Howdy,
Got a bit bored over the weekend and updated my build script to build wine daily while I'm at work, then automatically run the test suite under different (currently 6) configurations.
Currently, it tests: no winedebug (ae-ub-810) Wine running under a virtual desktop (ae-ub-810-virtdesktop) # this one was broken in today's test...err, not run under a virtual desktop, fixed for tomorrow's. Needed a sleep 5s to get the virtual desktop setting saved before running the test suite. WINEDEBUG=+heap (ae-ub-810-heap) WINEDEBUG=+relay (ae-ub-810-relay) WINEDEBUG=+seh (ae-ub-810-seh) WINEDEBUG=+all (ae-ub-810-all)
It's got some interesting results: http://test.winehq.org/data/3fa05ae6a7c08f9534badfeffbe02db9e00ab47a/#group_...
Currently, several tests timeout because of it (not worried about those...besides, I just ordered parts for a quad core on newegg, so those may be gone soon).
The more interesting thing is that some test fail with memory corruption, etc. Most notably, crypt32:chain, crypt32:encode, crypt32:sip, d3d9:visual, ddraw:visual, oleaut32:tmarshal, riched20:editor, riched32:editor, and rpcrt4:cstub.
There are several crashes as well (netapi32:access, jscript:run, msi:msi) all with a similar error. I filed http://bugs.winehq.org/show_bug.cgi?id=17415 for that.
I filed a few other bugs for some of these issues, but not all of them...hopefully some of them will be fixed before I get the chance to :-p.
I'd appreciate if everyone took a couple minutes to look at any areas of the code they're familiar with and see if they can fix these failures.
I plan to test around with various optimizations, e.g., -O0, etc., since I recall at WineConf 2008 there were some failures only showing up under those configurations. If that shows anything consistent, I'll add them.