<div class="gmail_quote">On Thu, Jun 3, 2010 at 11:04 AM, Boudewijn Rempt <span dir="ltr"><></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div><div></div><div class="h5">On Wednesday 02 June 2010, Dmitry Kazakov wrote:<br>
> Here is a small script for catching crashes. It stores the result in<br>
> ./crashReport file in the current folder.<br>
> As it can't catch "glibc memory corrupted" crashes, so they all go to<br>
> stdout.<br>
<br>
</div></div>Hmmm... Does this show many different crashes, or are all the crashes the<br>
result of the same problem?<br>
<br></blockquote><div><br>There are at least two left: in KoProperties and in KoColorSpaceRegitry.<br><br>Actually, Thomas said yesterday that KoLibs doesn't have any guarantees about thread safety, so, i think, all the Ko* classes can cause some problems and should be considered.<br>
<br>As for those two:<br>1) KoColorSpaceRegitry - i think adding a mutex is a must, because the KoColorSpaceRegitry::instance() is a static variable and can cause problems outside merge code even. So i'm waiting for Cyrille's word.<br>
<br>2) KoProperties - we can make a thread-safe wrapper like KisProperties and guard it with a mutex.<br> </div></div>-- <br>Dmitry Kazakov<br>