refactor(Core/Spells): QAston proc system#24233
Conversation
Port TrinityCore's modular spell proc handling system that replaces hardcoded proc handlers with database-driven spell_proc entries and AuraScript hooks. Changes: - Add spell_proc table with 300+ proc entries - Add DoCheckProc, PrepareProc, OnEffectProc hooks to SpellAuras - Convert class proc handlers to AuraScripts (DK, Druid, Hunter, Mage, Paladin, Priest, Rogue, Shaman, Warlock, Warrior) - Add 35+ item/trinket proc scripts - Add pet proc scripts (Hunter) TODO: - Implement remaining boss scripts (spell_deathbringer_blood_beast_blood_link, spell_putricide_ooze_tank_protection, spell_twisted_reflection, spell_uk_second_wind, spell_xt002_321_boombot_aura) - Implement misc scripts (spell_q13413_wyrmrest_skytalon_ride_periodic) - Test the new proc system Closes azerothcore#9177 Co-Authored-By: QAston <[email protected]> Co-Authored-By: joschiwald <[email protected]>
- Add DisableEffectsMask to SpellProcEntry struct and spell_proc table - Update LoadSpellProcs to load and auto-generate DisableEffectsMask - Update GetProcEffectMask to use DisableEffectsMask for effect filtering - Add DamageInfo constructor with SpellMissInfo for hitMask computation - Add m_hitMask to HealInfo class with GetHitMask/AddHitMask methods - Update spell damage, healing, and proc flows to use hitMask - Update PROC_SPELL_PHASE_CAST and PROC_SPELL_PHASE_FINISH procs - Remove legacy createProcExtendMask function Based on TrinityCore commits: - TrinityCore/TrinityCore@e641d0c - TrinityCore/TrinityCore@2ff8550 - TrinityCore/TrinityCore@016b31a Co-Authored-By: ariel- <[email protected]>
- Move proc checks from old system to Aura::GetProcEffectMask (returns effect mask) - Add AuraEffect::CheckEffectProc for per-effect proc validation - Add PROC_ATTR_REQ_SPELLMOD validation in LoadSpellProcs - Update SPELL_PROC_FLAG_MASK to include periodic and trap activation flags - Update DONE_HIT_PROC_FLAG_MASK to include trap activation - Sync spell_proc table with TrinityCore (848 entries) - Add proc cooldown on item equip (AddProcCooldown in EquipItem) - Port missing AuraScript proc handlers for class spells - Port boss/dungeon proc scripts (Saurfang, Putricide, XT-002, etc.) - Remove stale SpellProcEventEntry references - Auto-generate default proc entries for spells with proc flags Based on TrinityCore commits: - e641d0c7d71 Core/Spells: Implementation of QAston proc system - 2ff8550 Core/Spells: AuraScripts coverage - 016b31a Core/Spells: Proc cooldown on equip Co-Authored-By: ariel- <[email protected]>
- spell_dru_eclipse: Add 30-second internal cooldown for lunar/solar procs - spell_pri_body_and_soul: Use EFFECT_1 for Abolish Disease proc chance - spell_pal_sheath_of_light: Use GetEffectiveHeal() and GetMaxTicks() - spell_item_persistent_shield: Add shield strength comparison check - Remove obsolete 'reload spell_proc_event' command from database
Add comprehensive unit test suite for the spell proc system: - 75 SpellProc tests covering CanSpellTriggerProcOnEvent logic - Data-driven tests validating all 869 spell_proc entries - Integration tests for proc flag combinations and hit masks - Mock infrastructure (UnitStub, AuraStub, SpellInfoTestHelper, etc.) Tests cover: - Proc flag matching (melee, spell damage, heal, periodic, kill) - Hit mask filtering (crit, dodge, parry, block, absorb) - Spell family name/flag matching - Spell type and phase mask validation
- Add !IsTriggered() check in Spell::Cast() to skip CAST phase procs for triggered spells (e.g., periodic damage effects like Blizzard ticks) - Add PROC_ATTR_REQ_SPELLMOD check in GetProcEffectMask() to require spell to be affected by aura's spellmod before proccing - Add PROC_ATTR_REQ_MANA_COST check in CanSpellTriggerProcOnEvent() to filter procs based on mana cost requirement - Fix Arcane Potency charges being consumed by channeled spell ticks This fixes issues where buffs like Arcane Potency and Clearcasting were incorrectly consumed by periodic damage effects instead of only on player-initiated spell casts.
TAKEN procs (like Backlash, Redoubt) were being checked on the attacker instead of the target. This caused procs that should trigger when you get hit to not work correctly. Changed TriggerAurasProcOnEvent to call GetProcAurasTriggeredOnEvent and TriggerAurasProcOnEvent on actionTarget for target procs instead of on the attacker (this).
…L procs Direct heals like Healing Wave were incorrectly classified as PROC_SPELL_TYPE_DAMAGE instead of PROC_SPELL_TYPE_HEAL because the spellTypeMask calculation only checked for PROC_EX_INTERNAL_HOT (periodic heals) but not for direct heals. This caused procs like Tidal Force (55198) that require SpellTypeMask=2 (heal) to never trigger on direct healing spell crits. Fix: Also check if healInfo is present to identify direct heals.
Sheath of Light was checking GetEffectiveHeal() which returns 0 when overhealing. The talent should proc on any critical heal, using the base heal amount before overheal is calculated. Changes: - Use GetHeal() instead of GetEffectiveHeal() in CheckProc and HandleProc - Update script to match TrinityCore's OnEffectProc pattern - Add spell_proc entry for Sheath of Light from TrinityCore
The spell should heal whoever the marked enemy attacks, not the marked enemy itself. Use eventInfo.GetActor() (marked enemy) to cast heal on eventInfo.GetActionTarget() (attack victim).
…ration Move CAST phase proc to after handle_immediate() so spell mods (like Missile Barrage's duration reduction) are applied before the aura is consumed by the proc. Previously, the proc was firing before handle_immediate(), causing the aura to be removed and its spell mods deleted before the duration modification could be applied.
Core changes: - Fix TAKEN procs being checked on wrong unit - Fix direct heals not triggering PROC_SPELL_TYPE_HEAL procs - Fix Sheath of Light proc to use base heal amount - Fix Mark of Blood healing wrong target - Fix Missile Barrage not reducing Arcane Missiles duration Spell script fixes: - Paladin: Refactor proc handlers for Sheath of Light, Sacred Shield - Priest: Fix Vampiric Embrace proc handling - DK: Fix Mark of Blood target selection - Rogue: Fix Cut to the Chase spell registration SQL fixes: - Fix Cut to the Chase registered to wrong spell (-35541 -> -51664) - Fix Decimation DisableEffectsMask to disable EFFECT_1 TODO: Backlash (34936) buff not consumed on Shadow Bolt cast despite correct spell_proc entry (Charges=1, AttributesMask=8). Needs C++ investigation of PROC_ATTR_REQ_SPELLMOD charge consumption path.
At CAST phase, spellTypeMask was calculated from damage/heal info, but no damage has occurred yet. This caused procs like Backlash to fail validation because they expect SpellTypeMask=DAMAGE but received SpellTypeMask=NO_DMG_HEAL. Use PROC_SPELL_TYPE_MASK_ALL for CAST phase procs to allow validation based on spell info rather than actual damage/heal values. Fixes Backlash (34936) and similar spellmod-based procs not consuming charges on spell cast.
Remove redundant positive spell ID entries (18094, 18095) for Nightfall that were conflicting with the negative ID entry (-18094). The negative ID already covers all ranks via spell_warl_glyph_of_corruption_nightfall. This fixes Shadow Trance not proccing from Corruption ticks.
… Anger The proc handler only added a Mote of Anger stack on first application, but did nothing when the aura already existed with fewer than max stacks. Changes: - Always cast Mote of Anger to add/stack on each proc - Check stack count after adding, consume at 8 (normal) or 7 (heroic) - Add 50% chance for off-hand attack when dual wielding - Clear motes when trinket aura is removed
Several proc scripts were accessing eventInfo.GetDamageInfo()->GetDamage() without verifying that GetDamageInfo() returns a valid pointer, which could cause crashes. - spell_dk_necrosis: Add CheckProc for damage validation - spell_dk_unholy_blight: Add CheckProc for damage validation - spell_dru_t10_balance_4p_bonus: Add damageInfo null check in HandleProc
Core changes: - Change AuraScript::Remove() default from AURA_REMOVE_NONE to AURA_REMOVE_BY_DEFAULT - Fix proc system phase/type mask handling in SpellAuras Proc scripts added/fixed: - spell_sha_lightning_shield (Static Shock now works) - spell_dk proc scripts refactored - spell_hun proc scripts updated - spell_rog proc scripts (Turn the Tables, Master of Subtlety) Tested and verified fixes for GitHub issues: - azerothcore#5967 Tiny Abomination in a Jar - azerothcore#16062 Turn the Tables buff targeting - azerothcore#13330 Ignite extending combat - azerothcore#21401 Druid shapeshifting proccing Omen of Clarity - azerothcore#19709 Blazing Speed proc rate - azerothcore#11462 Arcane Concentration/Blizzard ticks - azerothcore#23326 Thunderstorm Clearcasting - azerothcore#22279 Righteous Weapon Coating ranged - azerothcore#9400 Blade of Eternal Darkness crit - azerothcore#19793 Darkmoon Card: Crusade stacking - azerothcore#24146 Seal of Command + Ravager - azerothcore#12115 Replenishment threat - azerothcore#24116 Flametongue rank stacking - azerothcore#22427 Cleave + Sweeping Strikes - azerothcore#17564 T5 Crystalforge 4-set - azerothcore#10993 Dislodged Foreign Object - azerothcore#19807 Priest T5 Sadist buff
…ory safety - Add 3 new SchoolMask filter tests to SpellProcIntegrationTest - Fix ProcScenarioBuilder to use shared_ptr instead of raw pointers to prevent double-free issues when builder objects are copied
github-actions
Bot
added
DB
blinkysc
added
the
Requires WIKI Update
|
Does this PR also correctly fix the Rogue T5 set bonus (4-piece: “Your attacks have a chance to make your next finishing move cost no energy”)? |
This system does handle the proc for it. Configured for 1ppm |
… ICD ApplyEquipCooldown was using procEntry->Cooldown (the trinket's own ICD) instead of the standard 30s equip cooldown, causing trinkets like Greatness (45s) and Mirror of Truth (50s) to take longer than expected before first proc.
The cheat cooldown flag only bypassed regular spell cooldowns and GCD but not proc internal cooldowns on the Aura system. Trinket procs and equip cooldowns now respect the CHEAT_COOLDOWN player flag.
|
While I was testing this PR, I noticed a discrepancy in the PPM calculation in Player::CastItemCombatSpell for Druids and other shapeshifters. Replacing proto->Delay with GetAttackTime(attType) should fix this, since GetAttackTime returns the base attack speed of the form (e.g., 1000ms for Cat) and is independent of Haste. Currently, the code uses proto->Delay from the ItemTemplate, which reflects the static weapon speed (e.g., 3.6s). For Feral Druids, this leads to incorrect scaling as their proc rates should be based on their form's base speed, as confirmed by WarcraftLogs and WoWSims models. Proposed Code Refactor: By implementing this change, the behavior for weapon enchantments will be aligned with the rest of the proc system refactored in this PR, ensuring that PPM is handled consistently across all types of procs. Note: |
This does seem to line up with WoWsims. It would be divergent from TC as well. Let me research just a bit more, but do think this would be correct if so. If you have more links or research please share. Overall the unit test still will show the same PPM though, this simply would change the "nuances" of how it procs and seems would only specifically affect feral likely. |
|
To be honest, I'm basing this on my analysis of my own WotLK Classic logs and simulation models, as they are the most accurate data we have today. Edit: |
|
GetAttackTime is the correct way to get the basespeed of the attack. However, that spot manages the proc chance of item combat spells so it will have no impact on omen. Most procs go through Aura::CalcProcChance |
…mbatSpell Enchant PPM in CastItemCombatSpell used proto->Delay (static weapon speed) instead of GetAttackTime(attType). For shapeshifted druids this used the weapon delay (e.g. 3.6s staff) instead of the form speed (1.0s Cat, 2.5s Bear), inflating enchant proc rates by up to 3.6x. The item spell PPM path already used GetAttackTime correctly. This aligns the enchant path to match. No behavior change for non- shapeshifted players since GetAttackTime returns proto->Delay for them.
|
@hennors yes you're correct =D thank you for your input! |
Changes Proposed:
This PR ports the QAston proc system from TrinityCore, which is a major refactor of how spell procs are handled. The key changes include:
Core Changes:
spell_proctable with enhanced filtering:SpellTypeMask,SpellPhaseMask,HitMask,AttributesMaskHandleDummyAuraProcswitch statements to individual spell scriptsBenefits:
This PR proposes changes to:
AI-assisted Pull Requests
Important
While the use of AI tools when preparing pull requests is not prohibited, contributors must clearly disclose when such tools have been used and specify the model involved.
Contributors are also expected to fully understand the changes they are submitting and must be able to explain and justify those changes when requested by maintainers.
AI Tools Used: Claude Code (claude-opus-4-5-20251101) thinking with azerothMCP
Issues Addressed:
SOURCE:
The changes have been validated through:
TrinityCore/TrinityCore@e641d0c
TrinityCore/TrinityCore@2ff8550
TrinityCore/TrinityCore@016b31a
https://github.com/TrinityCore/TrinityCore/commits/3.3.5?author=ariel-&before=d0fd8e683ca02f3d2d9ceefe55482a49aab59c6c+700&branch=3.3.5
Tests Performed:
This PR has been:
Tested ~400 proc scenarios including:
How to Test the Changes:
Use this gist https://gist.github.com/blinkysc/b6b388a62cb88efe987833ac340215e6 it should have primary coverage as well as comprehensive (every single spell_proc spell). These tests are generated by the MCP so some of them might be a little odd but I think testers should get the point on most.
Known Issues and TODO List:
How to Test AzerothCore PRs
When a PR is ready to be tested, it will be marked as [WAITING TO BE TESTED].
You can help by testing PRs and writing your feedback here on the PR's page on GitHub. Follow the instructions here:
http://www.azerothcore.org/wiki/How-to-test-a-PR
REMEMBER: when testing a PR that changes something generic (i.e. a part of code that handles more than one specific thing), the tester should not only check that the PR does its job (e.g. fixing spell XXX) but especially check that the PR does not cause any regression (i.e. introducing new bugs).
For example: if a PR fixes spell X by changing a part of code that handles spells X, Y, and Z, we should not only test X, but we should test Y and Z as well.