CVS regression.

Gerard Patel gerard.patel at nerim.net
Sat Dec 22 18:32:20 CST 2001


At 02:30 PM 22/12/2001 -0800, you wrote:
>I'm in the processess of doing a cvs regression to find out what broke
>half-life in the main wine. The first problem appears in the "2001-11-8
>EDT" cvs, but it's not the same promlem as I was having with the current
>cvs. That one occured between the "2001-11-15 EDT" cvs and the "2001-11-13
>EDT".
>
>But unless anyone has any ideas I though I'd start working on what patch
>did it, and I was wondering if "2001-11-13 EDT" was the cvs at the
>begining or end of the day (before or after that days pathces were
>applied). And if the dates in the wine.cvs newsgroup are accurate.
>
>I'm rather new at this so any advice would be welcome.

Did you read documentation/cvs-regression ?

I updated some weeks ago to add precisely that you can
use times up to the second so that it's possible to apply each
patch separately.

I use 
http://www.winehq.com/hypermail/wine-cvs/
to find exact patches times.

>
>P.S. The error i'm looking for is a timeout of a CriticalSelection when
>you try to join a mulitplayer game. The one I found in the "2001-11-8 EDT"
>cvs is an unhandeled exception when it first goes to the "Loading..."
>screen.
>

Unfortunately, it is not impossible that a first breakage is 
completed by another a few days after. The more important
breakage to find is the first (unless the second breakage is
an attempt to fix the first)

Another possibility is that the 2 problems are only different
manifestations of a single problem; you can have a thread
crash nicely, triggering the debugger, or in different circumstances
nastily, without much messages. In the latter case, it can happen
that another thread is blocked by the  crashed thread and displays
a critical section wait timeout.

It will not be possible for me to help you more until next year, 
I'll not have internet access until then.
Good luck.

Gerard





More information about the wine-devel mailing list