Was waiting for all ~50 tiles (radius 4) to fully prepare + finalize
before entering the game. Now loads only the immediate surrounding tiles
during the loading screen, then restores the full radius for in-game
streaming. setLoadRadius just sets an int — actual loading happens lazily
via background workers during the game loop.
- Worker threads: use (cores - 1-2) instead of cores/2, minimum 4
- Outer upload batch in processReadyTiles: ALL model/texture uploads per
frame share a single command buffer submission + fence wait
- Upload multiple models per finalization step: 8 M2s, 4 WMOs, 16 doodads
per call instead of 1 each (all within same GPU batch)
- Terrain chunks: 64 per step instead of 16
- Skip redundant M2 file I/O: thread-safe uploadedM2Ids_ set lets
background workers skip re-reading+parsing models already on GPU
- processAllReadyTiles (loading screen) and processOneReadyTile also
wrapped in outer upload batches
Every uploadBuffer/VkTexture::upload called immediateSubmit which did a
separate vkQueueSubmit + vkWaitForFences. Loading a single creature model
with textures caused 4-8+ fence waits; terrain chunks caused 80+ per batch.
Added beginUploadBatch/endUploadBatch to VkContext: records all upload
commands into a single command buffer, submits once with one fence wait.
Staging buffers are deferred for cleanup after the batch completes.
Wrapped in batch mode:
- CharacterRenderer::loadModel (creature VB/IB + textures)
- M2Renderer::loadModel (doodad VB/IB + textures)
- TerrainRenderer::loadTerrain/loadTerrainIncremental (chunk geometry + textures)
- TerrainRenderer::uploadPreloadedTextures
- WMORenderer::loadModel (group geometry + textures)
Terrain finalization was uploading all 256 chunks (GPU fence waits) in one
atomic advanceFinalization call that couldn't be interrupted by the 5ms time
budget. Now split into incremental batches of 16 chunks per call, allowing
the time budget to yield between batches.
M2 instance creation had O(N) dedup scans iterating ALL instances to check
for duplicates. In cities with 5000+ doodads, this caused O(N²) total work
during tile loading. Replaced with hash-based DedupKey map for O(1) lookups.
Changes:
- TerrainRenderer::loadTerrainIncremental: uploads N chunks per call
- FinalizingTile tracks terrainChunkNext for cross-frame progress
- TERRAIN phase yields after preload and after each chunk batch
- M2Renderer::DedupKey hash map replaces linear scan in createInstance
and createInstanceWithMatrix
- Dedup map maintained through rebuildSpatialIndex and clear paths
- Async creature model loading: M2 file I/O and parsing on background threads
via std::async, GPU upload on main thread when ready (MAX_ASYNC_CREATURE_LOADS=4)
- CharSections.dbc lookup cache: O(1) hash lookup instead of O(N) full DBC scan
per humanoid NPC spawn (was scanning thousands of records twice per spawn)
- Frame time budget: 4ms cap on creature spawn processing per frame
- Wolf/worg model name check cached per modelId (was doing tolower+find per
hostile creature per frame)
- Weapon attach throttle: max 2 per 1s tick (was attempting all unweaponized NPCs)
- Separate texture application tracking (displayIdTexturesApplied_) so async-loaded
models still get skin/equipment textures applied correctly
- Add water/lava/lighting rendering features to README
- Add transport riding to movement features
- Update status date and rendering capabilities
- Note interior shadow and lava steam known gaps
- Add magma/slime rendering path to water shader (fbm noise, crust/molten/core coloring)
- Fix WMO liquid height filter rejecting high-altitude zones like Ironforge (Z>300)
- Allow interior WMO magma/slime MLIQ groups to load (skip only water/ocean)
- Mark LAVASTEAM.m2 as spell effect for proper additive blend, hide emission mesh
- Add isLavaModel flag for M2 ForgeLava/LavaPots UV scroll fallback
- Add isLava material detection in WMO renderer for lava texture UV animation
- Fix WMO material UBO colors for magma (was blue, now orange-red)
- Add case-insensitive "glass" detection for WMO window materials
- Make instance (WMO-only) glass highly transparent (12-35% alpha)
so underwater scenes are visible through Deeprun Tram windows
- Keep normal world windows at existing opacity (40-95% alpha)
- Disable shadow mapping for interior WMO groups to fix dark
indoor areas like Ironforge
- Fix NULL renderer pointers by moving TransportManager connection after
initializeRenderers for WMO-only maps
- Fix tram direction by negating DBC TransportAnimation X/Y local offsets
before serverToCanonical conversion
- Implement client-side M2 transport boarding via proximity detection
(server doesn't send transport attachment for trams)
- Use position-delta approach: player keeps normal movement while
transport's frame-to-frame motion is applied on top
- Prevent server movement packets from clearing client-side M2 transport
state (isClientM2Transport guard)
- Fix getPlayerWorldPosition for M2 transports: simple canonical addition
instead of render-space matrix multiplication
- Add movementSuppressTimer to camera controller that forces all movement
keys to read as false, preventing held W key from carrying through
loading screens (fixes always-running-forward after instance portals)
- Increase shadow frustum default from 60 to 72 units (+20%)
- Make shadow distance configurable via setShadowDistance() (40-200 range)
- Add shadow distance slider in Video settings tab (persisted to config)
stepX/stepY were transposed: columns stepped south instead of east and
rows stepped east instead of south, mirroring all overworld water. Also
fix per-chunk origin to use layer.x for east offset and layer.y for south.
Remove frame throttling that skipped shadow updates in dense scenes,
causing visible flicker on player and NPCs. Reduce shadow half-extent
from 180 to 60 for 3x higher resolution on nearby shadows.
- Stronger wall collision push (0.35/0.15) and swept push (0.45/0.25)
for interior/exterior WMOs to reduce clipping through tunnel walls
- Use all triangles (not just pre-classified walls) for collision checks
- Allow invisible collidable triangles (MOPY 0x01 without 0x20) to block
- Pass insideWMO flag to all collision callers, match swim sweep to ground
- Widen swept hit detection radius from 0.15 to 0.25
- Restrict camera zoom to 12 units inside WMO interiors
- Fix /unstuck launching player above WMOs: remove +20 fallback, use
gravity when no floor found
- Slash and Enter keys always focus chat unless already typing
WMO origins can be far from their visible geometry, causing large city
buildings to be culled from the shadow pass. Use world bounding box for
instance culling and per-group AABB culling. Also increase WMO shadow
cull radius to match the shadow map coverage (180 units).
Fix VK_ERROR_DEVICE_LOST crash by allocating per-frame scene history
images (color + depth) instead of a single shared image that raced
between frames in flight. Water refraction can now be toggled via
Settings > Video > Water Refraction.
Without refraction: richer blue base colors, animated caustic shimmer,
and normal-based color shifts give the water visible life. With
refraction: clean screen-space refraction with Beer-Lambert absorption.
Disabling clears scene history to black for immediate fallback.
The glm::quat(w,x,y,z) constructor was receiving swapped X/Y components,
causing doodads like the Deeprun Tram gears to be oriented horizontally
instead of vertically. Also use createInstanceWithMatrix for instance WMO
doodads to preserve full rotation from the quaternion.
Classify light shafts, portals, spotlights, bubbles, and similar M2
doodads as spell effects so they render with additive blending instead
of as solid opaque objects.
Set camera yaw from server orientation on world load so teleports face
the correct direction.
Reduce area trigger minimum radius (3.0 sphere, 4.0 box) to prevent
premature portal firing near tram entrances.
WMO interior doodads (gears, decorations) were blocking player movement
via M2 collision. Skip collision for all WMO doodad M2 instances since
the WMO itself handles wall collision.
Also filter WMO wall collision using MOPY per-triangle flags: only
rendered+collidable triangles block the player, skipping invisible
collision hulls.
Revert tram portal extended range (no longer needed with collision fix).
Parse MOPY per-triangle flags in WMO groups and exclude detail/decorative
triangles (flag 0x04) from collision detection. This prevents invisible
walls from objects like gears and railings in WMO interiors.
Add WotLK area trigger IDs 2173/2175 to extended-range tram triggers.
- Equipment-driven geoset selection: read GeosetGroup1 from ItemDisplayInfo
for legs/feet/chest to pick covered mesh variants (1302+ pants, 402+ boots,
802+ sleeves) instead of always defaulting to bare geosets
- Prevent per-instance skin override from replacing baked/composited armor
textures on equipped NPCs
- Set bald NPC hair texture slot to skin texture so scalp isn't white
- Skip white fallback textures in per-instance hair overrides
- Remove debug texture dump, reduce NPC logging to DEBUG level
- Remove bogus 2-byte skip after materialId in MLIQ parser that shifted
all vertex heights and tile flags by 2 bytes (garbage data)
- Skip liquid loading for interior WMO groups (flag 0x2000) to prevent
indoor water from rendering as outdoor canal water
- Clear movement inputs on teleport/portal to prevent auto-running after
zone transfer (held keys persist through loading screen)
- Fix Stormwind barracks floor: interior WMO groups named "facade" were
incorrectly marked as LOD shells and hidden when close. Add !isIndoor
guard to all LOD detection conditions so interior groups always render.
- Fix water exit stair clipping: anchor lastGroundZ to current position
on swim exit, set grounded=true for full step-up budget, add upward
velocity boost to clear stair lip geometry.
- Re-enable NPC humanoid equipment geosets (kEnableNpcHumanoidOverrides)
so guards render with proper armor instead of underwear.
- Keep instance portal GameObjects animated (spinning/glowing) instead
of freezing all GO animations indiscriminately.
- Fix equipment disappearing after instance round-trip by resetting
dirty tracking on world reload.
- Fix multi-doodad-set loading: load both set 0 (global) and placement-
specific doodad set, with dedup to avoid double-loading.
- Clear placedWmoIds in softReset/unloadAll to prevent stale dedup.
- Apply MODF rotation to instance WMOs, snap player to WMO floor.
- Re-enable rebuildSpatialIndex in setInstanceTransform.
- Store precomputeFloorCache results in precomputed grid.
- Add F8 debug key for WMO floor diagnostics at player position.
- Expand mapIdToName with all Classic/TBC/WotLK instance map IDs.
Low-vertex groups (<100 verts) were incorrectly marked as distance-only
LOD shells in small WMOs like Stockades. Now only applies this heuristic
to large WMOs (50+ groups) where it's needed for city exterior shells.
- NPC hair/skin textures now use per-instance overrides instead of shared
model-level textures, so each NPC shows its own hair color/style
- Hair/skin DBC lookup runs for every NPC instance (including cached models)
rather than only on first load
- Fix keyframe binary search to use float comparison matching original
linear scan semantics
- Replace O(n) linear keyframe search with O(log n) binary search in both
M2 and Character renderers (runs thousands of times per frame)
- Smoke particle removal: swap-and-pop instead of O(n²) vector erase
- Character render backface cull: eliminate sqrt via squared comparison
- Quaternion validation: use length² instead of sqrt-based length check
Use cached model flags (isValid, isSmoke, isInvisibleTrap, isGroundDetail,
disableAnimation, boundRadius) on M2Instance instead of models.find() in
the hot culling paths. Also complete cached flag initialization in
createInstanceWithMatrix().
- Glow sprites now use dedicated vertex buffer (glowVB_) separate from
M2 particle buffer to prevent data race when renderM2Particles()
overwrites glow data mid-flight
- Move fadeAlpha from shared material UBO to per-draw push constants,
eliminating cross-instance alpha race on non-double-buffered UBOs
- Smooth adaptive render distance transitions to prevent pop-in/out
at instance count thresholds (1000/2000)
- Distance-tiered character bone throttling: near (<30u) every frame,
mid (30-60u) every 3rd, far (60-120u) every 6th frame
- Skip weapon instance animation updates (transforms set by parent bones)
- Split M2 instances into fast-path index lists (animated, particle-only,
particle-all, smoke) to avoid iterating all 46K instances per frame
- Cache model flags (hasAnimation, disableAnimation, isSmoke, etc.) on
M2Instance struct to eliminate per-frame hash lookups
- Replace full rebuildSpatialIndex on position/transform updates with
incremental grid cell remove+add, preventing 8.5ms/frame rebuild cost
- Advance animTime for all instances (texture UV animation) but only
compute bones and particles for the ~3K that need it
M2_UPDATE: 10.7ms → 2.0ms, FPS: 35 → 55-59
The single sceneColorImage races between frames with MAX_FRAMES_IN_FLIGHT=2:
frame N-1's water shader reads it while frame N's captureSceneHistory writes
it via vkCmdCopyImage. Pipeline barriers only sync within a single command
buffer, not across submissions on the same queue.
This caused VK_ERROR_DEVICE_LOST after ~700 frames on any map with water.
Disable the capture entirely for now — water renders without refraction.
TODO: allocate per-frame scene history images to eliminate the race.
The previous commit changed std::move to copy for terrain/mesh data to fix
the empty-cache bug. But copying ~8 MB per tile × 81 tiles caused a 60s
streaming timeout.
The tile cache was already broken before — putCachedTile stored a shared_ptr
to the same PendingTile whose data was moved out, so cached tiles always had
empty meshes. Remove the putCachedTile call entirely; tiles re-parse from
ADT files (asset manager file cache hit) when they re-enter streaming range.
The softReset cache clear from the previous commit remains as safety for
map transitions.
Three fixes:
1. Water captureSceneHistory gated on hasSurfaces() — the image layout
transitions (PRESENT_SRC→TRANSFER_SRC→PRESENT_SRC) were running every
frame even on WMO-only maps with no water, causing VK_ERROR_DEVICE_LOST.
2. Tile cache invalidation: softReset() now clears tileCache_ since cache
keys are (x,y) without map name — prevents stale cross-map cache hits.
3. Copy terrain/mesh into TerrainTile instead of std::move — the moved-from
PendingTile was cached with empty data, so subsequent map loads returned
tiles with 0 valid chunks from cache.
Also adds diagnostic skip env vars (WOWEE_SKIP_TERRAIN, WOWEE_SKIP_SKY,
WOWEE_SKIP_PREPASSES) and a 0-chunk warning in loadTerrain.
Add null checks for vertex/index buffers, pipelines, and zero-count
draws in WMO render path. The shadow pass already had buffer validation
but the main render() was missing it, which could cause GPU crashes
on WMO-only maps like Stockades (26 groups).
After VK_ERROR_DEVICE_LOST, beginFrame returns VK_NULL_HANDLE but
renderWorld() and renderHUD() were still called, passing the null
handle to vkCmdBindPipeline which triggered a validation abort.
When the GPU device is lost (unrecoverable Vulkan error), the app now
closes cleanly instead of looping with a black screen. Also adds
vk_context.hpp include for the isDeviceLost() check.
Two bugs in SMSG_MESSAGECHAT parser for MONSTER_SAY/YELL/EMOTE:
1. Sender name included trailing null byte from server (nameLen includes
null terminator). The embedded null in std::string caused ImGui to
truncate the concatenated display string at the NPC name, hiding
" says: <message>" entirely.
2. Missing NamedGuid receiver name for non-player/non-pet targets. When
the receiver GUID is a creature, the server writes an additional
SizedCString (target name) that we weren't reading, shifting all
subsequent field reads.
Also adds MONSTER_WHISPER, MONSTER_PARTY, RAID_BOSS_EMOTE, RAID_BOSS_WHISPER
chat types with proper parsing and display formatting (says/yells/whispers).
Root cause: LOGIN_VERIFY_WORLD path did not set areaTriggerCheckTimer_ or
areaTriggerSuppressFirst_, so the Stockades exit portal (AT 503) fired
immediately on login, teleporting the player back to Stormwind and crashing
the GPU during the unexpected map transition.
Fixes:
- Set 5s area trigger cooldown + suppress-first in handleLoginVerifyWorld
(same as SMSG_NEW_WORLD handler already did for teleports)
- Add deviceLost_ flag to VkContext so beginFrame returns immediately once
VK_ERROR_DEVICE_LOST is detected, preventing infinite retry loops
- Track device lost from both fence wait and queue submit paths
- Extract initializeRenderers() from loadTestTerrain() so WMO-only maps
(dungeons/raids) initialize renderers directly without a dummy ADT path
- Defer setState(IN_GAME) until after processing any pending deferred world
entry, preventing brief IN_GAME flicker on the wrong map
- Remove verbose area trigger debug logging (every-second position spam)
Reset descriptor pools in CharacterRenderer/M2Renderer/WMORenderer on map
change to prevent VK_ERROR_DEVICE_LOST from pool exhaustion. Defer re-entrant
SMSG_NEW_WORLD during active world load to avoid recursive cleanup crashes.
Gate swim bubbles on swimming state, skip redundant shadow pipeline re-init,
add WOWEE_SKIP_* env vars for render isolation debugging.
