The tick ran the OnUpdate walk without resolving the layout around it, so a resize made in an event handler fired `OnSizeChanged` after the walk, and some walk slots resolved on their own to compensate. Separately, `OnSizeChanged` gated on 2⁻²² in benilla's units, and an emptied FontString kept its old measured box in the incremental layout.
In the reference, the pump drains the layout queue before the walk and again after it, each drain firing `OnSizeChanged`, then re-picks the hover. ApplyRect compares sizes against 2⁻²² in its own units, where the screen diagonal is 1.
The tick now drains before and after the walk and drops the stand-in resolves. Size gates scale 2⁻²² by the layout unit, and a FontString's text going empty or back names it on the layout when its node changes.
Tests pin each.
Fixes#912Fixes#741Fixes#914Fixes#911
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Any write that changed the layout graph's roster (a frame or region created, freed, re-owned or re-pointed) made the next resolve derive the whole graph: about 1.35 ms for one `CreateTexture` or `CreateFrame` on the stock UI, against 0.05 ms for a no-op.
A write now names the nodes it touched, and the resolve re-derives just those in the cached graph: a joining node is added with its edges, a leaving one is removed and its dependents seeded, a staying one has its edges and inputs re-synced. The full derivation stays the fallback. A region's owner and an art change are now layout inputs that name their region. Verify now also compares the graph itself.
A randomized test checks rects and graph against a scratch solve after every batch of creates, frees, re-owns and re-anchors. `CreateTexture` now costs 0.14 ms, `CreateFrame` 0.15 ms, a backdrop rebuild 0.22 ms.
Fixes#903
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Hidden chat frames kept fading their lines, a scroll frame recomputed its range on every read and fired `OnScrollRangeChanged` on every `UpdateScrollChildRect`, hidden or not, and `OnUpdateModel` fired for panes with no size or off screen.
In the reference, each of these runs in the frame's own update slot, only while shown: message frames trim and fade there; a scroll frame recomputes a range its writers marked, stores it and notifies only on a real move. The model paint fires only for a resolved rect whose screen-clamped viewport has size.
Message frames and scroll frames are now walk kinds, the range is stored and marked, and the paint applies the rect and viewport gates. benilla's options scroll measures its body's own overhang.
Tests pin the hidden fades, the stored range and its notify, and the paint gates.
Fixes#182Fixes#175Fixes#892
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A backdrop was data drawn in the frame's own slot, so `GetRegions` and `GetNumRegions` never saw its pieces, and addons that skin by walking a frame's regions missed them.
In the reference, installing a backdrop creates nine Texture regions on the frame: the background on BACKGROUND, then eight border pieces on BORDER, linked into its region list in that order. A resize re-runs their texcoords and colours, and the colour setters tint the background or the border alone.
The pieces are now real regions, drawn through the region path, so edges draw over another frame's BACKGROUND at the same level. `<Layers>` and `<Backdrop>` load in document order. A region leaving its parent empties any Button, CheckButton, StatusBar, Slider or ColorSelect slot holding it, and a dead layout node now seeds only its dependents.
Tests pin the order, colours, resize, teardown and stock quads.
Fixes#883
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The UI scale was folded into the pixel seam, with the root and `UIParent` at scale 1. Frames outside `UIParent` (dropdown menus, need/greed rolls, the world map) shrank with it, `UIParent:GetScale()` answered 1, and `GetCursorPosition()` was off by the scale.
In 1.12.1 the scale is `UIParent`'s own frame scale (`0x494550` → `0x76ac10`, floored at 0.64) over a 768-unit root. `GetScreenWidth`/`GetScreenHeight` divide by it (`0x494590`), and `GetCursorPosition` (`0x48b820`) does not.
Now the seam is `height/768`, and the scale is applied to `UIParent` at the load's end and on a resize. The nameplates' compensation is gone. Backdrops, hit insets and minimap pings now follow a frame's own scale (`0x76b580`, `0x77e8d0`); tooltips at the cursor and `OnCursorChanged` divide by it (`0x530b20`, `0x77db23`). Tests cover each, and a stock-UI comparison shows nothing inside `UIParent` moves.
Fixes#172🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Code health in `benilla-ui`, found while its comments were rewritten. No behaviour a player can see changes.
- **Unused globals and queues removed.**
- The Rust `ChatFrame_OpenChat` and `ChatFrame_SendTell` queued text that nothing drained. The stock `ChatFrame.lua` defines both globals and replaces them when it loads, as the reference's FrameXML does.
- `BenillaGetItemStats` had no Lua caller. Its test of the item cache's miss, ask and push now runs through `GetItemInfo`, which shares that path.
- `UiScript::queue_social_request` had no caller.
- **One true/nil helper.** Seven modules each carried their own copy of `binding_abi::flag`, the reference's `1`-or-nil boolean. They call it now.
- **Three tests test their claim.**
- `click_focuses_regardless_of_autofocus_…` built both boxes with autoFocus on. One is now off.
- `level_shift_saturates_at_zero` never saturated. Its child now drops past 0.
- The layout profiler's `plan` bucket was never set, so it always printed 0. It is gone.
- **Smaller fixes.** An empty `if buyback {}` branch is gone, and a doc comment no longer sits on a `use` line instead of on `install_justify`.
Checked with `cargo clippy -p benilla-ui --all-targets -D warnings` and `cargo test -p benilla-ui` (1513 passed) before the land gates.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Part of #41: the comments of `benilla-ui` (the Lua UI engine, its FrameXML loader and its widgets) are rewritten to the rule in `docs/METHOD.md`, which is what the code does and the reference fact behind it, in a line. 43,140 comment lines become 15,161.
- **The mechanical strip.** The citations that were only a private record number go first, and each edit is shown to be a pure removal.
- **Wrong facts corrected.** The rewrite corrects the comments that stated wrong facts. Among them:
- `0x64a480` is the case-sensitive compare (a byte `repz cmpsb`), and `0x64a4c0` the case-insensitive one.
- A frame's id sits at `+0xb0`, not `+0xb4`. `EnableKeyboard`/`IsKeyboardEnabled` are `0x776ec0`/`0x776f90`.
- A region's unpinned edge has no fallback to its owner's rect: the owner supplies only the scale.
- `$parent` expands to the nearest named ancestor, or "Top" when none is named (`0x76c5b0`), not simply to the parent.
- A StatusBar fills from its internal normalised fraction, not from `GetValue()`.
- The addresses of `StartDuel` and `StartDuelUnit` were swapped.
- Many "untraced" notes now give the traced reference fact: the drag threshold, the per-grip sizing edges, the empty-`autoFocus` skip, and `SetTexCoordModifiesRect`'s default.
- Stock FrameXML line citations are checked against the stock 1.12 files and corrected. `ref-` file names become the stock file names.
- **Deviation labels.** "Deviation:" now marks only a choice with a reason. A surface that exists only for the retired Era addon target says so plainly: it is not 1.12.
- **Verification.** Every file strips to identical code, and `cargo fmt --check` is clean. Nine second readers compared every file old against new, restored the facts the rewrite had dropped, and checked each new citation against the code, the stock files, vmangos and the reference.
Behaviour differences the rewrite surfaced are filed as their own issues (#131–#142), and more are being validated. The comments here state what the code does. String literals are untouched; they are #92.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Comments that pointed into the maintainer's private reverse-engineering notes (a note by name, a
path, a section or finding id, the analysis's own name for a client function, a process word) now
carry the fact itself, cited by the reference client's address, by vmangos file and line, or as
"the reference". Nobody outside could follow those pointers.
Comment text only, plus the test-assert and log strings that carried the same pointers. No claim
changes: a comment the notes have since corrected keeps its wording here, and correcting it is
separate work.
The two reference generators, `scripts/gen-reference-globals.py` and
`scripts/gen-reference-verb-events.py`, named the private notes' layout to find their inputs. They
now take the input files as arguments (`--capture`, `--framexml`, `--names`, `--extents`,
`--disasm`). Run on the same inputs, both write the committed tables row for row; only the two
header lines that named the old environment variable change.
Checked per file:
- the code is identical: each file with its comments stripped compares equal before and after; a
changed string literal is compared under a mask that proves only literal text moved and every
literal kept its line count;
- no client address the old text cited was dropped;
- a second reviewer checked every added address against its derivation;
- a detector for the notes' names, paths, ids, function names and process words finds none left;
- `scripts/gates.sh` is green.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- the mail window's send tab takes typing again: the subject, body and money boxes were taking no input while the recipient box swallowed every key, and a sent letter clears its subject and item icon instead of putting them straight back (2145)
- seven events fire the way the reference does: UNIT_DYNAMIC_FLAGS fires at all, swapping two identical stacks between bank slots fires PLAYERBANKSLOTS_CHANGED twice, the pet sheet's training points repaint, and ITEM_LOCK_CHANGED loses its invented arguments (2140)
- nine widget methods 1.12 addons probe for are in: ShaguTweaks stops recolouring textures the reference leaves alone, pfUI and _Nameplates stop raising on a tooltip's GetAnchorType, and 63 AceEvent addons get RegisterAllEvents (2142)
- six post-1.12 methods leave the widget surface (SetSize, SetShown, HasFocus, SetFormattedText and two tooltip verbs), so an addon reaching for a later client's API finds what a real 1.12 client gives it (2142, 2146)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
- a slider with a value step snaps to it: a stepped drag fires once per row instead of once per pixel, a scrollbar cannot rest between two rows, and one wheel notch on the trainer list reaches the bottom (2133)
- setting a step on a live slider moves its value onto the new lattice and runs its handler, which an addon on the real client can count on (2143)
- a button pushed from script releases on the next mouse-up, so Tablet-2.0's rows stop staying pressed forever; a release off the button's rect un-presses it, and a locked state holds (2134)
- the mail body and the GM ticket box scroll as you type past their bottom, and the Use UI Scale checkbox reads a registered setting (2141)
- an engine warning, a call that was accepted and did nothing, reaches the /errors window in grey instead of one terminal line; a text row trimmed mid-character no longer kills the window's repaint (2135, 2138)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
- the logon address is the reference's own `realmList` CVar — name, `host[:port]` form and the "Address of realm list server" prompt read straight out of `WoW.exe`'s string table — persisted in `benilla-config/config.toml` and edited on the login screen: a Realmlist button in the slot `AccountLogin.xml` already reserves under the Quit stack, the current address in grey beneath it, opening the reference's own `hasEditBox` glue dialog; before this the address was `$WOW_HOST` read once at process start with no way to change it; one text field and no saved list (every client at this scale ships one field), a pasted `SET realmlist "…"` line is accepted, and the default is `localhost` because the retail host has not resolved since 2019
- the README's "Not built yet" becomes "Where it's going": what done means, no dates, and one list in probable order — the remaining basics (battlegrounds, guild charters, the stable master, meeting stones, GM tickets, master loot, cinematics), addons / options / performance as ongoing, the no-brainer VanillaFixes- and SuperWoW-class fixes, playable downloads with Linux first; not planned: other expansions, Warden
- a UI region's span comes from its content and the owner-edge fallback is deleted: the reference's resolver core has no parent pointer at all, a FontString floors at one unit on both axes, a texture takes its texel span, and a region resolves from its own anchors and span or not at all — a live census over 8772 rects: 0 stopped resolving, 923 FontStrings shrank (842 to exactly 1×1 — empty labels that had been the whole screen, or an inverted rect), 0 textures moved; and `Region:GetWidth`/`GetHeight` are the same virtual getters the resolver calls (authored value wins per axis, natural width, wrapped height, texel span for art), so a script reads the number its rect was built from
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- a unit's pick geometry is every model chained to it — body, held items, mount — one candidate per root resolving to the same unit, which is how the reference registers its pick (it walks the attachment tree and registers each model); Naxxramas' floating weapons could only be selected by their name plate before (no body, no candidate), and every armed unit gains its gear as clickable surface (our own body 12 → 21 tested parts); the broad phase stays the body's sphere, so a very long polearm on a small body is the named residual
- the selection ring's writer has two branches and we had transcribed one: a model whose stand box is degenerate (zero extent in x and y) rings at the literal 1.2 × scale — 2.7 yd under a trigger creature — instead of `sqrt(0.5·sqrt(dx²+dy²))` = 0 clamped to an invented 0.05 floor; both invented fallbacks are gone
- an HTML book image with no width/height is its art's texel square (one texel = one FrameXML unit, bit-for-bit the `<AbsDimension>` converter), not the rest of the page — *A Treatise on Military Ranks* drew its crest twice too big and stretched to the page's aspect; a texture-size oracle answers off the same decoder the renderer draws with, and an unfloated height-less `<IMG>` reserves its art's height in the flow; 797 shipped textures swept, exactly two zero-size axes move and neither is visible
- the dev debug panel killed the client on open with "Attachments have differing sample counts": Bevy's 2-D depth texture is keyed on the render target alone, so two Camera2ds on one window at different MSAA counts cannot both be valid, and the egui overlay camera had silently kept the default 4× after the player-UI camera went to Off; it now says `Msaa::Off`, and a preflight check reports any two 2-D cameras that disagree on a shared target, every run
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- a tooltip change costs ~1.3 ms — five whole-interface recomputes, each working as designed: the quad re-convert at 474 µs, the Lua tick at 287, the layout resolve at 200, bevy's `Assets<Mesh>` tax at ~270 and the mesh rebuild at 146 — pinned with new hover probes (a jitter dial, a `WOW_PROBE_HOVER_DUTY` duty cycle so run drift cancels, an `AssetChanged` tax meter), and the layout preamble's two whole-roster scans become lists maintained at the writer (52–83 µs → 1–11)
- the splice compared this frame's entries against last frame's by index, so one inserted highlight shifted the 160 entries behind it and read as "everything changed"; it now aligns on `z` — an identity, not a position — and a `Text` entry that writes only quads became spliceable, so a tooltip change re-converts 22 entries instead of 520 (`convert` +439 µs → +107), with `WOW_UI_SPLICE_VERIFY` running the full conversion behind every spliced frame: 0 mismatches over 1,078
- a spell-less macro is a usable action button, not the 0.4 grey — the reference's usable compute has a dedicated leg answering usable for the bound-spell zero, cross-checked at the bytes in the reference's disassembly — and the button's Name line is written unconditionally from `GetActionText` as the reference does, so the macro's name sits under its icon and a `/cast` of a spell the character does not know now greys, which is also 1.12's behaviour (B340)
- a capture bounds its run in wall-clock seconds, because a frame cap is no cap at the ~1 fps a starved `nextDrawable` leaves — 1800 frames is thirty minutes there — and the teardown backstop runs on every run as a liveness guarantee, carrying the exit code the run asked for; the Multisampling row reads Off / 2x / 4x, since "1x" contradicts the no-multisampling byte both of the reference's enumerators normalise to and the bit-depth prefix was identical on every row
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- the glue scene widens only as far as its art goes: none of the seven login dioramas is drawn wider than about 3:2 — measured per texel by `benilla-extract glueextent`, which rasterises every batch into the camera's own tan-space, painting where the reference's alpha-key test passes — so the frame zooms down to the reference's own 16:9 opening (never tighter) and pillarboxes past the aspect the art cannot fill, instead of showing the clear colour past the backdrop's slanted edges; 4:3 stays byte-identical
- the character screen opens on whoever you last logged in as: the reference's `lastCharacterIndex` CVar, written at Enter World and only there — clicks and arrow keys never touch it, which is what makes the value mean "last logged in" rather than "last clicked" — restored on every character-list rebuild and clamped into range
- spam-hovering tooltips stops costing 2 ms: two write sites gave up naming their node and re-derived the whole layout graph (1,812 µs against the ledger's 44, on 919 of 1,262 resolves) — the wrap pin now touches the one region it wrote, and re-pointing a tooltip at the button under the cursor retargets the node's edges in place instead of throwing the cached graph away
- two corrections off the project's first GPU-bound report (a Steam Deck at 94% GPU busy on 8.4 W of a 15 W budget, the memory-stall signature): `farclip` ships the reference's registered 350 instead of the clamp's 777 — the value also sizes terrain residency, so 777 streamed, held and drew 4.8× the area — and BLP textures upload WoW's own DXT blocks verbatim the way the reference does, instead of decoding to RGBA8 at 8× the bytes of DXT1; the terrain slider still reaches 777, and the lanes that read texels on the CPU (terrain arrays, the character-skin compositor) stay decoded and say why
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- the mix is bounded the way the reference's is: FMOD gave 1.12 exactly twelve software channels and nothing else limits a priest buffing five people — ported as a 12-voice ceiling (a probe capture read 42 live voices at +13 dBFS, the limiter pumping 13 dB to hold it), plus same-kit copies started in one frame capped at two, since five sample-aligned Fortitudes are one sound +14 dB; `$WOW_SOUND_PROBE` records the mix before and after the limiter with a timeline, F9 stamps the moment
- the raid tab draws its window art again — its four corner textures were written with single backslashes, which Lua drops silently, and a sweep now reads every archive path a shipped Lua chunk names; a drag whose release never arrives (pointer left the window, a second press) ends instead of gluing the frame to the cursor for the rest of the session, and `WOW_PROBE_DRAG` drives any UI drag headlessly in a live client
- the item tooltip fits its set bonuses again: an emptied line kept its last measured height forever (a pooled cell with no way to say "empty"), and the two blank rows above the set block are the reference's own literal (a space and a newline) — a whole row each, not the empty string
- an equipped BoE or BoP piece reads Soulbound: the bind line reads the item's own flags through the client's bound predicate instead of the template's bonding, one predicate shared with the enchant cursor's bind confirm
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- a text write names its own label, so the per-frame measure sweep drains a short ledger instead of hashing every font string twice — ~0.3 ms a frame at the Stormwind gates, gone
- the idle interface tick stops re-writing settled state: buff durations and the weapon-enchant frame write only on change, and the fade and unit-popup drivers park until woken
- the interface extract skips a hidden region before cloning its row, not after
- WOW_UPLOAD_BUDGET=<MB> brackets the arrival burst's upload pacing — a lever for measuring world entry, not a fix yet
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- the interface's layout graph is maintained as anchors change instead of rediscovered from scratch: 1.48 ms per mutated frame becomes 34 µs, and end to end a running cast now costs +0.14 ms where it cost +4.33
- the autocast shine draws its authored population — 1200 particles through the star's own projection, spinning as the model says
- the shine is clipped to its own widget: the model frame installs its rect as the viewport, so nothing it draws leaves the button
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- the pet bar's autocast shine cost +4.25 ms a frame for one button: its 92 script-side SetPoints forced two whole-roster anchor solves every frame — the shine now draws from a native lane, re-measured at +0.03 ms with what is drawn bit-faithful to before
- an NPC that spawned under its own floor — a cellar wolf, a pet at login — now measures its ground clamp from the server's seat instead of its own last answer, and climbs out
- a running cast cost ~4 ms a frame in layout gate walks; a mutated frame now takes one walk, not three, and the settled interface is pinned at zero
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- a single animating entry — the resting blink, a sweeping cooldown — now re-converts alone and stitches into last frame's interface quads instead of dragging the whole batch through conversion, and the staleness sweep stops allocating
- two physics levers measured and flipped: entity clusters and the pre-step transform propagation both default off, −0.28 and −0.40 ms at the Stormwind gates
- the per-frame hygiene basket: env lookups cached once, both mouseover walks filtered by archetype instead of touching every streamed object, ground-shade cards resolve to their nearest shaded root, the nameplate steady frame stops allocating and stops dirtying transforms
- new meters for what remains: an animation memo-sizing counter and a lever that prices the physics stack's rate
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- UI layout solves what moved, not what exists — hovering a full bag no longer re-solves the whole interface
- world albedo textures stop keeping a second copy of themselves in RAM
- collision gets a coarse grid for oversized slabs (32x fewer triangles tested), idle units stop re-asking the ground every frame, and anchors spawn on first consumer instead of by the thousand
- aura and loot watchers scan when the store changes rather than on the clock, and static-transform tracking and plate reaping join the sweep
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- SetBinding accepts the mouse wheel — the refusal was ours, not the client's — so wheel-bound action buttons press and release like the real thing
- one addon load law: a single arbiter for "can this addon load, and why not", with the out-of-date version gate live
- one AddOns screen — the character-select list, wearing the authored reference layout: roster dropdown, per-character tri-state checkboxes, force-load; the ESC-menu panel is retired, since 1.12's own menu has no AddOns row
- the glyph atlas follows the window scale, so maximised text stops washing out and a title stops measuring wider than it draws; an unpositioned frame positions nothing, instead of landing at the screen origin
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- GetStringWidth answers the moment the text is set — the client's width law is a per-glyph table, no shaping pass — so Bagnon's character dropdown stops collapsing to 40 pixels
- a chat channel keeps its number when another channel leaves: Trade stays /2 through a teleport
- the login splash, the zone music and the minimap label wait for the world under the player instead of announcing whatever tile happened to be loaded
- ending a session empties the chat window and the input box's history
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Lua 5.0's generic-for and nested long strings are restored in a forked VM — 60 corpus addons stopped dying on one deleted opcode
- a day of demand-driven carving takes the 218-addon corpus from 24 addons surviving a session start to 114, with 144 loading without a single error
- Bagnon draws a real bag and answers real gestures — hover tooltip, pickup, use, drag — because our frames wear the reference's names, and the reference's own ContainerFrame.lua runs off the player's install instead of being transcribed
- the physics broad phase leaves: nothing in the client reads a contact, and it was measured at up to 41% of main-thread samples in real sessions
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- benilla loads real 1.12 addons from Interface/AddOns — TOC, XML and Lua through one loader, with per-addon saved variables, Bindings.xml and the lifecycle events addons expect
- both AddOns screens are in: the character-select list with per-addon enable state, and the ESC menu's in-game list
- the API surface is measured, not guessed — a corpus harness runs hundreds of real 1.12 addons and ranks exactly which globals, templates and verbs they demand
- the Lua dialect is the client's own, read off the binary — 5.0 semantics with the 1.12 additions, and UI source files are bytes, not UTF-8
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- the pet gets its own unit frame beside yours, and an Attack order no longer latches as the standing command
- five rooms that always read bone dry hold their water now, the Deeprun Tram's flooded stretches among them
- macros fire through the client's own chat door, and the macro window fits its tabs and scrolls its icon chooser
- the options window grows a Nameplates page and its first dropdown, and the name-display switches apply live in the world
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- one era-style Options window: window chrome, category list, search box and its own ESC rung, with the game menu reshaped to match
- settings are real CVars now: GetCVar/SetCVar reach Lua, and your changes persist as a diff in benilla/config.toml
- all local state lives in one visible benilla/ folder beside your Data/ — nothing hides in platform config dirs
- UI text stays crisp at any scale and survives a mid-session resolution change: glyphs measure in the environment they raster in
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- The login screen stops paying for a world nobody is in. It had been
streaming and simulating all of Elwynn Forest behind a static image — the
client began life as a terrain viewer and the streamers never had an off
switch. The world now boots when you enter it, aimed at the map your
character is actually on, and logging out releases it. Login frame cost:
p50 18.6 to 10.3 cpu ms, worst paced frame 308 to 20 ms; entering
Stratholme is a covered 1.2 s load.
- More login-screen honesty: the fullscreen glue image was re-uploaded to
the GPU every frame — 22 MB a frame for a still picture — and the
braziers' fire had been invisible since the 25th, simulating perfectly
into a camera that is off, because two booths had computed the same render
layer. Both fixed; the flames are back.
- Indoors got truer: the outdoor world draws only through the portal windows
actually in frame, as the original does; the painted-skybox gate settled
on the reference's real rule — any sky-flagged room the portal flood
reaches; and interior objects — gates, hanging signs, chains — take their
room's baked light, anchored at their box centre instead of an artist
pivot that often sits under the floor.
- Hovering with a tooltip up takes its full fix: a tooltip content change
used to run the whole-UI layout solve twice, and it runs once now, on top
of the morning's fix that made an unchanged tooltip cost nothing at all.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- The UI stopped paying to discover that nothing changed. Layout used to hash
~1,900 frames' worth of inputs every frame just to decide "quiet"; it now
tracks a mutation epoch and answers with one integer compare, and a settled
frame reuses last frame's rendered quads instead of re-rasterizing every
glyph. ~0.9 ms off every idle frame — with a verify mode, forced on across
all 489 UI tests, that re-runs every skipped resolve in full and asserts it
would have produced identical results.
- Across this run of snapshots, the one number everything is graded by: CPU
work per frame at the crowded dungeon spot went from ~29 ms to ~16.5 ms.
- A lighting question got the full treatment and an honest answer. Units
briefly took a different day/night lighting law; it read wrong on screen the
same evening, and a build-by-build hunt a full week back confirmed the
world's look never regressed in the first place — the change is reverted,
nothing about the look changes in this snapshot, and the question stays
parked until the one remaining unknown in the reference client is pinned.
- The character pane's model booth grew a headless eye — a dump of exactly
what it draws, no live session needed — built for that hunt, kept for the
next one.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Other players walk the world. A watched player's motion between server updates now runs through
the same grounded movement resolve as your own character — same capsule, same colliders, same
step-up and ground election — so they follow terrain every frame instead of every packet, and
stop at a wall instead of walking into it and being snapped back. Their jumps are swept against
the world too, and a standstill jump steered mid-air reports that steer, the one mid-air press
the reference client broadcasts rather than defers.
- The move relay rebuilt: relayed moves ride a per-unit chain on real time, a due move never jumps
ahead of its unit's queue, and SMSG_COMPRESSED_MOVES — the batched carrier the server sends under
load — is decoded rather than skipped. Our own send path now matches the reference client's:
facing streams while moving, Nagle is off so each packet is one write, and the strafe axis nets so
left and right can never ride the same packet.
- NPC and quest text speaks the whole $-macro grammar, carved from the reference bytes — your name,
class and race, the gendered forms, and the server's live world states — and the same
substitution feeds gossip, quest text, the quest log and unit frames. Greetings are a weighted
draw, rolled per NPC per open, and a quest giver with no quest to give is no longer clickable,
where it used to open a dialog holding nothing but a placeholder greeting.
- Spell tooltips finish the reference's line law: cast time and cooldown behind the real passive
gate, the required tool or form, the reagents block, the chance-to-X line, a multi-bit
equipped-item requirement named properly instead of by raw bits, and a buff's dispel class.
- The UI layout engine got a dense, reusable anchor solver, and a change gate that skips the resolve
entirely on frames where no layout input moved — in practice, almost every idle frame. Pets are
named by their pet number rather than a creature-template entry, so the client asks the query the
server can actually answer. And the asset extractor grew a stray-geometry census for tracking
down geometry that renders where it should not.
This snapshot replaces the six dated `sync` commits that stood here before it, whose subjects said
nothing but the date. If you cloned before now: `git fetch && git reset --hard origin/main`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
