 names, each in exactly one bucket (32 built, 8 withheld pending a decision that is not mine, 4 not ours to add, 18 to build), with the `comm -23` that reproduces it. The two at the head of the last bucket are the most-called nampower functions in the whole corpus, and they are coupled: SuperCleveRoidMacros registers SPELL_CAST_EVENT gated on `GetCurrentCastingInfo` existing, so defining the function without firing the event would move it from never registered to registered and silent. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
7861ab2710 |
ui: ClassicAPI's GetCraftReagentItemID
The one column 1.12's `GetCraftReagentInfo` does not answer: it gives the reagent's name, icon and the needed/held counts, and no item id at all, so nothing that keys a table on an item — pfUI's reagent counter — can use the stock verb. Same `(craftIndex, reagentIndex)` pair and the same rows that verb walks, so the two cannot disagree about which reagent is meant. Out of range on either index is nil, matching the stock verb's own nil-first return. All nine gates green. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
de3a3411e7 |
ui: ClassicAPI's GUID player info, no-toggle cast and stable-id skill rank
Third DLL tranche, from ClassicAPI's side of the work list. Its 70 corpus-called names are mostly **not** functions — `Clear`, `Fire`, `Size`, `Armor`, `Nature`, `Normal`, `None` and the rest are widget methods and enum keys its DLL carries as strings, and several more (`GetSpellSchool`, `RunMacro`, `GetSpellLink`) the corpus defines for itself. Reading the call sites left nine genuine gaps; three are landed here and the rest need state this client does not feed yet. - `GetPlayerInfoByGUID(guid)` -> `localizedClass, englishClass, localizedRace, englishRace, sex, name, realm`. 1.12 has no verb that takes a GUID at all. Answered through the same `Model::unit` as the whole unit surface, so a raw GUID, a `0x`-prefixed one and every other token spelling resolve alike; an unknown GUID is seven nils, not a raise. `realm` is **nil** on purpose: the field means the realm the *named* character is from, which 1.12 has no concept of, and our own realm name there would be a guess that reads as data. - `CastSpellNoToggle(spellId)`: the plain cast by id. The toggle lives in the client's cast path and benilla's queue toggles nothing — a press is a send — so this is what the caller asked for, on the same queue a bar press uses. Same reasoning as nampower's `CastSpellNoQueue`. - `C_SpellBook.GetSkillLineRank(skillLineID)` -> `rank, maxRank`. 1.12 reaches a skill only by its index into the skills *window*, which moves as lines are learnt and categories collapse — a `SkillLine.dbc` id is stable, which is why pfQuest asks by one. It reuses `skills::displayed_ranks`, so the stable-id getter and `GetSkillLineInfo` fold bonuses and the mono override identically and cannot disagree. A line the character has no row for is `nil, nil`, not `0, 0`: untrained and unknown are different answers. All nine gates green. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
801bc09c05 |
ui+docs: SuperWoW's quest-log getters, and the DLL work list worked name by name
`GetQuestIDForLogIndex(index)` and `GetQuestLinkForLogIndex(index)`: 1.12's `GetQuestLogTitle` answers the title and never the id, so neither the id nor a chat link for a log row is derivable from the stock API. The id reads the same `quest_log.entries` ClassicAPI's `C_QuestLog.GetQuestIDForLogIndex` does, so an addon that feature-detects one and falls back to the other gets one answer; a header row is id `0` and **no** link, rather than a link to quest 0. `docs/ADDON_COMPAT.md` now carries the whole work list with a status per name, and the most useful finding in it is not a missing function: `SuperCleveRoidMacros/NampowerAPI.lua` is a capability registry whose `HasFeature` checks the version **and** `_G[name] ~= nil`, so nothing on the list was broken and raising `GetNampowerVersion` would not have helped. Defining the global is the only honest way to make one live. Counts, so the list cannot be read as more alarming than it is: 16 done across this tranche and the two before it; 12 that were already present and only looked absent because the script compares against `dump_globals` on a bare VM (and because an *event* is never a global); 6 that need a feed, each with the specific thing missing; 3 that are the spell queue this client deliberately does not have; and 7 withheld. The seven withheld are `ExportFile`, `ImportFile`, `WriteCustomFile`, `ReadCustomFile`, `CustomFileExists`, `ExecuteCustomLuaFile` and `EncryptPassword`. Six of those give addon Lua read, write and **execute** on files under the client folder and the seventh takes a password. They are real in nampower and used for config export, but they are a different kind of surface from a getter and the right default for a client that did not have them yesterday is not to grow them quietly. `GetCastInfo` and `GetCurrentCastingInfo` are left out for a reason worth keeping: their consumers divide by the cast and GCD timings, and this client has a spell id, a deadline and a guard bit — no start instant, no channel flag, no GCD end. A table of three real fields and seven zeros is worse than the nil the addon's own fallback already handles correctly. 269 data-gated failures, identical to the baseline. All nine gates green. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
5555bf9c89 |
ui+app: nampower's bag group, aura cancels, no-queue casts, unit-event CVars
Second tranche. Eight more verbs and six CVar rows, all from state this client already owns. The bag group — `FindPlayerItemSlot(idOrName)`, `GetBagItem(bag, slot)`, `GetItemIdCooldown(itemId)`, `UseItemIdOrName(idOrName)` — shares one lookup, deliberately: the three that search must agree about *which* of two identical stacks they mean, and they answer the first, reading bags and slots ascending, as the reference's own auto-store search walks (`0x4f8030`) and as a player reads their bags. A name matches through the item cache, case-insensitively, since a bag slot carries no name of its own. `GetItemIdCooldown` answers the stock `0, 0, 1` for a cold or uncarried item rather than nil, because the consumer does arithmetic on the triple. `UseItemIdOrName` calls `UseContainerItem` itself rather than reaching past it, so the cursor-clear and repair-mode rules the stock verb applies still happen. `CancelPlayerAuraSpellId(spellId, [ignoreMissing])` and `CancelPlayerAuraSlot(auraSlot)`: the two keys into the `CMSG_CANCEL_AURA` queue `CancelPlayerBuff` already fills, so the three cannot disagree. The slot is the **raw 0-based** descriptor slot upstream documents and the buff events report, not the Lua index. `ignoreMissing` is accepted and changes nothing — it asks upstream not to complain, and nothing here would have complained. `CastSpellNoQueue(spellId, spellBook [, unit])` and `CastSpellByNameNoQueue(name [, unit])`: the plain cast, which is the honest mapping and not a stub — benilla has no spell queue to bypass, so "cast this now, do not queue it" is exactly what happens. The point of them is the first argument: a **spell id** where stock `CastSpell` takes a book slot. `"mouseover"` routes to the queue SuperWoW's `CastSpellByName(name, "CLICK")` already uses. The six `NP_EnableUnitEventsX` switches are registered stored-only, one step further back than the four `NP_Enable*` rows beside them: benilla raises no GUID-keyed unit event at all, so there is no stream for a token filter to narrow. Registering them stops the unknown-CVar warning and lets `SuperCleveRoidMacros`'s writes persist; it claims nothing about the events arriving, and `cvars.rs`'s own reader gate carries that reason. All nine gates green. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
f120e8003d |
ui: six of nampower's reads, and the item-icon probe behind two of them
First tranche off the DLL work list. The method that produced it is worth recording with the code: `SuperCleveRoidMacros`'s `NampowerAPI.lua` is a capability registry — a `VERSION_REQUIREMENTS` table of every nampower name with its minimum version, reached through `HasFeature`, which tests the version **and** `_G[name] ~= nil`. So none of this was broken, and raising `GetNampowerVersion` would not have helped: the only honest way to make one of those features live is to define the global. - `PlayerIsMoving()`, `PlayerIsRooted()`, `PlayerIsSwimming()` — `1` or nil, as upstream's own doc lines state, never a boolean. All three read the per-tick posture snapshot `SelfMovement` already carries, which gains `rooted` (`MOVEFLAG_ROOT`) and `moving` (any of the six movement bits, the keyboard turns included — they are input to the same word the client tests, and a turn in place does interrupt a cast). - `IsAuraHidden(spellId)` — from a new `SpellLookup::hidden_aura`, filled with exactly `ui_aura::shown_in_aura_ui`'s predicate inverted, so it cannot contradict the aura list an addon reads beside it. An unknown spell is nil: "not hidden" is the honest answer about an aura this client knows nothing about, and the lists show a catalog miss too. - `GetItemIconTexture(itemId)` — two hops, because an item template carries a display id and not an icon: the cache row's `displayInfoID`, then `ItemDisplayInfo`'s `Icon` through a new `ItemIconProbe`. A known item with no display row is nil (a data gap); an item the cache has never seen raises and records the ask, as every `item_record` verb here does. ClassicAPI's `GetItemIcon` wants the same probe, which is why it is a probe and not a closure. - `GetTrinkets()` — slots 13 and 14 in `GetEquippedItem`'s own row shape, an empty slot absent rather than nil-valued, since the consumer walks it with `if t[i] then`. Each is in `reference_surface.rs`'s `allowed_beyond_1_12` with why 1.12 has no spelling to prefer — the client reads its own movement word and tells Lua nothing, applies the aura attribute test inside `UnitBuff`/`UnitDebuff` without exposing either, and can only reach an item icon through a bag slot it already holds. All nine gates green. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
ada08720bb |
ui+docs: the GUID token's real spelling, 40-yard interact, and the whole patch map
Working down the 29 `.tdata` hooks from the realm's patched `WoW.exe`.
Most turned out to be already ported; two were wrong here, and one of
those was wrong in a way that mattered.
**The GUID token was the wrong string.** `unit::guid_token` wrote 16
lowercase hex digits and no prefix. The realm writes `0x` + 16 upper-case
digits: its formatter at `0xd06500` — which `UnitExists`'s second return
goes through — lays down `0x30 0x78` then sixteen nibbles each `+0x37`
above 9, and its parser at `0xd06650` demands the `0x` and takes upper
case only. grom's own addons agree from the other end: BigWigs decides a
string is a GUID with `string.sub(guidOrName, 1, 2) == "0x"`
(`MageTools.lua:963`), and AI_VoiceOver reads fields from position 3 and
mints its own with `format("0x%04X%06X%06X", …)`.
Same number, wrong string, and wrong where it mattered most:
`UNIT_CASTEVENT` and the aura events carry these tokens, so every such
test read our GUIDs as unit *names* and every `sub(3, …)` field read came
out two characters off. Now written the realm's way; both spellings and
either case are accepted on the way in, since refusing one buys nothing.
Pinned literally rather than by round trip — a round trip held fine
through the whole bug.
**`CheckInteractDistance` type 4 is 40 yards, not 30.** `0x48bafd` jumps
to `0xd0686c`, which tests the type against 4 and compares against the
float the section carries at `0xd06868`: `1600.0`, squared thresholds.
Stock 1.12's `900.0` is still what the other three types use.
`docs/ADDON_COMPAT.md` now carries all 39 hook sites of both sections
with a status each: 10 in `.octoql` (one feature, the quest log, ported),
and 29 in `.tdata` — the aura ids, the master looter, `UnitExists`'s
GUID, the char-create mask and `CHAT_MSG_HARDCORE` already done; the two
challenge-mode functions needing nothing, because the realm's own
`ChallengeList.lua` defines them in Lua for exactly our case; four read
but unported with the reason (unnamed struct fields, no consumer in the
corpus); and three that cannot apply to a client that loads no DLLs.
Also `scripts/dll-surface.py`, which asks the injected DLLs the same
question. Their export tables are one `DllMain` each, but a DLL that
registers a Lua function must point at the name: collecting NUL-delimited
identifier strings the image holds a pointer to cuts `ClassicAPI.dll`
from 2872 CamelCase candidates to 706, and subtracting the 1.12 census,
Lua's own names and benilla's real surface leaves a candidate list the
addon corpus can rank. It is a list to read, not a specification —
statically linked libraries contribute their own strings, and
`UnitXP_SP3.dll` yields nothing. It found 38 names in `nampower.dll` that
grom's addons call and this file did not know about; they are written
down rather than guessed at.
269 data-gated failures, identical to the baseline, all nine gates green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
cf0e56a4cf |
ui: give the worst-call ordering test margins a stall cannot invert
`the_worst_call_is_one_calls_cost_and_not_the_mean` reads both its orderings off a wall clock, and the margins were too thin to survive the full suite: it failed about one run in three, on an idle box as well as a loaded one. One of the steady handler's ten calls would pick up a ~0.9 ms scheduler stall and overtake the spike it is meant to rank below — `self_us` 5109 / `worst_us` 1405 against the spike's 1398 / 1361. Twenty steady calls instead of ten raise the window cost without raising any single call's, and an 8x spike instead of 3x puts the worst-call margin about four times beyond the stall that broke it. Six consecutive full-suite runs clean, then a full gate chain. The doc comment carries the measured numbers and says to keep both ratios, since the absolute durations are not what the test rests on. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
49b4f7b55e |
ui+app: say what a fired event reached, and whether the UI answered
grom's stale-minimap report survived the same-interior-skip fix, and his log says why that fix was the wrong suspect: at the gate the zone-text interior claim reads `outdoors`, so the skip was never consulted there. The module doc no longer claims otherwise. What his run leaves is sharper and stranger. The election cannot have declined — leaf 1537 → 809 "Gates of Ironforge" takes the zone from 1537 to 1 and the subzone text with it, so both `ZONE_CHANGED_NEW_AREA` and `MINIMAP_ZONE_CHANGED` fire — and two separate addons call `SetMapToCurrentZone()` on those while the map is closed (`pfQuest/map.lua:1466`, `pfUI/modules/minimap.lua:183`). Zero calls arrived, with nothing raised anywhere in the session. Offline, against the same addons and saved variables, both events reach every registered handler and are answered every time, so the difference is live-only and the engine cannot currently see which of three things it is. So it can now. `UiScript::fire_event_reach` answers an `EventReach` — frames registered when the dispatch began, frames whose handler ran, and the name of the frame a cut-short walk stopped on (our documented deviation: a handler that unregisters the walk's saved next ends the dispatch, and every later registrant silently never hears the event). `UiScript::frame_shown` reads the shown bit addons branch on, which no amount of engine-side logging could otherwise explain. `fire_event` is unchanged for its nineteen callers. `area::feed_zone_events` logs the whole handshake at each fire: what was elected, reached/registered per event, the `SetMapToCurrentZone` delta, and `WorldMapFrame`'s shown bit. `0/0` is nothing registered, `reached` below `registered` is a cut-short walk, and every handler run with `0×` asked is the guard both answerers share. `addon_harness --fire` reports the reach too. Gates green (459 s). His corpus offline: 10/10, 8/8 and 4/4 handlers reached. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
9e13856cfd |
app: say in the log when the world map stops following the player
The stale-minimap report needs one fact: did the selection follow the player into the new zone? Everything map-derived rests on it, because `GetPlayerMapPosition` projects onto the **displayed** map — a selection left a zone behind seats pfQuest's minimap pins against the wrong rect, and plausibly rather than obviously wrong wherever two rects overlap, as Ironforge's does inside Dun Morogh's. The `zone:` line could not answer it even after it started printing both pairs, because the pairs disagree constantly and mostly correctly: for a frame or two after every crossing, since this feed publishes the new zone and the `ZONE_CHANGED*` an addon answers by reselecting fires after it; and for as long as the map is open on another zone, which is what browsing is. A marker on that line would have cried wolf on both. So the rule is its own thing, `SelectionLag::step`: an episode starts only on a frame where the *player* moved to a zone the map is not on, and only an episode outliving a 1 s grace is reported — once, at WARN, with the catch-up line when something finally reselects. A selection that wanders under a standing player is never an episode. Five tests, including the 600-frame browse that must stay silent. The line also forks the cause, which is the part that saves a round trip: `UiScript::world_map_resolver_calls` counts Lua's `SetMapToCurrentZone` calls, so the report can say whether the UI never asked to follow — the event did not reach the addon — or asked and still landed wrong, which is the verb reading a zone the feed had not published. Same symptom, different fixes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
ce8992bc3e |
Draw flight masters: C_TaxiMap.GetTaxiNodesForMap over TaxiNodes.dbc
grom: "can't see flight master location marks, even tho I have them on."
Nothing was broken — the layer was silently empty. pfQuest-octo has deleted its
shipped pfDB.meta.flight table in favour of this call (db/meta.lua:181 is a stub
whose comment says the data "always comes live from the client (ClassicAPI)"),
and SearchFlightNodes opens with
if not (C_TaxiMap and C_TaxiMap.GetTaxiNodesForMap) then return maps end
so with the namespace absent the option stays ticked, no error is raised, and no
icon is ever drawn. Same silent-by-construction shape as C_Map.GetAreas, same
consumer.
The contract is half read off the bundled ClassicAPI.dll's string pool
(C_TaxiMap, GetTaxiNodesForMap, nodeID, faction, reachable, state and the
FlightPathFaction/FlightPathState enum names) and half off the caller: the
areaID/mapX/mapY triple is not separately visible there, since the pool holds
one copy of each string and C_Map.GetAreaTriggerInfo already uses all three, so
for those SearchFlightNodes itself is the authority. It reads six fields, gates
on `node.areaID and node.mapX and node.mapY`, and feeds the pair into
pfMap:AddNode as meta.x/meta.y, which UpdateNodes divides by 100 — so they are
percentages of the zone rect. Both files say which half is which.
Faction comes from the two MountCreatureID columns, the reference's own gate
(0x4a2650 asks for the mount of the player's half): one column filled is that
faction, both is neutral, and neither means the row is not a flight master at
all — on this install those are exactly the eight Transport/Ferry/zeppelin
endpoints, and passing them on would put "Flight Master" pins on the Menethil
docks.
Attributing a node to a zone needed the file's own names. Geometry over
WorldMapArea boxes proposes — these are axis-aligned boxes around irregular
zones, so a node near a border sits inside two or three of them — but no purely
geometric rule works: "smallest box" misfiled a dozen of ~70 rows (Crossroads
under Mulgore, Aerie Peak under Alterac Mountains), and "most central box",
much better and what zones_at now sorts by, still loses Talonbranch Glade to
Darkshore by about one percent of inset. TaxiNodes.dbc names its rows
"<place>, <zone>" and a city row leads with the city, which is itself a zone, so
trying each comma-separated part against the candidates gets Ironforge over Dun
Morogh *and* Felwood over Darkshore from one rule. zones_at also reaches the
instance-map list, so Alterac Valley's two nodes place on the battleground map
instead of nowhere.
Seeded before the interface loads, after the map catalog it reads the rects
from. That deadline is tighter than file scope: pfQuest replays its saved
tracking selections once on PLAYER_ENTERING_WORLD and then unregisters every
event (database.lua:131-151), so a push one frame later is a push after the only
read.
Verified against the real addon offline, not just by unit test:
SearchFlightNodes now adds 69 nodes through the probe harness. Three tests pin
the rule against the DBC itself — the capitals and the border nodes by hand, and
every named row held to the zone its own name claims, which is the test that
rejected both earlier geometric rules. It also pins the four rows that name no
zone, three of which are this realm's own additions (Northshire Abbey and two
Darkshore canoe routes), so the next custom node shows up as a failing
assertion rather than a pin in the wrong zone.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
b4a5d3b38f |
ui: "do I know this spell?" asks the known set, not the spellbook
Every "Already known" check read `spellbook.slots`, and `SpellDisplay::in_spellbook` filters out `SPELL_ATTR_IS_TRADESKILL` — correctly, since recipes live in the trade-skill window, not the spellbook. So the question "do I know this spell?" answered **no for every crafting recipe in the game**: a recipe item never showed its `ITEM_SPELL_KNOWN` red line, and nothing stood between the player and consuming it a second time. grom learned Plans: Copper Chain Vest (item 3609) twice. `PlayerReqState` gains `known_spells`, filled from `PlayerActions::spells` — `SMSG_INITIAL_SPELLS` plus every learn, which is the set the reference's own check reads (`0xb700b0`). It belongs there: that struct is already "the player state the tooltip's red lines and the usable gate compare against", already change-gated on the action/spell inputs, and already pushed by the same feed. Both readers move over: the taught-spell line and `item_usable_by_id`'s spell requirement, which had the same hole. The existing `verified_families_signable_locked_resists_known` only passed because it made the recipe known by putting it in the spellbook, which no real session does — it now sets the known set instead, and that is the change's own regression story. The new test leaves the spellbook deliberately empty, which is the real shape and what fails if the check ever reads it again; it asserts the not-yet-learned case too, so it cannot pass by always reading red. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
a2da3e4da7 |
ui: a skill lock you can open stops reading red
The GameObject tooltip's skill-lock line took the opener-*unknown* arm unconditionally — `LOCKED_WITH_SPELL` in red — whatever the player's skill. The known arm was simply not built, and the code said so where it was missing. The reference forks on whether any known spell opens the slot's LockType (`0x52abcb`): known takes `LOCKED_WITH_SPELL_KNOWN`, which carries the requirement as a `%d` and the colour of the skill-difficulty ramp `0x529fa0`. Built now, both halves. grom reported it on a copper vein at Mining 70, which under the ramp is +69 over the vein's requirement of 1 — the green rung, not the grey one he expected; grey starts at +100, so a copper vein greys at Mining 101. Either way not red, and red is what it showed. The resolve is hoisted out of the `flag_locked` branch, because the "Locked" line's colour and this line's known arm are two readings of one answer and must not disagree with the click. `lock::skill_margin` recomputes the two numbers from the opener the resolver already matched, with the same two functions its own value test uses (`0x5f850f`), so hover and click cannot diverge on which opener is in play. Deviation, recorded at the constants: the three new rung RGBs are the standard trade-skill palette, not values read out of the binary's table. Only the ramp's shape is traced — five rungs, bands at +0/25/50/100 — and the rung a line lands on is what the player reads. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
e6ae2e0f67 |
ui: a scroll names the child it moved instead of giving up on the graph
`SetVerticalScroll` and `SetHorizontalScroll` called `Model::touch_layout` on the line before `reanchor_child`. That gave up naming for the whole graph and then *erased* the precise touch that followed: `reanchor_child` re-anchors the scroll child through `write_anchor`, which reads an offsets-only change (same point, same target) and names the child node — but `touch_layout_node` will not un-say an imprecise touch, so the give-up won on every call. Naming the child is the whole truth here. A scroll offset reaches the layout only as that one anchor: anchor targets are ids fixed at `SetPoint`, so no edge moves and the roster is unchanged, and the clip a scroll child draws within is the ScrollFrame's resolved rect rather than the offset — `clip::scroll_clip_sources` rebuilds it per call and caches nothing. Measured on grom's 2026-10-21 17:08 session, which was a map-and-panel run. These two lines were **472 of the run's 806 whole-graph derivations**, the largest single source by a factor of two, and the cost lands on the Lua that reads a frame edge afterwards rather than on the pass's own `resolve` phase: | | n | fps | UI pass | `frame:GetLeft` | |---|---|---|---|---| | windows with >= 10 scroll derivations | 12 | 68.5 (min 53) | 5.93 ms/f | 2.18 ms/f | | windows with no derivations at all | 11 | 92.4 (min 77) | 1.62 ms/f | 0.05 ms/f | `r(scroll derivations, GetLeft us/f) = +0.907` over 57 in-play windows, against `+0.014` for every other derivation site. `[ui-verbs]` has `frame:GetLeft` at 94-376 us a call in those windows and 0.6-1.5 us in the quiet ones, on ~14 calls a frame, because stock FrameXML reads frame edges constantly and each read past the clean epoch re-derived. That is why a scrolling panel — the world map, the quest log, a profession list — was the expensive thing on screen. The test scrolls both axes on the shipped interface and asserts zero derivations, with a positive control so it cannot pass on a dead counter. Put one `touch_layout` back and it reads exactly 10 for 10 scrolls. `WOW_LAYOUT_VERIFY=1` over the real 92-addon interface gives `frames 13379 -> 13379, regions 29597 -> 29645` — identical to the recorded pre-existing baseline, so this narrowing introduced nothing. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
754d577eaf |
ui: a trade-skill list mutator signals inside its own call
`ExpandTradeSkillSubClass`, `CollapseTradeSkillSubClass`, `SetTradeSkillSubClassFilter` and `SetTradeSkillInvSlotFilter` fire `TRADE_SKILL_UPDATE` from inside the call now, as the one helper all four tail-call in the reference does (`0x4fd710`, `0x4fd730`, `0x4fd750`; `reference/1.12-verb-events.tsv`). It used to be a `trade_skill_touched` flag drained by `ui_tradeskill::drain_trade_skill` and answered the following frame. Inline is load-bearing. These four verbs are called *from* a `TRADE_SKILL_UPDATE` handler: a profession library reads the recipe list by unfolding every header, walking it and folding back what was folded. The signal is therefore re-entrant by construction, and every addon that does it carries its own guard — LibCraftingProfessions has `trade_skill_scan_in_progress`, ATSW an `atsw_scans < 2` counter. Those guards are written against an inline signal and only an inline signal lands inside the window they cover. Deferred to the next frame it landed after the guard had cleared: the library rescanned, the rescan called the verb, the verb re-armed the flag, and the two kept each other alive at one fire a frame for as long as the window was open. grom's 2026-10-21 session shows it directly — `[ui-events]` reporting `TRADE_SKILL_UPDATE · 1.01 fires/f · ~900 call(s)` across 37 consecutive windows, three unbroken minutes, at 0.58-0.84 ms a frame in handlers over four addons rebuilding their lists for nothing. It is the 118-second and 61-second stretches at 55-70 fps in a session whose in-play median was 86. Two tests: each mutator's fire is read with its own `run` still on the stack, so a deferred signal fails; and a library-shaped in-progress guard turns one real update into exactly one scan, with nothing owed to a later tick. The bench's `[events]` block also fires the five non-quest events the live session named. They cost ~0.2-0.5 ms there against 12-72 ms live, because this harness has no NPC in front of it, no trainer list and no open map — recorded as a gap, not an answer. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
e269e4dbb4 |
ui: name the event, not the library frame that dispatched it
`WOW_UI_EVENTS=<secs>` (`script/event_prof.rs`): per event, what one fire
costs across every handler it reaches, and which frame took the most of the
worst fire.
The gap it fills is a dispatcher. `WOW_UI_HANDLERS` keys on `(frame, script)`,
which is the right key for an `OnUpdate` and the wrong one for an `OnEvent`,
because every addon library registers *one* frame for *every* event it wants and
fans out in Lua. So the largest identifiable cost in grom's 2026-10-21 session,
eleven separate times, read as
[ui-handlers] 312.5 312.5 0.00 72522.3 AceEvent30Frame:OnEvent
— a true row that names a library's shared frame and no addon, no event and no
cause. `[sys-timing]` saw the same cost as `ui_quest::feed_quest` at 85.7 ms for
one run in 76, because every handler runs synchronously inside whichever system
fired, so the system holding the stack gets the bill.
Two conventions the report sticks to:
- `self` is nest-adjusted, in both columns. A handler's own time has any event it
fired taken back out (`event::fire` differences `event_prof::nested_ns` across
the call), or every event in a cascade reports the cascade and the row naming
the real cost is the one that looks innocent. `total` keeps it, which is what a
`[sys-timing]` span sees.
- A fire is one `fire_event`, however wide it fans out, because that is the unit
a dropped frame is measured in. `calls` is beside it for the width.
Measured on the player's own 92-addon interface (`tick_bench`'s new `[events]`
row, three fires per event from a settled start):
us/fire worst event
5068 14759 QUEST_DETAIL <- HCQuest.lua:139 took 14,097 us
764 1324 QUEST_LOG_UPDATE
662 1564 QUEST_FINISHED
Read `us/fire` against `worst` and that 14.8 ms is **one** fire of three: the
other two cost ~0.1 ms, so it is HCQuest building something once and keeping it,
not a cost paid per quest. With that read correctly, no quest event costs more
than ~1.5 ms recurring offline — so the bench does not reproduce the live 57-73
ms and says so, rather than offering a number that looks like an answer. Hence
the instrument, armed in `./measure.sh`.
Also adds `./measure.sh --frames`, which arms two meters that already existed and
have never been armed: `WOW_FRAME_PHASES=50` and `WOW_RENDER_MS=1`. A frame's
wall time is `max(main app, render app)` since bevy pipelines the render app onto
its own thread, so when the render app is the longer, `wall` climbs while `cpu`,
`main`, `gpu` and every per-system `worst` all read healthy — exactly what that
session's **161 ms frame in a second whose costliest system worst was 2.9 ms over
105 runs** looks like, and what 43 of its 64 spike-seconds look like. Its own
mode because both perturb what they measure.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
0c862838da |
ui: a SetScale names its subtree instead of giving up on the graph
`[ui-phases]`'s new write-site tally answered the combat-fps question on its
first armed session: 463 whole-graph layout derivations across 34 windows, and
**232 of them** at `object/frame_state.rs:427` — the `touch_layout()` inside
`SetScale`. Splitting the per-window count by site and regressing against
resolve time over 43 in-world windows, resolve correlates with `SetScale`
derivations at r = +0.996 and with every other derivation at r = -0.041. The
four costliest windows of the run are 36-64 `SetScale` derivations each and
almost nothing else.
That also corrects the flat "~15.5 ms per derivation" of the round before: a
mint-triggered derivation costs ~1.1 ms, a `SetScale`-triggered one 12-13 ms.
The difference is that `SetScale` gave up *two* ledgers, not one.
The caller is pfUI's action-button press animation
(`pfUI/modules/actionbar.lua:182`, `:207`, `:233`, `:259`): four styles that all
run `this:SetScale(this:GetScale() + fade)` on `OnUpdate` for the ~60 frames a
button's fade lasts. Pressing an ability starts it, which is why a settled
interface never reproduced this and a fight always did.
Both give-ups had precise versions already, written for `SetParent`:
- `touch_layout()` -> `touch_layout_subtree_scale(h)`. A scale write moves the
subtree's `effective_scale` and nothing else: anchor targets are ids fixed at
`SetPoint`, so no edge moves and the roster is unchanged.
- `touch_measure_all()` -> the same walk now names the subtree's regions on the
measure ledger. This was the bigger half. `measure_dirty = None` makes the
next `fontstrings_needing_measure` walk all ~43,000 regions of a real
interface to issue, typically, zero requests; `stale_measure_key` reads only a
region's own data plus *its owner frame's* `effective_scale`, so the subtree's
regions are exactly the ones that can go stale.
Measured on the player's own 92-addon interface, 43,195 regions, one `SetScale`
per frame for 240 frames (`tick_bench`'s new `[scale]` row):
before changing resolve 0.601 ms/f · 0.996 derivations/f (239/240)
after changing resolve 0.005 ms/f · 0.000 derivations/f ( 0/240)
control same resolve 0.005 ms/f · 0.000 derivations/f ( 0/240)
`changing` is now indistinguishable from a control that writes the value it
already holds. The harness understates the live cost (its measure answers are a
6px-per-char stub), so the live saving is to be confirmed in a play-test, not
predicted from that row.
Correctness is left to `WOW_LAYOUT_VERIFY`, which `cfg!(test)` forces on for all
1975 benilla-ui tests: it re-derives the whole graph after every `resolve()` and
panics on any rect the incremental pass got wrong, and runs the full measure
roster beside the ledger drain and panics on any stale region the ledger missed.
So the two new tests assert only that the cheap path was taken, on a graph three
frames deep with a FontString on the leaf. The real-interface verify run fails
identically to its documented pre-existing red (`frames 13379 -> 13379, regions
29597 -> 29645`), so this change adds nothing to it.
Also repairs an assert message in `layout_gate.rs` whose line continuations had
been eaten by a heredoc, leaving runs of spaces mid-sentence.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
c1a56f98ab |
layout: name the write that asks for a whole-graph derivation
`[ui-phases]` could say a window paid for 61 whole-graph derivations. It could not say which of the forty-odd `touch_layout` sites asked for them, and that is the half a fix has to be aimed at. Measured on the player's interface, one derivation costs ~15.5 ms — from two directions: the derivation count explains total resolve time at r = +0.977 across twenty windows (solve count: r = -0.33, so solves are free and derivations are everything), and 26 of 78 worst-single-call figures over 1 ms land in the 15-16 ms bin across eleven unrelated handlers by six authors, two of which fired exactly once for their whole window. Eleven handlers cannot each independently be 15.5 ms slow; it is one shared charge landing on whoever reads the graph first, which is also why the same session shows `resolve` p99 at 14.4 ms and worst frames "mostly tick". One defect, two accounts. So `#[track_caller]` down the naming chain and a tally of `file:line`, printed in the report. Not a backtrace: `WOW_LAYOUT_DERIVE_TRACE` already captures those and has to be rationed to `n` prints because of what they cost. This is a `(&'static str, u32)` key off `Location::caller()`, cheap enough to leave armed, and take-and-rearm so a window reports its own rows and a VM replaced by `/rl` re-arms without anybody noticing it was replaced. Also a `[churn]` window in tick_bench, which records a negative: a plate roster turning over 4-in-40 every ten frames costs **zero** derivations on the player's own interface. `sync_nameplates` pools the plate frames, so a new key rewrites an existing plate and mints nothing. With `resolve_bench`'s hover-sweep, tooltip-content and moving-region tests all asserting zero, four candidate triggers are now ruled out offline and the live tally is what names the real one. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
112647b919 |
perf(ui): the handler profiler names the worst single call, not only the mean
`WOW_UI_HANDLERS` ranked handlers by self time over the window, which is the right ranking for what costs framerate and the wrong one for what costs a frame. A handler that stalls once every four seconds has a mean near zero over five hundred frames and never reaches the table at all. The player's session had a 15-28 ms tick in nearly every five-second window — one dropped frame every few seconds, which is the kind of thing a player reports as "it hitches" and no average will ever show. The table said `UIParent:OnUpdate 60.8 us/f` and nothing about it. So each slot keeps its costliest single call, the table gets a `worst` column, and the handlers whose worst call cost a frame are listed again underneath — but only the ones the table left out, so the block says something new or nothing. `tick_bench` prints the same ranking. It names the stall on the first run, offline, in this interface: ``` 15768.7 us once (self 17080.8 us over 240 calls) pfQuestMapTracker:OnUpdate 12339.4 us once (self 13042.0 us over 240 calls) AceEvent20Frame:OnUpdate 5380.9 us once (self 16166.0 us over 240 calls) pfQuest\quest.lua:348:OnUpdate ``` 92% of the tracker's whole cost over 240 frames is one call. Both of the top two are steady per-frame allocators, which made the collector the first suspect; the bench's own heap line refutes it — 222.8 MB live against 29.1 kB/frame of churn is ~131 s between 200%-pause cycles, not the ~4 s the stall recurs on. What the per-binding bill says is in there instead: one `pcall` of 7.5 ms, and `frame:GetLeft` at 92 us a call, which is a layout resolve against a dirty graph. `measure.sh` comes with it. The set of instruments worth arming for a play-test is long enough that a run with one of them missing is normal, and such a run looks exactly like a run that found nothing — the session this column came from was missing `WOW_SYSTEM_TIMING`, so it could say what the UI cost and not what the frame cost. The script arms the set and prints what it armed. The test asserts the property the column exists for and not a duration: a handler called ten times for 50k iterations each outranks one called once for 150k by self time, and the second still has the larger worst call. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
a8b6fc8ad6 |
perf(layout): the resolve's seed pass walked the graph, not the change
Each solve chooses which nodes to solve and which clean rects to seed it
with. It did that by asking **every id in the graph** — 18 203 frames and
42 284 regions on the player's interface — "is one of your dependents
dirty?", for a dirty closure whose steady-state median is 2 frames and 4
regions.
And nothing was measuring it. `[layout-pre]`'s last lap ends at the change
gate; `[layout-prof]`'s `frame_us`/`region_us` start at the round loop. This
pass sits between them, so neither line charged it — which is how a
`frame:GetLeft` that measures 17.5–18.4 us a call in a live session (six a
frame, through stock `MouseIsOver`'s four edge reads) could force a resolve
the profiler called 3 us. A `seedsel_us` column now closes that gap, and it
stays: a phase between two timers is a phase nobody owns.
Measured on the player's own 80-addon interface, last-quarter solves:
median p75 p90 mean max
before 458 498 607 488 882
after 0 3 9 38 3788
`solved` and `swept` come out identical to the digit (median 2/4, mean
156.3/408.3, max 18183/41415): the same work is selected. Before, 202 of 549
solves cost over 400 us and the dearest of them had dirty sets of 0–13 nodes,
which is the pathology itself; after, 29 of 549 exceed it and every one is
solving 14 000–16 600 frames at 0.48 us a node, against a `frame_us` of
10 170 us on the same pass. The cost moved from the size of the graph to the
size of the change, which was the point; the higher max is that, not a
regression.
**What it is not: an fps fix.** The bench's per-frame totals do not move —
1.294 -> 1.293 ms/frame grouped, `vm-tick` 0.667 -> 0.676 (noise, single
runs). The one column that shifts is `resolve`, 0.011 -> 0.006 ms/frame. The
458 us is per *solve*, and a steady frame does a fraction of one; the live
bound is the whole of `frame:GetLeft`, 105-110 us/f, against a 7.6-9.5 ms
frame. So this removes real waste and buys about 1%, which is worth having
and is not what grom asked for. Chasing it at all was a misread of mine:
`GetLeft` is 54% of the *binding* bill, and the binding bill is ~200 us of a
~9000 us frame.
The same set, by the same relation read the other way round: a clean node is
seeded iff some dirty node reads it, and a node's dependencies are exactly
the three `scope.edge` arguments in the roster walk — its anchors' targets,
plus a region's owner frame, whose scale it reads. `LayoutScope::dirty_ids`
carries the dirty set as a list, written wherever `dirty` turns true so the
two cannot disagree, and `seed_dep` enumerates the dependencies. A
`full_scope` pass skips the walk: everything is dirty, so there are no seeds
to find.
Order is not a property here, which is what makes the rewrite safe. Both
rosters are built from `HashMap` iteration (`frame_to_id`, `region_data`), so
no pass ever had a canonical order, and the region sweep is a fixpoint whose
round cap covers any chain — order moves `rounds`, not rects. Only the seed
lists are sorted, to drop duplicates a shared dependency would otherwise
re-`set_external` every round.
`verify_selection` is the falsifier: under `WOW_LAYOUT_VERIFY` it recomputes
all four lists the old way and requires the same sets. That is a tighter
question than `resolve_layout`'s rect comparison — it isolates this pass, so
a selection bug cannot hide behind an unrelated red, and an unrelated red
cannot be pinned on the selection. Verified non-vacuous by dropping the
region-owner dependency and watching it fail (`0 vs 1 entries`).
Note for whoever runs it: `WOW_LAYOUT_VERIFY=1` over `tick_bench` is **red at
HEAD**, independently of this change — an incremental pass ends with 48 fewer
region rects than a full derivation of the same frame (`regions 29597 ->
29645`), byte-identical with and without this commit. That is a missing
`touch_layout()` somewhere, not a selection bug, and it had never been run
before because `tick_bench` is `--ignored` and the flag is in no gate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
7322a95a9d |
fix(ui): WOW_UI_VERBS must not wrap setfenv — it unloaded a third of the interface
A wrapper puts its own frame between the caller and the callee, which silently changes the meaning of every stack-level-relative call. `verb_prof::install` wraps every function in `_G`, and `_G` at that point holds Lua's own stdlib as well as our bindings — so it wrapped `setfenv`. pfUI's `api/api.lua:4` is `setfenv(1, pfUI:GetEnvironment())`, and level 1 from inside a wrapper is the wrapper, a C function, so Lua raises ``setfenv' cannot change environment of given object`. One armed session logged **91** of those and lost pfUI, WhatsTraining_Turtle, AI_VoiceOver, MasterTradeSkills, pfUI-addonskinner, pfUI-locplus and PizzaSlices: an instrument that unloaded a third of the interface it was there to measure, and whose numbers were therefore of a client without them. `STACK_RELATIVE` names the four `wrap_table` must leave alone. `setfenv` and `getfenv` take a level; `error(msg, level)` positions its message by one; and `debugstack([start])` is the reference's traceback by level, which `FuBarPlugin-2.0.lua:752` asks for as `debugstack(6, 1, 0)` and `AceDB-2.0.lua:742` reads a specific line of. `pcall`/`xpcall` stay wrapped: no level argument, and an armed session shows them working and worth seeing. The test wraps `_G` and then runs pfUI's line, so it is the bug rather than a restatement of the list; deleting the skip reproduces grom's exact error. The module doc also claimed `_G` "holds our bindings and nothing else" at install time, which is what made this easy to miss. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
5764837e35 |
perf(ui): give WOW_UI_VERBS a live window, so the instrument exists in a session
`verb_prof` wrapped every global and every widget method and charged each call two `Instant::now`s — and the only caller of `log_report` was the `addon_harness` **example**'s `--together` load. Armed in the client it paid for every binding call and printed nothing at all. The one question it answers — which binding a handler's self time went into — is exactly the open one: `WOW_UI_HANDLERS` names `UIParent:OnUpdate` at 503 us/f in combat against 48 idle, and stops there. So `WOW_UI_VERBS=<secs>` now prints a window from `report_window`, called at the end of `UiScript::tick` beside the handler and raise reports, one level down from them. The bare `=1` the harnesses already pass reads as one second, which is why `period()` parses the same variable rather than taking a new one. A window is a **delta** against a snapshot (`PREV`), not a `reset`: `ROWS` is the one table and `log_report`'s callers read it whole at a moment of their own choosing — `--together` ticks ten times before printing its load bill, so a window that cleared the table would have handed it whatever happened after the last window instead of the load. The arithmetic is `diff`, split out and tested on its own because `window_delta` reads two process-wide statics and `diff` is the whole of the logic. It earns its keep immediately. `tick_bench` over the player's real 94-addon folder, 121-frame windows: `frame:GetLeft` costs **0.42 us/call** on a settled layout and **29.95 us/call** the moment anything moved — the same two calls a frame, 55% of the frame's whole verb bill — because the four edge getters call `settle()` and a resolve past tier 1 walks every live frame. During the load burst the same call is **370 us**. `GetWidth`/`GetHeight` already carry an authored-span shortcut for precisely this; the edge getters never got one. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
497570a68c |
addons: an on-demand load says what it loaded
The startup walk names every addon it loads; `LoadAddOn` named none. So when the client froze on 10-19 while learning recipes, the session log could not say whether the trainer window was even open — and `Blizzard_TrainerUI` is LoadOnDemand, so that was the whole question. `benilla-ui` has no logger, so the line travels the way its errors and warnings already do: a `notices` channel on the model, drained to `info!` beside them in the same pass. Silent for a redundant load, because the stock `*_LoadUI` glue calls `LoadAddOn` on every open and a line per call would bury the one that did the work. A refusal carries the reference's own reason token, and a name that is not installed says so. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
9782b07fd4 |
lua: a frame that never leaves the VM fails its call, not the client
The load has had an instruction bound since `3b6975fd` — "so a hung load reports instead of freezing the loading screen" — and the end of the load disarmed it, "so no handler is killed for being slow". Right about slow, and it leaves a handler that never returns at all able to freeze the client until the player kills it, which is what happened to him this morning. So the bound does not go at the end of the load, it changes shape. Wall time rather than instructions, because the number has to be obviously beyond anything legitimate and a count is not: a frame still inside Lua after fifteen seconds is not slow, it is stuck. `tick` restamps every frame, so no frame is charged the one before it and a handler reached later in the frame — a click, a fired event — measures from that mark, which can only make the bound tighter. The raise names the chunk and the line, so his log says which addon rather than that something somewhere looped. It shares the one count hook, which costs 5-13% of pure Lua execution time (measured: 30M instructions, 235-260 ms bare against 267-274 ms hooked, which also puts his machine at ~125M VM instructions a second — what sizes the load's 200M at about 1.6 s). Lua is a small part of a play frame, so that is a fraction of a percent of frame time, and the load was already paying it. `WOW_LUA_RUNAWAY=<secs>` moves the bound, `0` disarms it. Two tests: a loop with no exit fails its own call, names its chunk and leaves the VM usable; and a frame is not charged the wall time that passed before it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
b7ddc18782 |
load: a reload says what it cost and what the chunk cache did for it
Both lines sat on the world-entry path, so the load the player actually waits through — the `/reload` — reported nothing at all. His 2026-10-19 session has two of them, 4,981 ms and 6,276 ms against the 7.5-9 s he started from, and the log cannot say where either went or whether the bytecode cache was hit, although the cache directory proves it was: 1,585 entries written at first load, 14 added across both reloads. `report_load` now runs on both roots and names which one. The reload's clock starts before the saved-variable writes, because those are part of the one blocked frame he sees, which is also what the `frame hitch` warn measures. `chunk_cache_line` resets its counters on read, so each line is that load's own work rather than a running total in which a warm reload hides behind the cold login before it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
d5ddaa1286 |
lua: a compiled chunk is kept, so a load parses only what changed
The compile is a UI load's biggest removable cost and the only half of it that can be kept — running the chunks is what an addon *is*. Measured through the addon harness's `--together` load of the player's own 91 addons, three passes: ``` no cache 4,788 chunks compiled 3,324 ms cold 1,592 compiled, 3,196 from cache 1,482 ms warm 0 compiled, 4,788 from cache 471 ms (+24 ms of keys) ``` **About 940 ms off every login and every `/reload`**, and the first load after an addon changes pays only for the files that changed. `benilla-config/cache/lua/`, one file per chunk, keyed by a hash of the chunk's own bytes and its name. Both chunk paths go through it — `run_chunk_named` and the XML loader's own runner, which is how the biggest files in the corpus arrive (pfQuest's database is 33 MiB of `<Include>`d Lua). It cannot go wrong quietly. A key is a property of the bytes, so an edited file is a different entry and a stale one is never asked for; an entry that is truncated, corrupt or from another Lua build fails `lua_undump`'s header check and the source is parsed instead, so a bad entry costs the parse it meant to save. The dump keeps its debug info — `debugstack` is an addon-facing verb and a stripped chunk would lie about its own lines — which the test asserts by raising from a cached chunk and demanding its file and line back. The folder is capped at 512 MiB (one load's worth is 159 MiB) and emptied wholesale when it is over: a cache that needed an eviction policy would need bookkeeping that costs more than it saves. The first build of this looked like it worked and saved nothing, because `ahash::AHasher::default()` is seeded per process: every session hashed the same file to a different name, found nothing and rewrote the whole cache (4,776 entries, 477 MiB, a steady 1,592 misses a run). Fixed seeds, and the instrument that caught it — `benilla_ui::script::chunk_cache_line()`, hits and misses with the time each spent — stays, because "the cache is installed" and "the cache is working" turned out to be different claims. Verified: the harness's verdicts over all 91 addons are **identical** with the cache and with `WOW_LUA_CACHE=0`, the full gate chain is green, and the `WOW_DATA` failure sets are unchanged (benilla-app 270, benilla-formats 80). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
0fe0a99d8a |
layout: a size the author set is not a question for the solver
`GetWidth`/`GetHeight` answer against current layout, so they resolve first — and a resolve that follows any `CreateFrame` re-derives the whole layout graph, because minting a frame id drops the cached one (`Model::frame_id` → `touch_layout`). A UI load is nothing but creation interleaved with reads, so the player's interface paid **839 whole-roster derivations in one load, 2.84 ms each: 2.4 s**, against 650 `GetWidth` calls. `WOW_LAYOUT_DERIVE_TRACE` named the giver-up: every sampled one was `create_frame`. Most of those reads cannot learn anything from a resolve. An edge takes its anchor scan first and only then `combineEdge(center, opposite, size · scale)` (`resolve_left`, `resolve_right`), so an authored size IS the span unless the anchors pin **both** of that axis's edges. `LayoutInput::authored_span` is that rule, written beside the solver whose precedence it mirrors, and it refuses the two cases where only the solver knows: both edges anchored, and a clamped frame, whose rect the screen can squeeze. Measured on the player's own 91 addons through the harness's `--together` load, three passes each: ``` derives per load 839 → 235 preamble per load 2,414 → 560 ms frame:GetWidth 686 µs → 60 µs per call (1,241 ms → 118 ms over three loads) ``` **1.85 seconds off every login and every `/reload`**, and the steady state gains the same way: an addon reading a size it set in an `OnUpdate` no longer walks the roster for it. Proven as well as measured: six UI captures (character, bag, merchant, quest, quest log, tooltip) are **pixel-identical** before and after, MAE 0.000, max 0; the layout suite passes under `WOW_LAYOUT_VERIFY`, which re-derives every frame from scratch and demands the same rects; and `resolve_bench` now asserts both halves — an authored, singly-anchored span answers with no gate walk, while one whose anchors pin both edges still forces the resolve and reports the anchors' distance, not the authored number. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
c5911ccd6f |
verbs: an instrument that charges a load's time to the binding it went into
`WOW_UI_HANDLERS` names the handler and `WOW_UI_TICK` names the tick phase, and
between them they left the question this answers. A reload of the player's
interface spends **4.47 s inside handler bodies**, and single calls in there cost
hundreds of milliseconds — `pfBag:OnHide` 394 ms in ONE call — which is not Lua
arithmetic. A handler's self time includes the engine work it provokes, so that
time is ours: it is inside the verbs the handler called, and nothing charged it to
them.
`WOW_UI_VERBS=1` wraps every global and every widget method at the end of
`UiScript::new` — before FrameXML or any addon runs, so everything it finds is
ours — and times each call. Off, nothing is wrapped and nothing is paid.
It works offline, which matters more than the live case: the addon harness's
`--together` mode loads the player's whole folder into one VM in load order, which
is the closest thing to a login burst that needs no server. The first run over his
91 addons, three passes:
```
[ui-verbs] 519 verbs, 17,966,414 calls, 9,148 ms in all
4384 6393 685.8 us pcall (addon code, nested)
1498 16071556 0.1 us type (5.4M calls a load, Ace's)
1241 1950 636.2 us frame:GetWidth ← 650 calls a load at 0.6 ms each
548 33093 16.6 us CreateFrame
219 37524 5.8 us fontstring:SetPoint
95 99 960.3 us frame:GetHeight
67 123 545.9 us frame:GetLeft
17 3 5628.6 us frame:GetCenter
```
A geometry read costing 636 µs is the whole finding, and the fix is the next
commit.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
6affd61c35 |
doc: the model pane's skins link points at the struct that has them
`doc-links` was the last red gate, on one link dating from the fork's import (`81d96af3`): `script::types.rs`'s `QuadContent::ModelPane::skins` cited `widget::kinds::ModelState::skins`, a field that never existed on that struct. The three creature skin variations live on `script::CreatureModelView`, which is what a display id resolves to for a pane, so the link now goes there — by the public re-export rather than through the private `model` module. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
e95de4a5ad |
log: the session's warnings are 330 down to 129, and the 201 that went were ours
A full inventory of the player's 2026-10-18 session: **330 WARN lines, 67 distinct
shapes, and not one ERROR**. Read one at a time against the code and the
reference, most of what remained was the client complaining about itself. Three
causes, measured by replaying the new rules over that same log: **330 → 129**.
**A label is not a key (33 lines).** `text=` resolves as a GlobalStrings key and a
miss shows the raw value, which is faithful; the warning on a key-shaped miss is
ours. But a miss is already a string no locale defines, so a single all-caps word
is a label the author typed — `OK`, `UP`, `DOWN`, `MT`, `TPS`, `TOP`, `BOTTOM`,
`WIM`, `TWA`, 42 lines of them in one session, every one rendering exactly as the
reference renders it. The predicate now wants an underscore, which keeps the shape
that cannot be anything else: `ATSW_ALLREAGENTLISTFRAME_CHN`, an addon's own locale
row missing from the locale it shipped, which really does paint its key name on
screen. The shipped-XML sweep keeps its own broader rule, where `text="DELETE"`
must resolve and a collision is a bug either way; its comment now says so.
**Prose is not a missing file (12 lines, and phantom load failures).** A `.toc` is
read line by line, `#` opening a metadata line and everything else a filename, so
a wrapped `## Notes:` paragraph and the `--` section headers many addons write
become entries that no file answers. The reference opens them, fails, and says
nothing. We called each one a load failure, which put it in `/errors` and counted
it in the "N addon load failures" line at every login — ColorPickerPlus shipping
"two files it does not contain" that are really one two-line sentence. A line that
cannot name a file is now a debug line; one that looks like a file and is absent
stays a warning, because then the package really is short a piece (BigWigs'
`Raids\Other\Nerubian.lua`). The addon harness shares the predicate, so its
package audit stops making the same claim.
**One cause speaks once (168 lines → 9).** The stock inspect panel's sixteen slots
each name a frame `Blizzard_InspectUI` has never been asked to build, and
QuickHeal's one undefined template is inherited ten times — per UI load, so 78
lines a session for two causes. A document's log warnings now collapse by shape:
the first two keep their detail, the rest become one "N more like …" line, and
`/errors` still gets every one, because the panel is where detail belongs.
Three shapes were checked and deliberately left, because the client is right:
`SetScript("OnInputLanguageChanged")` must keep raising (an unfired script kind
raises by this project's own rule, and nothing can fire an IME edge here); the
`SetOwner: unknown anchor 'BOTTOMRIGHT'` line is TurtleWoW's own
`Turtle_ShopUI.lua:530`, where our mode-0 fallback already matches the reference
and the line is already once per session; and a `.toc`-listed `Bindings.xml` walks
as FrameXML with `<Binding>` an unknown frame type, which the reference logs too
(the bindings themselves load by their own path and work).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
3e616e5cb3 |
reload: one VM and no flush storm, 615 ms off every rebuild
Two costs read straight out of the player's own `WOW_UI_HANDLERS=1` session log (2026-10-18 20:33), where a `/reload` blocks the main thread for 7.84 s. **The saved-variables flush, 305 ms.** The reload's first act is to write every loaded addon's saved variables, one file each, and `write_atomic_bytes` ends each with `sync_all()` — a disk round trip per file, 60 of them, inside the one blocked frame. Measured on this project's own btrfs home: **327 ms flushed, 6 ms not**, which is the +27→332 ms the log spends between `ReloadUI — ending the UI session` and the last `wrote …/saved/*.lua`. `write_atomic_bytes_bulk` keeps the temp file and the rename (so a file is whole or absent) and gives up only a power loss in the seconds after; the reference writes these with plain buffered CRT writes and a `.bak` rotation whose result it discards, so this is still the stronger of the two. **A whole VM built and thrown away, 310 ms.** `end_ui_session` installs a fresh boot VM, which is right for the roots that stop there — the character screen runs on it — but `run_pending_reload` then calls `load_ingame_ui_on_world_entry`, which mints its own in `mint_entry_vm`. So every reload ran `UiScript::new`, `GlobalStrings.lua`, the emote table and the font registry twice and let the first one go without running a line: the log prints `GlobalStrings loaded` at +350 ms and again at +662 ms. `end_session` now takes what to leave behind, and the reload leaves the dying VM until the mint replaces it — safe because the two calls are one exclusive system with nothing in between, the old VM is still dropped before any file loads, and `fold_dying_vm_cvars` drains what it folds, so the mint's second fold of the same VM finds nothing. Only when the rebuild is certain to mint: with no UI wanted (a capture) the load returns early and the world would be left with no VM at all. `chunkbench` is the instrument the next item needs, and the one that priced this one's neighbours: it runs a file list three ways — a VM per file, one VM with the stock collector, one with it stopped — so parse, execute and the GC's share are separate numbers. Over his enabled addon set (1,471 files, 70 MiB) the parse is 703 ms against 74 ms to compile the same chunks from bytecode, which is what a bytecode cache would be worth; the collector is not the problem (829 ms against 731 ms stopped), and pfQuest's 23 database files are 417 ms and 110 MiB of heap. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
2a51c9817a |
frames: EnumerateFrames, the walk over every frame the client has
A real 1.12 global we did not have (`reference/1.12-globals.tsv`), and the natural way for an addon to ask the live UI which frames carry some script or property — the player's backlog item, raised because nothing answered. Read off `0x705f60`. The frame manager owns a Storm list (`[0xcf0bd8]+0xcc0`, link offset `0x304` written by its ctor at `0x7641b8`) and the `CSimpleFrame` ctor links every new frame through `0x7646a0`, which puts the newcomer in the sentinel's own slot `[0xcc4]` and the frame it displaced in the newcomer's `+0x308`. So the list is walked **newest to oldest**: with no frame the binding answers `[0xcc4]`, with one `[frame+0x308]`, and the first frame ever created holds the tagged sentinel `&head|1` there, which is the `test $0x1,%al` that ends the walk with nil. Three details that are not guesses: - The only type test is `lua_type(L, 1) == LUA_TTABLE` (`cmp $0x5`), so a number, a string or a boolean is not an error — it falls through to the head and starts the walk over. - A table is resolved through `T[0]` exactly as our `decode_id` does, and both rejections raise through `luaL_error` (`0x6f4940`, which never returns) with the client's own wording: `0x872f08` for a table with no object behind it, `0x872ecc` for one that is some other kind of object — a region, which is not in this list at all. - Regions and the walk's order aside, hidden and anonymous frames are in it: the link is the ctor's, not the draw list's. The arena grows the list it already implied. `Frame::insertion_seq` could not serve: that is the draw-order stamp, and a `Show` or a strata change restamps it (`resequence_to_tail`), where creation order never moves. So `created` is the handle list in creation order with `created_seq` the frame's index into it, and a step is one lookup. A destroyed frame's slot is kept, not compacted — its handle stops resolving and the walk skips it — because an index shift would break every other frame's `created_seq`, and nothing in play destroys a frame anyway, as nothing in 1.12 can. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
e527931885 |
worldmap: WORLD_MAP_UPDATE is signalled inline, so the map opens on the player's zone
The world-map flash, and it is one frame of the wrong paint rather than a wrong selection. `WorldMapFrame.xml`'s OnShow runs `SetMapToCurrentZone()`, and `WorldMapFrame_OnEvent` answers `WORLD_MAP_UPDATE` with `WorldMapFrame_Update()` **only** `if ( this:IsVisible() )`. Our `store_selection` pushed that event on the deferred lane, so it was heard a tick later — by which time the panel was already on screen, still carrying whatever tiles the last `WorldMapFrame_Update` left. Those are the World sheet's: `WorldMapFrame_OnLoad` runs its own `WorldMapFrame_Update` before the player's zone is known, and a nil `GetMapInfo()` takes FrameXML's own `-- Temporary Hack` branch, `mapFileName = "World"` (`WorldMapFrame.lua:47-55`), painting `Interface\WorldMap\World\World1..12` into the twelve detail tiles. Signalled inline instead, as the setter `0x4a67a0`'s own `SignalEvent` is, so the repaint happens inside the OnShow and the first frame drawn is the player's zone. Signalling moves out of `store_selection` — the writer cannot do it, only a caller holding the Lua state can — and each of the four call sites does it: `SetMapZoom`, `SetMapToCurrentZone`, `ProcessMapClick` (only when the click actually moved the selection, since the setter is the only signal site) and the engine-side world-enter sync. `worldmap_update_event_fires_on_next_tick` asserted the deferred behaviour deliberately, on the reason "a binding cannot re-enter dispatch". That was not so: `UpdateSpells` had already been doing exactly this with the same `fire_event_into`, with its own test pinning the synchronous dispatch. Replaced by one that asserts the inline fire and one that covers the two other legs. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
8e0d1bae8d |
lua: a numeral ends where its digits do, so 1then is two tokens
`AdvancedTradeSkillWindow2.lua:3878` is `if Item.ReplaceWords == 1then`, with no space, and it compiles on the reference. Its `read_numeral` (`0x6ffee0`) is 5.0's: digits, at most one `.` and its digits, then — given an `e`/`E` — the marker, an optional sign (the `0x2b`/`0x2d` test at `0x6fff04`, reached with the marker already saved) and its own digit loop. The only character-class test in the whole routine is the ctype digit mask at `0x6fff47`; past it the buffer is terminated and handed to `luaO_str2d` (`0x6f5900`). 5.1 instead ends with `while (isalnum(ls->current) || ls->current == '_') save_and_next(ls);`, which swallowed `1then` into the numeral and failed it — so we raised `malformed number` and lost a 4000-line file and every global it defines, and its own XML then raised four more times calling them. The addon was dead. The ninth hunk of the `lua-src` fork, which is where a grammar difference belongs: `BENILLA.md` is explicit that rewriting the chunk source means parsing Lua with a regex. A replacement, not a deletion — 5.1 had no exponent-digit loop of its own, so deleting the slurp alone would have cost `1e2`, which is what the new test caught. The same boundary is the deliberate half: 1.12 has **no hex literals**. The digit loop stops at the `x`, so the reference reads `0xFF` as `0` and the name `xFF` and cannot parse the statement around it. A comment- and string-stripping scan of the installed corpus finds hex in code in exactly two files, `pfQuest/toolbox/pngLua/deflate.lua` and `numberlua.lua`, which no `.toc` row reaches — zero loaded sites, the same standard the fork's three other deletions were held to. Our own shipped Lua and his saved variables have none either. Two of our own test chunks did, and are now decimal: no chunk of benilla's Lua may use a construct the client's parser rejects. Verified: the whole chain still loads with no failure, the `WOW_DATA` failure set is unchanged, and `AdvancedTradeSkillWindow2` now publishes `ATSW_OnLoad` and `ATSWTypeColor`. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
f185f3b549 |
camera: CameraZoomIn and CameraZoomOut, and a bare one is a yard
The first two rows of the reference's camera table `0x84f7a0` (`0x50b400`, `0x50b450`), and the one verb the chain's own barbershop, arena and guild-bank windows call that we did not publish — `Turtle_Barbershop.lua:211` zooms the world camera out two yards as its preview opens. Read off the binary rather than guessed, because the argument handling is not the one every sibling in this file uses: both thunks test is-number (`0x6f34d0`), take `tonumber` (`0x6f3620`) when it holds, and otherwise store `1.0f` outright (`0x50b42d`, `0x50b47d`). So `CameraZoomOut()` steps a yard out, where `SetView()` and `FlipCameraYaw()` return silently. A test pins that, since it is exactly what reading the callers would have got wrong. Deviation, named in the module: the reference turns the yards into a timed glide — divide by the rate at `[0xbe0fbc]+0x24`, scale, truncate (`0x40a2b0`) into a step count — and a call during a running glide adds into its pending-zoom field (`0x50fca9`). Our rig holds one target distance and glides to it every frame, so a second call re-aims the same target instead of queueing behind the first. The clamp is the rig's own, the one the wheel and the zoom keys already share. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
02265cc224 |
compat: nampower's equipment and item-cache getters, for the character panel
"then we gotta fix bettercharacterpanel, that one is driving me nuts."
BetterCharacterPanel is built on five nampower getters and had none of them.
`GetEquippedItem` was 8 raises in one session of his log; the other four failed
inside the addon's own `pcall`s, so its item-level column, quality colours,
missing-enchant warning and two-hand test were all dead at once. DoiteAuras
reads the same getter in six places.
The spec is the installed `nampower.dll` itself rather than inference, which is
what unblocked this — the earlier note here said to find the upstream spec
first, and the DLL *is* the spec:
- its usage strings give the signatures,
- "Slot must be a number between 1 and 19" gives the band, so the slots are
1.12's own `GetInventoryItemLink` numbering and `NampowerAPI.lua`'s
`slot - 1` in `IsItemInSlot` is that shim's bug, not the contract,
- its string pool gives the key names in order, and
- "Item not found in DBCache" is a raise, which is exactly why the addon
`pcall`s these three and calls `GetEquippedItem` bare.
`itemLevel` is the one field 1.12 cannot reach at all: `GetItemInfo`'s fourth
value is the *required* level, and later clients are the ones that insert
`itemLevel` there — so the panel's item-level column had no possible source
before this. The cache row now carries it, with `buyPrice` and `displayInfoID`
beside it, and an equipped item carries the raw descriptor numbers the tooltip's
resolved enchant lines could not answer (an id whose `SpellItemEnchantment` row
hides its name is still an id).
Fields left absent rather than guessed are listed in code and in
`docs/ADDON_COMPAT.md`; nothing installed reads any of them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
13989fb09b |
messageframe: a scroll that moves the view fires OnMessageScrollChanged
WIM sets this script on every window it builds — the history frame at load, then one per correspondent (six whisper windows in one session of his log) — and we refused the name by raise, which killed the rest of the builder function each time, not just the handler. It was on the deliberate "real 1.12 kinds nothing fires raise rather than silently accept" list. That rule is right and this was the wrong side of it: the slot now fires, from the ScrollingMessageFrame's own scroll verbs when a scroll actually moves the view, which is what a chat addon hangs its scroll-down arrow on. A scroll refused at either end moves nothing and says nothing; whether the reference fires one there is untraced. The raise list keeps its rule and its test — `OnInputLanguageChanged` and the movie frame's are still raises, and a name leaves that list only when something fires it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
15f069667e |
saved: a widget writes as its name, so a saved frame comes back a frame
Decursive raised `attempt to call method 'AddMessage' (a nil value)` on every
login (`Decursive.lua:1393`). It keeps its output window in its own saved table
and guards the reload with `if Dcr_Saved.Dcr_OutputWindow == nil or not ...`
(`:1245`) — which a non-nil *data* table passes.
Our writer walked the frame like any other table: his `Decursive.lua` came back
holding ChatFrame1's own fields (`buttonSide`, `channelList`, `editBox`, …),
twice the size of the real client's file and with no methods on it at all.
The reference tests the value's `T[0]` identity before its fields —
`lua_rawgeti(v, 0)` then `lua_touserdata`, `0x704613`-`0x704628`, reading the
lightuserdata `0x701bd0` wrote — and never walks a widget. A named one writes as
its **bare global name**, a Lua expression, so the reloaded chunk's own lookup
hands the live object back; an anonymous one as `nil --[[ unnamed object ]]`
(`0x872e0c`). His own real-client file is the proof:
["Dcr_OutputWindow"] = ChatFrame1,
Deviation: a name that is no Lua identifier writes as an anonymous object, where
the reference writes it raw and would leave a file that will not parse.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|
||
|
gromandClaude Opus 5
|
ed92574cea |
tooltip: SetOwner hides the plate, so a rebuild gives the addon its OnShow back
Reported from play with a video: hovering a bag item draws the enhanced tooltip — MasterTradeSkills' recipe block, AdvancedTradeSkillWindow2's Bags/Bank/Total — for about nine frames of sixty, then the plate falls back to the bare one and stays bare while the cursor sits still. The loop is stock. `ContainerFrameItemButton_OnUpdate` re-enters `..._OnEnter` every `TOOLTIP_UPDATE_TIME` (`ContainerFrame.lua:693` in his own patch-4.mpq; pfUI's bag slots are `ContainerFrameItemButtonTemplate`, so they run it too), and each turn is `SetOwner` + `SetBagItem` — a clear that wipes every line an addon appended. What puts them back in the reference is a hide/show pair, and we had neither half: the `SetOwner` binding hides the plate before it even reads the owner argument (`0x531179` zeroes the content gate at `+0xd0`, `0x531183` calls `[vtable+0x84]` = `0x530a60`, which is the `SetOwner(NULL)` core plus the base hide `0x76ad50`), and the content `Set*` re-shows it through the tooltip's own Show override (`0x530a80`). So every rebuild raises `OnHide` then `OnShow`. Ours cleared in place, and the base show early-outs on an already-visible frame (`0x76ae3b`), so the plate never hid and the `OnShow` never came. That is why exactly two addons lost their lines while others kept theirs: `LibItemTooltip-1.0`, the tooltip contract of MasterTradeSkills and Gromissingcrafts, hangs entirely on a scratch child frame of `GameTooltip` listening for `OnShow`, where pfUI's ItemID and sell-price lines hook the `Set*` methods themselves and so re-run on every refresh. The reference also clears twice per `SetOwner` (once inside the hide, once in the core) and so fires `OnTooltipCleared` twice; that falls out of calling the same hide, and the existing lifecycle test already allowed for it. Addresses read off the player's own `WoW.exe` with objdump, not from memory. The new test is the hover loop in miniature — three turns of `SetOwner` + a content `Set*` under a `LibItemTooltip`-shaped child probe — and fails on the old behaviour with one `OnShow` instead of three. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
0807144f38 |
modelframe: a unit pane holds the unit it was given
`SetUnit` resolved nothing: it stored the token and the extract re-read it every paint. unitscan's alert is the live consumer — it `SetUnit'target'`s the rare it just found by name, and the addon restores the player's own target on the very next line — so the card showed the rare for one frame and then went empty. It only looked right while our by-name targeting was leaving the target on the rare, which is the accident the previous commit removed. The reference resolves the token in `SetUnit`'s own rebuild worker (`0x505b50`, the one `RefreshUnit` also runs) and nothing re-resolves it per paint. Three things say so independently: FrameXML re-issues `SetUnit` from `PaperDollFrame_OnShow` and on `UNIT_MODEL_CHANGED`, which would be pointless against a live token; `RefreshUnit` exists at all; and pfUI detects a portrait change with *no event* by pinning a scratch pane and comparing `GetModel` against the shown one (`api/unitframes.lua:2069`), a test that says nothing unless a pane holds what it was given. `ModelState::unit_guid` is that pin, read through one helper (`Model::unit_pane_subject`) so the paint gate and the extract cannot disagree about whether a pane has a subject. A pane seated *before* its unit streamed in keeps re-reading its token — ours arrives asynchronously where the reference's is already in the object manager — and `RefreshUnit` re-pins, as its worker does. `modelframe.rs`'s own extract test passed only because its `SetUnit` ran with no target; its comment claimed the general rule, so it now states the real one and pins both halves. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn |
||
|
gromandClaude Opus 5
|
3cc71f4c7d |
scrollframe: the child's re-anchor is a SetPoint at the setter, not a per-frame override
`SetVerticalScroll`'s own note here already said it: `0x786db0` stores the offset,
**re-anchors the child (0x787100)** and fires its event, "all only on a change".
0x787100 passes the offsets to `SetPoint` — so the engine writes the child's
TOPLEFT once, when the scroll state changes. This applied it instead as a layout
override, recomputed every frame, which is a different thing: it made the
engine's anchor permanent and unoverridable.
That is why AI_VoiceOver's options could be opened and read but not scrolled.
AceGUI-3.0 scrolls by moving the child **itself** — `SetScroll` is
`content:ClearAllPoints()` then `content:SetPoint("TOPLEFT", 0, offset)`
(AceGUIContainer-ScrollFrame.lua:62-77) and it never calls `SetVerticalScroll` at
all — so the override put the content back at the top on every frame. Measured
through the harness: `SetScroll(1000)` computes offset 170 (content 549 − view
379) and the content's top now moves 563 → 733 where before it did not move.
`SetScrollChild` and both offset setters now write the merged TOPLEFT through
`layout_methods::write_anchor`, which is the commit `SetPoint` itself uses, so
the merge rule is shared rather than reimplemented and a child authored
TOPLEFT/TOPRIGHT keeps the TOPRIGHT that gives it its width. The whole override
path goes away with it: no third element on the layout plan, no override in the
node hash, no special case in the edge walk or in `retarget`, and
`LayoutSolver::set_frame_anchored` is gone.
One behaviour genuinely changes and the test says so now: `SetScrollChild(nil)`
no longer restores the child's authored anchor, because the authored TOPLEFT was
replaced the first time the frame took the child. Nothing was ever saved to
restore — that expectation was an artefact of the override model, not a trace.
The child keeps where the last scroll left it, and its own `SetPoint` wins from
there.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
|