This checkout had no git history at all, so this one commit carries the whole
of the Everwood work as a single import on top of chairmanpow/benilla-octo's
`main` (706cd71). The two trees have genuinely diverged — each has files the
other does not — so this is a fork point, not a catch-up.
Today's work, which is the reason a baseline exists at all:
- Booth previews no longer cast world shadows (`NO_BOOTH_SHADOW`). Portrait
characters were entering the sun's shadow cascades, so every skinned-rig
shadow specialization compiled live mid-frame; the hitch detector logged
"main thread blocked 1316 ms" beside a `compiled LIVE` warning. This was
the Fashion/Tmog stutter.
- Per-pane dress-up: each `DressUpModel` pane assembles and stages its own
tried-on look instead of every pane mirroring the live player.
- Item template cache is append-only JSONL written on the IO pool, replacing
an 11.6 MB whole-map JSON rewrite on the main thread every 10 s.
- `WOW_SYSTEM_TIMING` rebuilt per-thread: it took 2 mutexes per system enter
and 4 per exit, costing 15 ms a frame (590 fps -> 60) and corrupting the
measurements it was taken to produce.
- The visibility-churn sweep moved to its own `WOW_FLASH_CHURN` switch, so a
frame-cost reading no longer pays 2.7 ms a frame for it.
Local state (`benilla-config/`, `logs/`, `target/`) is excluded by the
existing .gitignore; the account credential it holds is not in this commit.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
Same fork policy: upstream structure wins, local-unique re-instated.
Cargo.toml unions both measured dev-profile philosophies (codegen-units=1 +
lto=off from samwhosung, debug=line-tables-only + split-debuginfo from the
fork's disk measurements).
glue_booth.rs takes samwhosung's typed GlueScene (token/of_race/from_model_path)
and re-adds the fork's TurtleWoW races (Goblin=9, BloodElf=10) as variants;
their glue fog stays on Human's row, exactly as the fork's scene_fog fallback did.
ui script takes samwhosung's mouseover/hyperlink rework (world_mouseover,
minimap_blip, hyperlink.rs) over the fork's inline hover_link/span_at/pointer
path — the rework covers the same OnHyperlinkEnter/Leave behaviour.
game_plugins.rs keeps the fork's re-measured ceilings (466 / 5,509); mail.rs
keeps the explicit 13-tuple with a float daysLeft; MAP.md takes the superset.
Fork policy (docs/METHOD.md): upstream structure wins; local-unique work
re-applied on top. This commit is the upstream half of the reconciliation.
Upstream is the same program at a rewritten history (884dff77 == 4e0e89cb
tree), so the merge base was replaced to keep a real three-way diff. The
large conflicts are upstream moving first: the UI manifest (benilla.toc ->
the binary LAYER_MANIFEST/layer.toc split), the CVar table (cvars.rs ->
cvars/{mod,table}.rs, +24 video rows), the editbox flush model, the
host-side GetTime, the Options* -> BenillaOptions* rename, expanded
ui_unit.rs, and a rebuilt volumetric_light.rs.
Local-unique pieces re-instated on the upstream structure:
- OptionsFrame.xml: the 1.12 OptionsFrameSliders table restored under the
Benilla rename, with the TurtleWoW deviation note (patch-4 replaces
OptionsFrame.lua).
- layer.toc: CharacterFrameAdapters.xml (the sixth TurtleWoW character tab).
- ui script: wipe/sizeof, BenillaGetMerchantItemStats, and the attach/mount
teardown re-walk all survive.
- game_plugins.rs keeps UPDATE_ACTIONABLE_CEILING=5_509 (re-measure pending).
The LFT/Turtle_TransmogUI rows benilla.toc named off the player's install
now come from the chain's own FrameXML.toc via the new core loader.
In 1.12.1, logging out or a refused character login returns to character select on the same world connection and re-lists there (`0x5aaeb0` → `0x401ee0`, then `0x46b880` → `CMSG_CHAR_ENUM`). There's no realmd logon and no new `CMSG_AUTH_SESSION`. Enter World closes the realmd link (`0x46b70c` → `0x5b3320`), and only the glue's own re-dials reopen it.
- benilla used to drop the socket and log on again from cached credentials. Now the world session splits and rejoins on one connection, and the header cipher carries its state across.
- The realmd link closes at the pick.
- A leftover packet at character select is skipped rather than fatal.
- The write thread applies a pending writer swap before any command.
- A kept-link teardown ends the character without a "server down".
Fake-server tests pin one logon, one auth and one more enumerate. The logout smoke fails without a same-link relist.
Fixes#984Fixes#983🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A build cache serves a dependency rebuild after a branch switch or a `git clean` instead of
recompiling it, which is most of what the in-loop verify spends. It stays a convenience and
never a requirement: with nothing on `PATH` nothing is exported and a clone needs only stable
Rust and a C compiler, exactly as before.
`check.sh`, `gates.sh` and `smoke.sh` source it, so the loops pick it up without being asked
and a machine that would rather not have it sets `BENILLA_NO_SCCACHE=1`. An existing
`RUSTC_WRAPPER` wins. Only rustc is wrapped — the vendored Lua build wants its own
`CC="sccache cc"`, so it is not cached by this.
The core now loads the chain's `Interface\FrameXML\Bindings.xml` after the `FrameXML.toc` rows, as the reference's UI load does. The load checks the file at `0x48fff9` and loads it through the addons' loader `0x4b6f70` at `0x490018`. The default keys come from the chain's `WTF\DefaultBindings.wtf` (`0x4b62b0`). Every command now runs the stock file's own body, and the registry compiled into benilla, with its 91 retyped bodies, is deleted.
- **The loader's rules (`0x4b6f70`).** It reads the root's children only and skips empty, `debug="true"` and non-Windows rows. A second definition of a name is skipped, and duplicate headers are refused.
- **Keys and commands are separate tables, as in 1.12.** `SetBinding` never reads the command (`0x4b7490`). `RunBinding` runs the body in the VM. `RunCommand` sets `keystate` around the body and clears it after (`0x4b7c42`).
- **Engine behaviour sits behind the Lua globals the bodies call.** These are the movement, turn and mouse-steer bits. Focus loss clears the direction bits and keeps the mouse pair (`0x514490`). One key's stop clears its bit, as `0x514b70` does, so releasing W while UP is held stops the character.
- **Behaviour changes that follow the stock bodies.**
- The nameplate keys (V, Shift-V, Ctrl-V) now run the stock bodies.
- ALT-Z runs `CloseAllWindows(); UIParent:Hide()`.
- The minimap, sound, stance and bonus-bar keys call their stock functions.
- The wheel zooms one `CameraZoomIn(1.0)` per notch.
- **Platform.** benilla answers as the PC build, so a `platform="mac"` row does not register on any OS.
Verified:
- New tests: a fixture `Bindings.xml` (reordered, edited, and with debug, empty, nested and Mac rows) drives the command list, defaults and bodies. The V keys follow the stock bodies. No stock body's text remains in the tree. Every stock body's calls resolve after a production load. Each test fails with its change undone.
- `scripts/gates.sh` and `scripts/smoke.sh` are green.
- A live run with `WOW_GM=off`:
- W and Space moved and jumped.
- B toggled the backpack.
- 1 reached `UseAction`.
- V, Shift-V and Ctrl-V behaved as the stock bodies do.
Built on macOS; no platform seam touched.
Fixes#570Fixes#563🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The core now loads the player's own `Interface\FrameXML\FrameXML.toc` off their chain, row by row in its order (`0x6edb90`, from `UI_Init` at `0x48ffed`), instead of benilla's copy of the list, and `benilla.toc` is gone. Order is observable in 1.12.1: event listeners run in registration order, children are kept in creation order, and hit-test ties go to the earlier frame.
- The world VM starts bare: `GlobalStrings.lua` and `Fonts.xml` load as the toc's first rows. The boot VM keeps `GlobalStrings.lua` and the emote tokens, and no longer loads `Fonts.xml`, which nothing reads there (#307).
- A chain without `FrameXML.toc` builds no stock UI and still loads the layer and addons, as the reference does (`"Couldn't open %s"`, severity 2, return unread); benilla reports it as an error.
- Every test kit now loads in the production order, `FrameXML.toc`'s rows then `layer.toc`'s, and a guard fails any kit that loads out of it. Two tests whose premises the stock order disproves are gone.
- `scripts/smoke.sh` counts VMs by the VM marker, since `Fonts.xml` now loads once per world VM.
Verified: new tests for the load list against the chain toc and a reordered fixture toc, the missing-toc case, and two stock frames on one event running in the toc's order (ExhaustionTick before PlayerFrame on `PLAYER_UPDATE_RESTING`), each failing on main's order. A before/after dump of globals, frames, draw order, event listeners and saved variables differs only in stock-order effects: the reordered listeners and children, draw-order ties between non-overlapping frames, and `UnitPopup`'s quality colours now sharing the live tables. `scripts/gates.sh` and `scripts/smoke.sh` are green, and live before/after runs show 0 errors and identical warnings. Built on macOS; no platform seam touched.
Fixes#555Fixes#307🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The comment and string passes' last leftovers, and one lookup made to match the reference.
**Comments that stated a wrong reference fact**, now stating the right one:
- A cast failure clears the GCD, and a failed cooldown-on-event spell also drops its parked record (`0x6e73cc`).
- `SetUnit` is `0x505d70`; `0x476cb0` is the component clone it reaches.
- Interior sound follows the character, where the listener sits, not the camera.
- Template inheritance re-runs `LoadXML` on the one object against each template (`0x76985c`), and reads `name` and `id` from the instance alone (`0x769770`).
- On the ghost-to-alive edge the reference clears its corpse cache and sends nothing (`0x5ee990`, `0x491f50`).
- The name cache has no high-bit miss. An empty-name answer evicts the record in the reference (`0x55f6f0`).
- The combat-log deferred queue (`0xc4e208`) waits only on an item name. A unit name never waits.
- A flying spline's segments are paced by curve length (`0x454320`, `0x453760`), not by chord length.
- A stop within 3 yd of its stop point keeps the arrow nocked (`0x6019e4`).
- A missile's sound id is at `CMissile+0x40`.
- A nil or unresolved `AssistUnit` token is a gap against `0xb8` ERR_GENERIC_NO_TARGET, not a deviation.
**Strings:**
- Test, log and tool messages lose leftover history wording and placeholder labels.
- The extractor's scan summaries describe the current tint and sequence-clock paths.
- The smoke script's failure text carries no private numbers.
**Line continuations:** twelve log, panic and assert strings get back the `\` continuations they had lost. Each printed a run of spaces mid-message. Fixes#106.
**One lookup:** a region's `inherits=` is looked up as one verbatim name, matched case-insensitively (`0x6ee6f0`, `_strnicmp` at `0x6ee747`), as `framexml::expand` already did for frames. The gate before it split on commas and matched case-sensitively, so a texture or font string whose `inherits=` spelled its template in another case never received the template. `a_region_takes_a_mis_cased_template` fails before this change and passes after.
**Tests:** the quest tests stop loading three FrameXML files that no quest file uses.
**In game:** nothing changes for the stock UI. Log lines read as one sentence, and an addon texture that names its template in another case now gets it.
Verified with `scripts/check.sh` on the changed crates, plus the loader and quest tests with the install present.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The comment and string passes' last leftovers, and one lookup made to match the reference.
**Comments that stated a wrong reference fact**, now stating the right one:
- A cast failure clears the GCD, and a failed cooldown-on-event spell also drops its parked record (`0x6e73cc`).
- `SetUnit` is `0x505d70`; `0x476cb0` is the component clone it reaches.
- Interior sound follows the character, where the listener sits, not the camera.
- Template inheritance re-runs `LoadXML` on the one object against each template (`0x76985c`), and reads `name` and `id` from the instance alone (`0x769770`).
- On the ghost-to-alive edge the reference clears its corpse cache and sends nothing (`0x5ee990`, `0x491f50`).
- The name cache has no high-bit miss. An empty-name answer evicts the record in the reference (`0x55f6f0`).
- The combat-log deferred queue (`0xc4e208`) waits only on an item name. A unit name never waits.
- A flying spline's segments are paced by curve length (`0x454320`, `0x453760`), not by chord length.
- A stop within 3 yd of its stop point keeps the arrow nocked (`0x6019e4`).
- A missile's sound id is at `CMissile+0x40`.
- A nil or unresolved `AssistUnit` token is a gap against `0xb8` ERR_GENERIC_NO_TARGET, not a deviation.
**Strings:**
- Test, log and tool messages lose leftover history wording and placeholder labels.
- The extractor's scan summaries describe the current tint and sequence-clock paths.
- The smoke script's failure text carries no private numbers.
**Line continuations:** twelve log, panic and assert strings get back the `\` continuations they had lost. Each printed a run of spaces mid-message. Fixes#106.
**One lookup:** a region's `inherits=` is looked up as one verbatim name, matched case-insensitively (`0x6ee6f0`, `_strnicmp` at `0x6ee747`), as `framexml::expand` already did for frames. The gate before it split on commas and matched case-sensitively, so a texture or font string whose `inherits=` spelled its template in another case never received the template. `a_region_takes_a_mis_cased_template` fails before this change and passes after.
**Tests:** the quest tests stop loading three FrameXML files that no quest file uses.
**In game:** nothing changes for the stock UI. Log lines read as one sentence, and an addon texture that names its template in another case now gets it.
Verified with `scripts/check.sh` on the changed crates, plus the loader and quest tests with the install present.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A failure message pointed at records and bug ids that are not in this
repo; each now says the fact alone. Only message text changes.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A failure message pointed at records and bug ids that are not in this
repo; each now says the fact alone. Only message text changes.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
With no `.probe-identity`, the scrub kept the shell's WOW_USER, WOW_PASS
and WOW_CHAR in the environment, so the realm leg inherited WOW_CHAR and
`realm_select::smoke` refused it: the documented
`WOW_USER=… WOW_PASS=… WOW_CHAR=… scripts/smoke.sh` could never pass.
The account is resolved into the script's own variables first and each
leg passes it by name, so it now leaves the environment either way.
Verified: the scrub run alone with the account in the shell leaves only
WOW_HOST in a child's environment, with and without a declared identity.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
With no `.probe-identity`, the scrub kept the shell's WOW_USER, WOW_PASS
and WOW_CHAR in the environment, so the realm leg inherited WOW_CHAR and
`realm_select::smoke` refused it: the documented
`WOW_USER=… WOW_PASS=… WOW_CHAR=… scripts/smoke.sh` could never pass.
The account is resolved into the script's own variables first and each
leg passes it by name, so it now leaves the environment either way.
Verified: the scrub run alone with the account in the shell leaves only
WOW_HOST in a child's environment, with and without a declared identity.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Each header is a one-line summary, the environment it reads and its
usage; the history, measurements and record numbers go. Only comment
lines change: with comments stripped the scripts are byte-identical.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Each header is a one-line summary, the environment it reads and its
usage; the history, measurements and record numbers go. Only comment
lines change: with comments stripped the scripts are byte-identical.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
What a clone lacked: the gates went green with a thousand skipped tests and said nothing, every
login script assumed a worktree-pool slot and its probe account, the client defaulted to the
maintainer's account, a docker path and a home directory were in the code, the temp files broke
on GNU coreutils, and two generators could only run against a sibling repo.
Now the test gates print how many data-gated tests skipped and why; a scripted run logs in as
the account its checkout declares in `.probe-identity` or as the three variables name, never a
default, and the client refuses a scripted login on any other account; the event and message
tables regenerate from vendored inputs, byte for byte; the docs state the real prerequisites.
Out of the tree, into the maintainer's own tooling: the worktree pool, its hooks, the perf leg
runner, the cross-platform compile and the Vulkan container leg — a way of working, and setups
not every machine has. One hook stays: no writes into the install.
Two tests that could never run are gone with their subject: the quest-timer geometry diff
compared a copy of the window that no longer exists (and joined a chain path onto a folder, so
it always skipped), and the FrameXML flag gate now reads the stock files off the player's own
patch chain instead of an extracted folder only one machine had.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
What a clone lacked: the gates went green with a thousand skipped tests and said nothing, every
login script assumed a worktree-pool slot and its probe account, the client defaulted to the
maintainer's account, a docker path and a home directory were in the code, the temp files broke
on GNU coreutils, and two generators could only run against a sibling repo.
Now the test gates print how many data-gated tests skipped and why; a scripted run logs in as
the account its checkout declares in `.probe-identity` or as the three variables name, never a
default, and the client refuses a scripted login on any other account; the event and message
tables regenerate from vendored inputs, byte for byte; the docs state the real prerequisites.
Out of the tree, into the maintainer's own tooling: the worktree pool, its hooks, the perf leg
runner, the cross-platform compile and the Vulkan container leg — a way of working, and setups
not every machine has. One hook stays: no writes into the install.
Two tests that could never run are gone with their subject: the quest-timer geometry diff
compared a copy of the window that no longer exists (and joined a chain path onto a folder, so
it always skipped), and the FrameXML flag gate now reads the stock files off the player's own
patch chain instead of an extracted folder only one machine had.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Until now this repo was an export of a private tree, pushed as squashed snapshots. From here on
it is the working tree itself, and the private repo is archived.
The root holds AGENTS.md and the Rust and GitHub files only. Everything an agent needs to orient
is in docs/: METHOD.md (the rules and how the project is built, 123 lines, project method only),
MAP.md (generated from the code at every land) and CONTRIBUTING.md (what gets in and how a
change is judged). One skill, `probe`, holds the procedure for running the client unattended
against the local server.
Brought over because the project needs it: reference/ (the 1.12 name catalogues 34 source
files read; the export failed its own tests without them), the gate chain and the pool
(scripts/), the hooks that enforce the install rule and the worktree rule, the full .gitignore,
the `cargo play` alias, and the original wording of eleven source files the export had reworded.
Left out: the decision records (archived with the private repo), the export tooling, the
operator's Discord and triage loop, the lab-laptop scripts and the instruments nothing
references, which live outside the repo now. What read the records is gone from the tooling:
no corpus lint, no records index, no `reserve` verb, no number-collision check.
Three hooks are gone: fmt-at-commit and the file-size nudge were redundant with the gates, and
the assets/ui guard is now a test (`assets_ui_does_not_grow`) that fails for everyone, not only
for a Claude Code session. The remaining pool hooks stand down on a machine without a pool.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Until now this repo was an export of a private tree, pushed as squashed snapshots. From here on
it is the working tree itself, and the private repo is archived.
The root holds AGENTS.md and the Rust and GitHub files only. Everything an agent needs to orient
is in docs/: METHOD.md (the rules and how the project is built, 123 lines, project method only),
MAP.md (generated from the code at every land) and CONTRIBUTING.md (what gets in and how a
change is judged). One skill, `probe`, holds the procedure for running the client unattended
against the local server.
Brought over because the project needs it: reference/ (the 1.12 name catalogues 34 source
files read; the export failed its own tests without them), the gate chain and the pool
(scripts/), the hooks that enforce the install rule and the worktree rule, the full .gitignore,
the `cargo play` alias, and the original wording of eleven source files the export had reworded.
Left out: the decision records (archived with the private repo), the export tooling, the
operator's Discord and triage loop, the lab-laptop scripts and the instruments nothing
references, which live outside the repo now. What read the records is gone from the tooling:
no corpus lint, no records index, no `reserve` verb, no number-collision check.
Three hooks are gone: fmt-at-commit and the file-size nudge were redundant with the gates, and
the assets/ui guard is now a test (`assets_ui_does_not_grow`) that fails for everyone, not only
for a Claude Code session. The remaining pool hooks stand down on a machine without a pool.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
gromandClaude Opus 5