[BUG] All Root-Auras can be removed due to damage

[Auntie Mangos]

Mangos Version: 10400

Custom Patches: none

SD2 Version: none

Database Name and Version: UDB392

How it SHOULD work: There are Spells which apply a root-aura, that shouldn't be broken due to dmg

How it DOES work: All root-auras are broken with dmg (and chance)

Suggested Fix: http://getmangos.eu/community/viewtopic.php?id=15283

Reason in mangos:

Unit::DealDamage
   if (!spellProto || !IsSpellHaveAura(spellProto,SPELL_AURA_MOD_FEAR))
       pVictim->RemoveSpellbyDamageTaken(SPELL_AURA_MOD_FEAR, damage);
   // root type spells do not dispel the root effect
   if (!spellProto || !(spellProto->Mechanic == MECHANIC_ROOT || IsSpellHaveAura(spellProto,SPELL_AURA_MOD_ROOT)))
       pVictim->RemoveSpellbyDamageTaken(SPELL_AURA_MOD_ROOT, damage);

* Called by DealDamage for auras that have a chance to be dispelled on damage taken. */
void Unit::RemoveSpellbyDamageTaken(AuraType auraType, uint32 damage)
{
   if(!HasAuraType(auraType))
       return;

   // The chance to dispel an aura depends on the damage taken with respect to the casters level.
   uint32 max_dmg = getLevel() > 8 ? 25 * getLevel() - 150 : 50;
   float chance = float(damage) / max_dmg * 100.0f;
   if (roll_chance_f(chance))
       RemoveSpellsCausingAura(auraType);
}

this shows, that all root-auras or fear auras can be broken with dmg.

This is not correct, especially for some creature-used special auras, like

Soulstorm - FoS, Bronjahm

this aura shouldn't get broken with dmg.

Unfortunately I didn't find any common thing of this spell and alike spells

(I did a search with NOT description LIKE "%may interrupt%" onto spells, to get a pack of candidates for non-interruption)

[laise]

Fear, Entangling Roots and Frost Nova have procflags and Soulstorm doesn't..I think thats what adds first 3 the ability to be removed from damage, so it shouldn't be explicit in DealDamage

[Schmoozerd]

hmm, this doesn't look too bad, however if the procflags are right, then quite a few spells have wrong description:

Spells with "%interrupt%" description are (for fear and root): 49, whereas only one (61385) has not the procflags

Spells without interrupt description (for root) 399, but from them have 60 spells the procFlags

Spells without interrupt description (for fear) 166, but from them have 7 spells the procFlags (5246-wowhead comments indicate that the desc. is missing, 19134, 19725, 35474)

[laise]

also note that spells like 46102 Spell Fury might have proc flags for different effect unrelated to root/fear removing

[Schmoozerd]

hmm, most of these spells have 664232 as procflag.

this is

PROC_FLAG_TAKEN_MELEE_HIT = 0x00000008, // 03 Taken damage from melee auto attack hit

PROC_FLAG_TAKEN_MELEE_SPELL_HIT = 0x00000020, // 05 Taken damage by Spell that use melee weapon

PROC_FLAG_TAKEN_RANGED_HIT = 0x00000080, // 07 Taken damage from ranged auto attack

PROC_FLAG_TAKEN_RANGED_SPELL_HIT = 0x00000200, // 09 Taken damage by Spell that use ranged weapon

PROC_FLAG_TAKEN_AOE_SPELL_HIT = 0x00002000, // 13 Taken AoE damage spell hit (not 100% shure unused)

PROC_FLAG_TAKEN_NEGATIVE_SPELL_HIT = 0x00020000, // 17 Taken negative spell (by default only on damage)

PROC_FLAG_ON_TAKE_PERIODIC = 0x00080000, // 19 Taken spell periodic (damage / healing, determined from 14-17 flags)

if these descriptions had some motivation - to use them would be convinient here

[Schmoozerd]

sr for double post, but I started a _draft_ how it might be possible to "fix" the problem:

http://paste2.org/p/973176

I don't know how to check if spell-dmg is caused by melee-spell, ranged-spell or AoE spell, and it surely will have some side-effects, but perhaps it is a step to the right direction to handle it

- However there is no spell which doesn't have all three of these procFlags, so this won't hurt

[laise]

why you dont use procFlags where they suppose to be used - in spell proc system?

[Schmoozerd]

do you suggest to remove the chain of DealDamage, remove auras at all?

sorry, as I don't know the code well enough, for me it was natural to modify these functions.

If you suggest that it should be handled entirely different, I won't be able to give serious help about this issue ;(

sorry for that, hope you will have the time and motivation to implement it at all (or likely even slower to explain what needs to be done)