stfx

Was fixed more generic in [11690] + [11692]. Thanks

Renaming the sd2 dll is not needed - was it ever? Also wont work on Unix and is really not needed. Thanks for your work though

Building on Mac works fine with the new cmake build system added in [11167] - also sockets are handled with ACE now

No idea when this was fixed but compile on VS 2010 works now. Thanks nonetheless

Was fixed in mangos in [10302] and sd2 in [1752] + [2271]

This was fixed in [11783] by letting scripting library decide if an encounter is in progress in instances. Thanks nonetheless

PCH on Unix was added in [11167] with the new cmake build system. Thanks nonetheless

another problem related to fishing, from a well know wiki:

As of 3.1, there is no longer a minimum skill requirement to fish in any zone, and all casts will catch something and award a chance for a skill-up. However, if your skill level is too low for a zone, you will catch mostly vendor trash items. The skill level required to guarantee "no junk" catches is equivalent to the old "no get away" level.

This was fixed in [10989].

The original fishing skillup formula was still not changed though

Was added in [10620]. Thank you

Alternative fix added in [11783] for master and backported in [s1366] for One. Thanks

1) ACE has been updated to 5.8.3 in [10874] and we try to not modify external libraries if they dont break things

2) Likely outdated custom code

3) If the GO fix was proper it was probably added by now - if not Schmoozerd probably still has it around

Well patch link is broken but I move it to rejected based on given info and age.

Please never combine different non-related things in a single patch. Use seperate patches and one thread per patch, thanks

Added by zergtmn himself in [11636]

Cmake file has been added with vmap3 in [10164]

ACE has been updated to 5.8.2 in [10643] and 5.8.3 in [10874].

Didn't see this until now. Fortunately it was already correctly added like that in [s1419]. Thank you

blizzlike?

"I noticed that too. Interestingly it can also be used while mounted causing your player model to disappear, and summoning a mount while 'dispersed' cancels the dispersion effect."

"When using dispersion on a mount, your character dissppears and simply leaves the mount on its own. You don't appear like the shadowy figure you normally would, you totally vanish."

"It works even after you dismount, so you can be completely invisible for the duration."

http://www.own3d.tv/video/25112/Shadow_Priest_-_Dispersion_Glitch

Now that this is implemented in Zero I wonder about One and master. Is the current behaviour intentional? was it changed during wotlk or is it simply a bug?

I would appreciate any info, thanks

1. Please read the posts/patch before you post a solution as its essentially the same as yours which was the same as TC's old method.

2. If you want to contribute it would be awesome if you could push the patches Shauren mentioned to your repo with debug info as how many more packets are sent of old version vs. new version

Thanks

What bug does the patch fix? What features does the patch add?

Fixes Devastate not causing the Sunder Armor Effect.

For which repository revision was the patch created?

[s1419] (mangos-one)

Who has been writing this patch? Please include either forum user names or email addresses.

Me

Code is taken from master - the only difference is that on One there is no single Sunder Armor Effect spell which only causes the effect. Therefore we need to get the highest Sunder Armor spell from player spell book and cast it on enemy.

The issue I have here though is that it attacks the enemy twice - dealing physical damage twice (once for Devastate and once for triggered Sunder Armor) and also playing the animation twice. Is that blizzlike? Maybe someone more experienced with the spell system and/or warriors at 2.4.3 can take a look at it

https://gist.github.com/1303471

Meaning that 0 (BG_BRACKET_ID_FIRST) is a bracket on One but on master it actually is none, because first id is 1 in PvpDifficulty.dbc - not sure about that though.

To correct myself this is actually not true for master. On master a bracket_id of 0 is also an actual bracket because the ids in PvpDifficulty.dbc are between 0 and 15. Therefore setting the initial m_BracketId = BG_BRACKET_ID_FIRST is incorrect!

This is the patch for master: https://gist.github.com/1301721

The patch for One will be the same after its been added to master repo.

Added in [s1419] with fixed percentage of damage. Thanks to Nighoo for testing it.

Thanks for your research. I spent a few hours trying to find the correct fix and came up with the following:

https://gist.github.com/1299877

This is what I got so far:

At first the template bgs are being initialized (1 template for each bg map type).

When initializing a bg on One and Zero we use MAX_BATTLEGROUND_BRACKETS for default m_BracketId which is used as a marker for us to see if that bg is a template or not.

When removing the bg templates, DeleteClientVisibleInstanceId is being called and tries, as you already found out, to erase an entry which an index out of the boundary of the array.

Now setting m_BracketId to BG_BRACKET_ID_FIRST as it is done in the master is probably not such a good idea as on master the bracket id actually is a pointer to dbc structure associated with bg copy, in One on the other hand this is the id itself. Meaning that 0 (BG_BRACKET_ID_FIRST) is a bracket on One but on master it actually is none, because first id is 1 in PvpDifficulty.dbc - not sure about that though. See this commit for more details: https://github.com/mangos/mangos/commit/b9712d36d2c285be14cf7bb9bd5cc1f4c9c58cf3

So yeah please try the patch at the top of this post, thanks

Probably fixed as of TBCDB 1.2.0: http://udb.no-ip.org/index.php/topic,12793.0.html

I see, I didnt know that. Thanks, I cant believe I didnt think about this but I am not sure if this is not overkill as the server will now send the updated fields everytime they quickly changed to one value and back to the original again as it seems to only cause an issue on a very rare cases like shapeshifting - though maybe thats blizzlike... hm

Yea, setting it to m_uint32Values_mirror[index] = m_uint32Values[index] + 1; seems like a good idea (taken from mangosr2)