pasdVn

Ah, I see. I did not thought about that^^. But what du you mean by "use channelling target"? Which channeling target? Thanks for feedback :-)

Hey Guys, I discovered the following problem e.g. with "Penance" (http://www.wowhead.com/?spell=47758). Just some examples in short form, to understand what's that about: * target yourself and use "the hand" to cast penance at another friendly or unfriendly unit -> you will damage or heal yourself (ok, worked only in 3.0.9 because now you are abled to cast penance on yourself) * select a friendly or unfriendly unit and use e.g. the healbot (addon) to heal any of your party members -> you will heal whatever you have in your target selection currently * don't select anything and try the thing with the hand, healbot... whatever -> spell will channel but nothing triggers Problem: All periodic, aura triggered spells have the target type 77 (currently called TARGET_SINGLE_ENEMY). Auras triggering such spells are handeled as SingleEnemyTargetAura. They store the guid of the current selection of a player when this aura gets applied. This guid is used, to get the correct unit target in the triggering methods in Aura. Sollution: I discovered that class spells using this construct (so spells casted by players) always provide a second dummy aura, that is casted at the real, client-provided target. To get acces to the target I linked those dummy auras to the caster. Other of those auras, that are no class skills (so mainly casted by mobs or used by quest items) cast the triggering aura directly on their target. The triggered spells simply have the same target as the triggering aura in this case, so that a dummy aura is not needed for target selection. Update 30.07.09: I rechecked the advices of thenecromancer and made it possible , that different dummy auras can stack on the target now. I also added SPELL_AURA_PERIODIC_TRIGGER_SPELL, SPELL_AURA_PERIODIC_TRIGGER_SPELL_WITH VALUE. This makes it possible, that also periodic triggering auras, that are directly casted on the victim (not on the caster, mostly those are spells used by creatures, e.g. http://www.wowhead.com/?spell=60906), are possible to stack. http://github.com/pasdVn/mangos/commit/0210751587c9038ed9c40d428f364c55b961bca6.patch In adition I discovered, that the hunters "tame beast" spell, that also uses this spell mechanic, won't work with this patch, because in CheckCast() the target of the trigger spell is checked (what is still the caster himself before the target map is filled). I just moved this check to the original channeling spell. http://github.com/pasdVn/mangos/commit/62c1ba1f2110fea1dc6f3023180f2c8c7cbfa7e1.patch Update 16.02.2010: Updatet links. Patches rebased on [9397]. I'm not sure about the the second part with the tame pet effect. In fact it's a kind of hack, but as there is just one real spell with those effect it does not matter for now. Don't discovered how other implicit targets are handle in checkcast (target map is filled after checkcast() call). pasdVn

Dispersion (http://www.wowhead.com/?search=dispersion#talents) got an additional effect in 3.1.x (with a new spell... why ever!?). http://github.com/pasdVn/mangos/commit/6566ec2bbd3654921446b6522e9800ef835ffffe greetings, pasdVn

I also thought also about using the first aura for the first effect. But as basevalue of this aura is not that useful (they could simply add the the correct value there: 15, 20 25) and this aura has also classmasks, that fit to the priest renew spell I thought this is (maybe) a left, but not used part out of their developement (maybe this spell was originally planed to do some other things). But that is all only guess So thank you for reviewing :-)

Hmm, ok. At least the HandleAuraSchoolAbsorb() would look mor clean then^^ I had the idea with using a new damage effect type. Maybe the use is more clear then, when reading SpellDamageBonus(). I just did it with spelldamagebonus as before (maybe we could hardcode it again for power word shield if anybody has a proof, that it still uses healing bonus but not damage bonus). http://github.com/pasdVn/mangos/commit/25e66aef6b68a124bfbb3561694eb4e0689a7b6c The diff looks quite huge, but only because I shifted a code part right one tab. Edit: SPELLMOD_SPELL_BONUS_DAMAGE is added multiplicative in spelldamage bonus. I did it additive before. Don't know what is right. Edit2: Maybe this is the reason why blizz did not handle it as a spellmod... So we could could just hardcode that in HandleSchoolAbsorb() !?

Hmm, ok I see. I could swear that the other fix worked... seems as I am wrong :-/ But as I see before the proc flag patch it was also called with a unit target. Just wondered and was unsure, because it was explicitly coded there as a NULL pointer.

Great work Not much left to do for me^^ The sollution with the spellfamily flags is really great. I did not now that some of those dummy auras have SpellClassMask values (as they are used for spellmods). Maybe we could also think about to handle such "on aura remove" effects a bit more general in the future. E.g. "Glyph of Freezing Trap" works similar, but as I discovered, the ClassMask value does not fit to to the "freezing trap effect" :-/ Hmm, ok. Needs a bit more research, but that is off topic now... Updated the first with the new working/tested patch. Edit: Hmm, I'm motivated now^^ As this is nearly the same part of code here is also the dummy proc for "Glyph of PW: S": http://github.com/pasdVn/mangos/commit/cb372ac65fc14cc71608ff6fc59289304422be11 (again inspirated by maxxie)

Hmm, ok. I didn't concentrate on the sql part. But should not be any problem then to change this values easily. I think those are the values I took out of the code when I wrote this patch (although 10% are really few!?). If anybody has more reliable values for the coefficients of different absorb spells I could add them to the first post.

Hmm, maybe I understood the tooltip wrong. I thought, that the proc on on the target is "onCast" but not on removing the absorb aura... but maybe you are right. Edit: Ok, you are right. I read some more comments at wowhead. I'm going to change taht and will test it this time^^

Updated patch to apply it without merge problems.

Updated. But possible that it does not work. I don't have a 3.1.3 client here currently (and so no dbc's). I don't like this big hardcoded part on removing PW:S, but I don't see any other possibility for now. Don't forget to apply the sql part.

Hmm, sry. I overlooked that. But are you shure with your forula then? Would mean that you would profit more from steady shot if you are using a slower weapon, while the casting of steady shot is equal with all weapon speeds.

Yes, i can confirm this formula. Here is my version that includes a bugfix for steadyshot <-> autoshot (the hacky way, as I don't find any other sollution), if anybody likes it ;-) http://github.com/pasdVn/mangos/commit/72a89c836068d60ab104abcc35d6f62560e5ad72 Sorry for german commit message. I didn't made this commit for this forum originally ;-)

Hmm, git sems to interpret the tabulators I do in vc++ the wrong way :-/ Anyway fixed and updatet link.

Some spells differentiate between a friendly (unit-) target or unfriendly (e.g. penance, holy shock). This patch implements that feature. http://github.com/pasdVn/mangos/commit/87240d3944d254dfef79871d32886280004455ad Maybe this could be changed at some more places in the code where GetSpellMin/MaxRange() is used, but as there are only few spells using friendly range this should do it for now.

I changed the patch at the repo. Link and hash of the commit keeps the same. To reapply it, best would be to revert the old commit at your repo and than cherry-pick the new one (or however you are used to do it^^). Edit: Hmm sry. The commit hash and link changed... very strange. In fact this linked commit should not exist anymore... just a second... Edit2: Ok, link updated.

Thank you for feedback. I missed a type conversion. Maybe different compilers have/don't have problems with that (vc++ has). Fixed that and should work for all ranks now :-)

Not with a clean mangos core ;-)

When uploading this patch to github and rechecking it I noticed, that the mana, returning from absorb did not work (wrong conversion from float to int). I could swear that it worked before and I did not changed anything :mellow: Anyway, here comes an updated working version (see first post)!

Updated patch to work with the changes in [7782].

Fixed this bug and updated patch and links.

Updated all links and patches. I finally managed to change to an own github account . So there won't be problems with expired links anymore!

Ah allright, thx. Didn't recheck the current code in detail^^ Edit: Updatet first post.

Still actual. Maybe if there is done a lot at the distance calculation currently, any dev could have a look on this.

Reuploaded. Edit 25th may: Reuploaded again to github repo