Implement GPU-driven Hierarchical-Z occlusion culling for M2 doodads
using a depth pyramid built from the previous frame's depth buffer.
The cull shader projects bounding spheres via prevViewProj (temporal
reprojection) and samples the HiZ pyramid to reject hidden objects
before the main render pass.
Key implementation details:
- Separate early compute submission (beginSingleTimeCommands + fence
wait) eliminates 2-frame visibility staleness
- Conservative safeguards prevent false culls: screen-edge guard,
full VP row-vector AABB projection (Cauchy-Schwarz), 50% sphere
inflation, depth bias, mip+1, min screen size threshold, camera
motion dampening (auto-disable on fast rotations), and per-instance
previouslyVisible flag tracking
- Graceful fallback to frustum-only culling if HiZ init fails
Fix dark WMO interiors by gating shadow map sampling on isInterior==0
in the WMO fragment shader. Interior groups (flag 0x2000) now rely
solely on pre-baked MOCV vertex-color lighting + MOHD ambient color.
Disable interiorDarken globally (was incorrectly darkening outdoor M2s
when camera was inside a WMO). Use isInsideInteriorWMO() instead of
isInsideWMO() for correct indoor detection.
New files:
- hiz_system.hpp/cpp: pyramid image management, compute pipeline,
descriptors, mip-chain build dispatch, resize handling
- hiz_build.comp.glsl: MAX-depth 2x2 reduction compute shader
- m2_cull_hiz.comp.glsl: frustum + HiZ occlusion cull compute shader
- test_indoor_shadows.cpp: 14 unit tests for shadow/interior contracts
Modified:
- CullUniformsGPU expanded 128->272 bytes (HiZ params, viewProj,
prevViewProj)
- Depth buffer images gain VK_IMAGE_USAGE_SAMPLED_BIT for HiZ reads
- wmo.frag.glsl: interior branch before unlit, shadow skip for 0x2000
- Render graph: hiz_build + compute_cull disabled (run in early compute)
- .gitignore: ignore compiled .spv binaries
- MEGA_BONE_MAX_INSTANCES: 2048 -> 4096
Signed-off-by: Pavel Okhlopkov <pavel.okhlopkov@flant.com>
Demote 44 more LOG_WARNING messages to LOG_DEBUG: warden module chunk
progress, entire shutdown/teardown sequence, transport manager connect,
inventory right-click slot, and warden handshake diagnostics. Keeps real
warnings (texture not found, slow handlers, stalls, integrity hash)
visible in the log.
When spline parsing consumes the wrong number of bytes, the subsequent
blockCount read lands on garbage data (e.g. 71 instead of ~5 for UNIT).
Previously the parser logged a warning but continued, reading garbage
mask/field data until hitting truncation. Now it returns false for
CREATE_OBJECT blocks with suspicious counts, letting the block loop
skip cleanly to the next entity.
Also downgrade ~44 diagnostic LOG_WARNING messages to LOG_DEBUG across
17 files (equipment, transport, DBC, heartbeat, chat, GO raypick, etc.)
to reduce log noise and make real warnings visible.
Resolve conflicts:
- audio_callback_handler.cpp: keep PR's animation_controller include
- movement_handler.cpp: use PR accessors with master's transportResolved logic
- world_packets.cpp: keep PR's decomposed version (functions moved to split files)
Apply overkill field fix to world_packets_entity.cpp (WotLK
SMSG_ATTACKERSTATEUPDATE missing uint32 overkill between damage and
subDamageCount).
AzerothCore sends a uint32 overkill field between totalDamage and
subDamageCount. The parser was missing this, causing subDamageCount to
read the first byte of overkill (0 for non-kills) and fail immediately.
This broke all melee swing animations except the killing blow.
Log parse failures with remaining packet size and successful parses with
attacker/target/player GUIDs, damage, and callback status to diagnose
why meleeSwingCallback is never invoked during auto-attack.
If the capability probe ran before the model was fully loaded, all melee
animation IDs would be 0 and auto-attack swings would silently fall back
to STAND (no visible animation). Now re-probes when a melee swing fires
but hasMelee is false.
Added WARNING-level logging to triggerMeleeSwing and CombatFSM to
diagnose the night elf stationary combat animation issue.
MSG_MOVE_TELEPORT_ACK now logs server-sent coordinates AND current
position at WARNING level (was LOG_INFO, invisible in log file).
Heartbeat position audit now logs every ~60 heartbeats (~30s) at
WARNING level to trace position drift before rogue teleports.
Emote animations: fix DBC chain for /laugh, /flirt, /sleep, /fart, /stink.
Previously all emotes with AnimID=0 used emoteRef as animId (wrong DBC
record IDs). Now resolves through Emotes.dbc properly, with per-emote
overrides for emotes whose DBC chain yields 0. Adds Emotes.dbc load
failure warning and diagnostic logging.
WMO culling: skip portal culling when camera is outside all groups (fixes
vanishing Stormwind ground tiles). Also handle indoor/outdoor AABB overlap
by showing all groups when position is in both indoor and outdoor AABBs.
Transport: clear ONTRANSPORT flag and transport state when transport not
found, preventing stale transport data from teleporting player to map
origin. Add area trigger safety net near (0,0,0) on Eastern Kingdoms.
When on-transport flag is set but the transport isn't tracked by
TransportManager, getPlayerWorldPosition() returns localOffset (a small
relative value) as a world position. This overwrites movementInfo with
near-zero coordinates, teleporting the player to map origin on Eastern
Kingdoms (Alterac/Hillsbrad area). Add transport existence checks in
sendMovement() and getComposedWorldPosition() before composing position.
Add game_interfaces.hpp with five narrow domain contracts that GameHandler now
publishes to its domain handlers, replacing the previous friend-class anti-pattern.
Changes:
- include/game/game_interfaces.hpp (new): IConnectionState, ITargetingState,
IEntityAccess, ISocialState, IPvpState — each interface exposes only the state
its consumer legitimately needs
- include/game/game_handler.hpp: GameHandler inherits all five interfaces;
include of game_interfaces.hpp added
- include/game/movement_handler.hpp: remove `friend class GameHandler`; add
public named accessors for previously-private fields (monsterMovePacketsThisTickRef,
timeSinceLastMoveHeartbeatRef, resetMovementClock, setFalling, setFallStartMs)
- include/game/spell_handler.hpp: remove `friend class GameHandler/InventoryHandler/
CombatHandler/EntityController`; promote private packet handlers (handlePetSpells,
handleListStabledPets, pet stable commands, DBC loaders) to public; add accessor
methods for aura cache, known spells, and player aura slot mutation
- src/game/game_handler.cpp, game_handler_callbacks.cpp, game_handler_packets.cpp:
replace direct private field access with the new accessor API
(e.g. casting_ → isCasting(), monsterMovePacketsThisTick_ → ...ThisTickRef())
- src/game/inventory_handler.cpp, combat_handler.cpp, entity_controller.cpp:
replace friend-class private access with public accessor calls
No behaviour change. All 13 test suites pass. Zero build warnings.
SMSG_GM_MESSAGECHAT has an extra gmNameLen+gmName field after the
standard header that was causing misaligned parsing for non-whisper GM
messages. Strip the GM name before forwarding to the regular parser.
Also improve party state notifications (show member names on join) and
add INFO-level logging for all incoming/outgoing chat messages and
WARNING-level logging for group invite/accept packets to diagnose
recurring phantom party state issues.
Auto-reply was sent on every incoming whisper with no dedup, causing
infinite loops when both players had auto-reply enabled. Now tracks
which senders have been replied to and only sends one auto-reply per
sender per AFK/DND session.
The area trigger system was temporarily moving the player to the trigger
center and sending a heartbeat before firing CMSG_AREATRIGGER. This told
the server the player was at a different location, causing unexpected
teleports (e.g. Stormwind to Hillsbrad). Just send the area trigger
packet directly — the player is already inside the trigger radius.
Was reading field 2 (InstanceType) which fell back to field 1 (internal
name like "Azeroth"). Field 4 has the localized display name ("Eastern
Kingdoms").
getMapName() returned empty for mapId 0 (Eastern Kingdoms) due to an
early return guard. Remove it since 0 is a valid map ID.
queryGuildInfo() required IN_WORLD state but the character screen is at
CHAR_LIST_RECEIVED. The server accepts CMSG_GUILD_QUERY before login,
so just check for a connected socket.
Load area names from AreaTable.dbc (canonical zone names) in addition
to WorldMapArea.dbc so zone IDs from SMSG_CHAR_ENUM resolve to names.
Use lookupGuildName() to query and display guild names instead of raw
guild IDs.
The decomposition PRs moved mail state to InventoryHandler but the GO
interaction code still set stale GameHandler fields. Add openMailbox()
on InventoryHandler and forward from GameHandler so the correct
mailboxGuid_/mailboxOpen_ are set and refreshMailList() works.
Calculate repair costs client-side using DurabilityCosts.dbc and
DurabilityQuality.dbc. Block repair when player can't afford it and
only apply optimistic durability/gold updates when cost is verified.
Show repair cost next to the Repair All button in the vendor window.
MAX_CULL_INSTANCES was 16384 but game object instances were allocated
at indices 20000+, beyond the cull buffer. Increased to 65536.
Also compute fallback boundRadius from vertex extents when M2 header
reports 0, and seed bones for global-sequence-only animated models.
handleValuesUpdate silently dropped VALUES updates for item GUIDs not in
entityManager, causing repair-all durability changes to be lost. Fall
through to updateItemOnValuesUpdate for items tracked in onlineItems_.
The minimum floor (3.0 for sphere radius, 4.0 for box dimensions) was
inflating narrow triggers like AT 5711 (boxWidth 1.06 → 4.0), causing
false area trigger fires near the Stormwind AH and unexpected teleports.
M2 GPU instancing
- M2InstanceGPU SSBO (96 B/entry, double-buffered, 16384 max)
- Group opaque instances by (modelId, LOD); single vkCmdDrawIndexed per group
- boneBase field indexes into mega bone SSBO via gl_InstanceIndex
Indirect terrain drawing
- 24 MB mega index buffer (6M uint32) + 64 MB mega vertex buffer
- CPU builds VkDrawIndexedIndirectCommand per visible chunk
- Single VB/IB bind per frame; shadow pass reuses mega buffers
- Replaced vkCmdDrawIndexedIndirect with direct vkCmdDrawIndexed to fix
host-mapped buffer race condition that caused terrain flickering
GPU frustum culling (compute shader)
- m2_cull.comp.glsl: 64-thread workgroups, sphere-vs-6-planes + distance cull
- CullInstanceGPU SSBO input, uint visibility[] output, double-buffered
- dispatchCullCompute() runs before main pass via render graph node
Consolidated bone matrix SSBOs
- 16 MB double-buffered mega bone SSBO (2048 instances × 128 bones)
- Eliminated per-instance descriptor sets; one megaBoneSet_ per frame
- prepareRender() packs bone matrices consecutively into current frame slot
Render graph / frame graph
- RenderGraph: RGResource handles, RGPass nodes, Kahn topological sort
- Automatic VkImageMemoryBarrier/VkBufferMemoryBarrier between passes
- Passes: minimap_composite, worldmap_composite, preview_composite,
shadow_pass, reflection_pass, compute_cull
- beginFrame() uses buildFrameGraph() + renderGraph_->execute(cmd)
Pipeline derivatives
- PipelineBuilder::setFlags/setBasePipeline for VK_PIPELINE_CREATE_DERIVATIVE_BIT
- M2 opaque = base; alphaTest/alpha/additive are derivatives
- Applied to terrain (wireframe) and WMO (alpha-test) renderers
Rendering bug fixes:
- fix(shadow): compute lightSpaceMatrix before updatePerFrameUBO to eliminate
one-frame lag that caused shadow trails and flicker on moving objects
- fix(shadow): scale depth bias with shadowDistance_ instead of hardcoded 0.8f
to prevent acne at close range and gaps at far range
- fix(visibility): WMO group distance threshold 500u → 1200u to match terrain
view distance; buildings were disappearing on the horizon
- fix(precision): camera near plane 0.05 → 0.5 (ratio 600K:1 → 60K:1),
eliminating Z-fighting and improving frustum plane extraction stability
- fix(streaming): terrain load radius 4 → 6 tiles (~2133u → ~3200u) to exceed
M2 render distance (2800u) and eliminate pop-in when camera turns;
unload radius 7 → 9; spawn radius 3 → 4
- fix(visibility): ground-detail M2 distance multiplier 0.75 → 0.9 to reduce
early pop of grass and debris
Drop PROT_EXEC from warden module mmap when using Unicorn emulation
(not needed — module image is copied into emulator address space). Use
MAP_JIT on macOS for the native fallback path.
Add --listfile option to asset_extract and SFileAddListFileEntries
support for resolving unnamed MPQ hash table entries from external
listfiles.
- Enable alpha-to-coverage on alphaTestPipeline for smooth hair edges
when MSAA is active (both init and recreatePipelines paths)
- Shader uses fwidth()-based alpha rescaling for clean coverage
- Skip group 17/18 geosets (DK/NE eye glow) when no geoset filter is
set — prevents blue eye glow on all NPCs
- Add missing <libgen.h> include for dirname() on Linux
Whisper sender name may not be in the player name cache when the packet
arrives. Store the sender GUID and lazily resolve the name from the
cache in getLastWhisperSender(). Also backfill lastWhisperSender_ when
the SMSG_NAME_QUERY_RESPONSE arrives.
- alphaTestPipeline_ uses blendDisabled() so surviving pixels are fully
opaque (was blendAlpha, causing hair to blend with background)
- Remove alphaCutout from alphaTest condition — opaque materials like
capes no longer alpha-test just because their texture has an alpha
channel
- Two-pass batch rendering: opaque (blendMode 0) draws first to
establish depth, then alpha-key/blend draws on top
Bone SSBO buffers were allocated for MAX_BONES (240) entries but only
the first numBones were written. Uninitialized GPU memory in the
remaining slots caused vertex spikes when any bone index exceeded the
model's actual bone count.
Also send MSG_MOVE_STOP before spell casts so the server doesn't reject
cast-time spells (e.g. hearthstone) with "can't do that while moving".
Global sequence bones (hair, cape, physics) need time values spanning
their full duration (up to ~968733ms), but animationTime wraps at the
current animation's sequence duration (~2000ms for walk). This caused
vertex spikes projecting from fingers/neck/ponytail as bones got stuck
in the first ~2s of their loop. Add a separate globalSequenceTime
accumulator that is not wrapped at the animation duration.
buildDefaultPlayerGeosets() was inserting all submeshIds 0-99 into
activeGeosets, showing every hair variation simultaneously. Now uses
the hairGeosetMap_ (from CharHairGeosets.dbc) to select only the
correct hair scalp geoset for the player's race/sex/style, matching
the existing NPC geoset filtering logic in EntitySpawner.
Hair, cape, and other physics bones use global sequences (continuously
looping timers independent of the character's current animation). The
character renderer was ignoring globalSequence entirely, causing these
bones to fall back to identity transforms and produce deformed/spiked
hair geometry. Added resolveTrackTime() to wrap global sequence time
correctly, matching the M2 renderer's existing behavior.
M2 vertex bone indices are indices into boneLookupTable, not direct bone
array indices. Without remapping, vertices weighted to higher bone
indices (cloak, cape, hair) get the wrong bone transform, causing
vertices to project wildly outward from the character.
Wait on in-flight upload batch fences at the start of each frame and
insert a memory barrier (transfer→fragment shader) so the graphics
queue sees completed layout transitions from the transfer queue.
Fixes VK_IMAGE_LAYOUT_UNDEFINED validation errors for freshly loaded
textures.
