The radial blur read the scene depth 28 times per full-resolution pixel. A
new mask pre-pass (`fs_mask`, R8Unorm at ceil(w/2) x ceil(h/2)) loads the
depth once per half-res texel and thresholds it with the same sky test
(depth <= 1e-7 on infinite reverse-Z); the blur then takes 28 clamp-to-edge
bilinear taps of the mask. Only the mask stage is MSAA-specialised (single
and multisampled depth layouts, one pipeline each); the blur pipeline is
keyed by the target format alone.
Verified: cargo check -p benilla; all four WGSL variants (blur, blur +
MULTISAMPLED, SHAFT_MASK, SHAFT_MASK + MULTISAMPLED) compose and validate
under naga_oil 0.20 / naga 27 offline. No GPU or WoW data in this
environment, so the image is not captured; built on Linux only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y6yUWgPosSakvL23gtgvpH
The daylight lane's tooltip had raw line breaks inside a Lua string instead of
an escaped blank line, a syntax error that broke OptionsFrame.xml's inline
script (an error dialog on the UI-fixture captures).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review B4: the Stormwind window split changed the default image with no switch.
New cvar daylightWindowSplit (default 1 = merged behaviour, a future High-only
preset member), bridged to daylight::set_window_split. Review B5: each settled
exterior torch slot is keyed by the resident tiles whose footprint reaches its
sphere (terrain_torch_generation_near), not a global stamp that re-rendered
every slot on any stream event.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review W4: the bend height was normalised per submesh, so a trunk + canopy
detail doodad tore apart at the seam; the span is now taken over all of the
model's submeshes. Review W5: clutter meshes carry UV_1 (VERTEX_UVS_B), so the
six clutter pipelines are also warmed on the POS+NORMAL+UV0+UV1+COLOR quad.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review P4: emissive_boost now lifts only above a 0.6 knee (equal to the old
gain at 1.0), so additive spells and lava keep their look and the excess feeds
the bloom. Review P3: the window gain applies to static_gx's SIDN light term
before fog, not to the whole fogged batch. Review P6: Low-tier extract taps
scale with the quarter-size target. Review P7: excess capped at 64 (no NaN).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review P1: the path mean lifted every clear sky pixel equally; it is now the
occlusion ratio times an aspect-corrected radial falloff from the sun.
Review P2: sky = depth <= 1e-7 (cleared / sky-pinned), no longer a nearclip-
dependent ramp. Review P5 (partial): read one MSAA sample per tap.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Foliage wind is a lighting-preset member but bypasses VideoConfig. Read back its world resource in the shared preset clamp test so the merged row is covered like every other member.
Co-Authored-By: GPT-5.6 Sol (Codex)
The wind lane's unverified closer handoff coordinates put the capture camera inside a Goldshire building. Restore the open-grass coordinates used by the lane's successful before/after captures.
Co-Authored-By: GPT-5.6 Sol (Codex)
Framed with the vista instrument: Elwynn above the canopy (the old eye stood
inside a tree), Redridge from the known-good Lakeshire dock eye with a wider
look, Hyjal above the terrain (the Light.dbc sphere centre z is underground),
Ironforge over a floor face inside the city (the old eye was outside the WMO).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sun-fog strength 0.25 / angle 0.4 / lean 0.4, end-fog lean 0.25: the first
pass turned dusk ridges into flat orange cut-outs against a blue sky. The
dome's low band now MIXES toward the far fog colour by 15 deg (an offset
over-pushed an already warm dusk ring). update_fog_model joins the
ordering test's ALLOWED list (it runs in the resolve set, after it).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Let deterministic captures read an explicitly selected BENILLA_HOME config while keeping every capture persistence path disabled. This makes the post ON/OFF homes effective without weakening capture hermeticity.
Co-Authored-By: GPT-5.6 Sol (Codex)
Classify static alpha-tested tree and bush leaf batches, log accepted model paths, and apply gentle anchor-phased crown sway and flutter in both retained and fade-twin render paths. Keep animated doodads and tree shadows unchanged, and add wind capture instruments.
Co-Authored-By: GPT-5.6 Sol (Codex)
Port the stateless gust and veer profile, scale it from live weather, and publish one sample through the WindField resource and MonkeyFrame each frame. Add the foliageWind tier CVar, High preset value, and Advanced Graphics row.
Co-Authored-By: GPT-5.6 Sol (Codex)
Preserve p0 fog and MonkeyFrame hooks, water's row-view ABI, and the post lane's emissive, bloom, sun-shaft, grading, cvar, and options work.
Co-Authored-By: GPT-5.6 Sol (Codex)
The combine's dither lane was armed only by the WOW_DITHER env var. It is now
the `skyDither` cvar (0/1, default 0 = reference look; Graphics preset High
= 1) with an Advanced Graphics "Sky Dithering" checkbox row (EN + RU). The
cvar lands on `benilla_world::ffx_glow::SkyDither` (extracted) through the new
`benilla-app/src/monkey_gfx.rs` bridge; WOW_DITHER=1 still forces it on.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
71 upstream commits: the repo-wide comment rewrite, the hover highlight as the
scene ambient, M2 Mod/Mod2x batches drawing the bare texel, test and contribution
changes. Conflicts were comment-vs-code in 36 files; where upstream changed only
comments our code was kept, and wow_model.wgsl carries both our lighting lanes
and upstream's new unlit/Mod/Mod2x paths.
Follow-ups folded in:
- our nine Advanced Graphics / water / fog options tests take upstream's
wow_data_or_skip! guard
- PostUpdate ambiguity ceiling re-measured under upstream's new census: 466,
with the per-family breakdown in the comment
- LIGHTING.md: new upstream base
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Water (Enhanced/High; Classic stays byte-identical to upstream):
- the enhanced water now lives in its own WGSL library, benilla::enhanced_water
(enhanced_water.wgsl), with its own bindings 103-106; upstream liquid.wgsl
keeps three small hooks only
- depth look: per-channel extinction over the real path through the water,
refraction of an opaque-colour copy bent by the waves (samples landing on
anything in front of the water are refused via the scene depth), caustic
web on sunlit shallow beds
- High: screen-space reflection of the scenery (bisected march against the
scene depth)
- the module fogs its own surface terms only (the refracted/reflected scene is
already fogged); depth+colour copies run both-or-neither
- credits: portions derived from WarcraftXL wxl-experimental-water by
iThorgrim, used with permission (WATER.md)
Volumetric fog (Advanced Graphics -> Volumetric Fog: Off/Low/High, cvar
volumetricFog, env WOW_VOLFOG): own fullscreen pass after the main pass,
near-field haze converging on the zone fog colour, denser at dawn and in
weather, faint indoors, plus sun/moon light shafts from the shadow map.
Tests: addon-harness oracle fixtures no longer require unix symlinks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Water (Enhanced/High only; Classic stays byte-identical to upstream):
- scene depth resolved after the main opaque pass and bound to liquids
(liquid/scene_depth.rs, water_depth.rs)
- procedural multi-band waves with analytic normals; long ocean swell moves
vertices, mirrored on the CPU (liquid/waves.rs)
- separate ocean and inland profiles; depth-driven colour and soft edge
- Fresnel sky reflection, sun path, moon path, point-light glints
- beach surf driven by the shore wave phase: continuous front, dissolving
lace, swash sheet; only where the authored sea bed is shallow, so sunken
objects never foam; thin bed clutter (grass, reeds) ignored for contact foam
- swimmers: wake foam, treading rings, visual-only bob with the swell
- river flow effect tried and removed at the owner's request
Lava: magma surfaces emit warm point lights through the lighting module
(lighting/lava_light.rs, lavaLightGain).
Options: Water Quality + Lava Glow rows (EN, RU strings prepared).
Docs: WATER.md. Credits: design reference WarcraftXL wxl-experimental-water,
author iThorgrim (reuse permitted with attribution).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Comments that named a chat reporter, cited a chat channel or message id, or quoted a report word for word now state the symptom in their own words. The test fixtures that used real character names use invented ones.
Verified: 52 of the 56 files strip to identical code with their comments removed; the other 4 differ only in the fixture names, and with those mapped back they strip identical too. `cargo fmt --check` is clean, and the `ui_party`, `unit_frame_tests` and `capture` tests pass with the install present and no skips. Record numbers still in these comments go with the comment pass (#41).
Fixes#34🤖 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 options window, the ESC menu and the key bindings page follow the 1.15 era client because
1.12's settings screens are much worse to use.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The 1.12 highlight is not a fixed 64/255 lift. SetHighlight samples the
scene's committed ambient (read at 0x614576-0x6145bd), and every batch
adds it to its colour before the final clamp, lit or unlit.
0xff404040 is only the fallback before a map's light loads. The
highlight now follows the time of day, zone and weather: roughly
(29, 60, 84)/255 at night to (104, 130, 154)/255 at noon, where it was
always (64, 64, 64)/255.
Unlit M2 batches in blend modes 0 to 4 now draw
texel × clamp(tint·M2Color + highlight), the unlit program's c28 + c29.
They took no highlight before. The reference holds the ambient sampled
when a highlight begins; benilla samples it live, a marked deviation.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
BENILLA_FALLBACK_ICON held the fourth archive path the sweep counts, and
it is gone. The floor exists so the sweep never passes by matching
nothing, and it comes down as paths leave.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
1,874 comment lines become 430; the code is unchanged. The note about
reference graphics CVars that pfUI writes and benilla leaves
unregistered described the CVar registry, not this file. It moves there
with its addresses.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Its readers were our own ActionBar and group-loot files, and both were
retired when those windows moved to the stock FrameXML. The stock files
leave a missing texture empty.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude