johnholiver

Well, now there's something strange... the /played command is being executed everytime. Even without player's consent =P

I would like to add that it seems that Seal of Vengeance is currently bugged.

http://www.getmangos.eu/community/showthread.php?p=78709#post78709

This problem is about execution of mangos-worldd but i just had this on linux:

Initialize data stores...
[mangos-worldd: ../../../../src/shared/Database/dbcfile.h:73: const char* DBCFile::Record::getString(size_t) const: Assertion `stringOffset < file.stringSize' failed.

I copied the DBC folder from my Windows extract, is that a problem? Should i re-extract in linux to work? I tried also to use an older rev [7252] using git-reset --hard.

Using Ubuntu 8.10

I'm having several crashes a day because of this method and reading the code from HandleGroupLeaveOpcode til the RemoveAllInvites i came to some conclusions that might help solving this problem.

I did this topic to discuss the matter, to see if there is some sense on what i think it is.

- RemoveAllInvites has little relation with all the code that comes before it, the variables used (m_invitees) seem to be untouched until this call.

- m_invitees is a list of Player* only.

- Player::SetGroupInvite(NULL) just modify Player.m_groupInvite variable setting it to NULL.

So my (weak) conclusion is that, been as simples as this call is, the problem might be with one Player* on m_invitees... for example, what happens with the player doesn't exist anymore for some reason? (Player* now points to somewhere that is no longer a player) And i think that would justify an ACCESS_VIOLATION.

Maybe the destruction of a Player (logging out, rejecting party, i dunno) in somewhere isn't removing him from Group.m_invitees.

Anyways, if i'm right on my conclusions then a check to guarantee that itr* has a good and nice player* would solve this? My opinion is that it would, but i would like to listen to other devs their opinion.

Thanks for reading til here... below i quoted free-back that reported that crash (again, 2 or 3 times others had reported before) and his crash dump.

rev 7092 + sd2 844 + udb 366

Revision: 2009-01-16 07:53:29 7092 b88bfbdcefbe457b28af34d5f750a148d7155807

Date 17:1:2009. Time 12:14

//=====================================================

*** Hardware ***

Processor: Intel® Xeon MP CPU 2.80GHz

Number Of Processors: 4

Physical Memory: 3144628 KB (Available: 1656008 KB)

Commit Charge Limit: 4194303 KB

*** Operation System ***

Microsoft Windows Server 2003 Enterprise Edition Service Pack 2 (Version 5.2, Build 3790)

//=====================================================

Exception code: C0000005 ACCESS_VIOLATION

Fault address: 0048CF6E 01:0008BF6E I:\\servers\\wowserv\\Mangos\\mangosd.exe

Registers:

EAX:3B7255E0

EBX:08947C08

ECX:34FADB14

EDX:3B7255E0

ESI:00000000

EDI:00000000

CS:EIP:001B:0048CF6E

SS:ESP:0023:0FF3F75C EBP:0FF3F794

DS:0023 ES:0023 FS:003B GS:0000

Flags:00010216

Call stack:

Address Frame Function SourceFile

0048CF6E 00000000 Group::RemoveAllInvites+7E

0048DAB9 00000000 Group::Disband+289

0048D5D6 00000000 Group::RemoveMember+356

004C0383 00000000 Player::RemoveFromGroup+23

0070864C 00000000 WorldSession::HandleGroupLeaveOpcode+11C

Tnx, i'll report as a bug then

Hi guys,

Today i have a nice doubt =)

I was playing with an ally and i was facing an Horde Lock in a Horde zone that had his PvP disabled. When he saw me he invoked an imp and the imp started to attack me but the lock remained with his PvP disabled so i couldn't attack him, only the imp... is that ok, or is that a bug? I'm asking cause i never played on offi and I thought that was pretty unfair =P

Thanks in advance

lol ^^ tnx!! =]

I also started to look elsewhere and found one for windows only... it uses windows.h

SYSTEMTIME now;

GetSystemTime(&now);

That's what i like about MaNGOS makes me look through the nastiest stuff lol

Hi all,

I'm trying to model/trace clients workload on my emulator and I'm using WorldLog for this. The thing is that in order to do obtain response time (and many other metrics, I'm solving a analytical model using Little's Law and bottleneck analysis, maybe MVA) i would like to put a time stamp on every entry of the WorldLog... well... that's done, but time.h lib and its function time() gives a time granularity of seconds and i want milliseconds to make my results more accurate and valid. So, any ideas about how to get time in milliseconds?