http://bugs.winehq.org/show_bug.cgi?id=20669
--- Comment #7 from Jarod wine@faltine.com 2010-02-25 03:52:36 --- Created an attachment (id=26471) --> (http://bugs.winehq.org/attachment.cgi?id=26471) Backtraces 1.1.18 to 1.1.39
I am not sure it is the same bug as #19659 but it might be related: it looks like there are several crashes that might come from the following patch:
commit 014c4bfc70a4d4e60f033d579d1be13a46f65170 Author: Stefan Dösinger stefan@codeweavers.com Date: Thu Apr 9 18:40:57 2009 +0200
wined3d: Save some memory in vertex buffers.
I prefer to post here because it is exactly the symptoms I experience: certain objects found in game cause crashes. It never crashes randomly shortly after entering the game: there is always an object or a character coming into view, and I can play all night if I stay away from such objects. I never experienced a crash with version 1.1.18 yet. The patcher doesn't work for me with this version, but if I patch using 1.1.39 I can get into the game by launching EverQuest2.exe directly. I haven't noticed any difference in game between versions yet, apart from the crashes. The hardware I use is a nvidia nForce chipset 7050 with integrated video. The graphical options in game are set for "extreme performance" because it is the only setting offering a fluid game on this hardware, be it on windows or linux. I might not be experiencing some crashes other people with better hardware and better settings experience.
I made the regression test between 1.1.18 and 1.1.19 as Scott Murray did but I specifically kept testing only one thing: crafting stations, because it is easy to get in the situation and allow to get rid of any randomness in the testing. I also tested with 1.1.39 and the feb, 24 version from the git repository. I noticed it is the action of crafting, or witnessing someone who crafts, that gives a crash 100% of the time. I sometimes can look at crafting stations for a time without crashing.
The backtrace I get is always the same, as far as I can tell (see attached backtraces.tgz). In all the versions I tested there are those same lines, and a few others: wine: Unhandled page fault on read access to 0xffffffff at address 0xa768f7 (thread 0009), starting debugger... Unhandled exception: page fault on read access to 0xffffffff in 32-bit code (0x00a768f7).
My conclusion is that the crash I get with crafting stations hasn't changed since this patch. The other crashes might have the same cause, or not.
What I plan to do is: - stress test 1.1.18 to confirm there is no crash, even while raiding with lots of people or crafting for a long time - test the first buggy version in all situations other than crafting and record the backtraces and the object who triggered it - try to provide +d3d,+tid traces, but it is not very easy for me at is slows the machine down to a crawl. - do whatever the devs ask :)