Josh wrote:
bulia byak wrote:
On 10/16/06, MenTaLguY <mental@...3...> wrote:
To the best of my recollection, the problem appears to be some detail with inappropriate cleanup of the tool's cursor resource in src/event-context.cpp; possibly a GDK refcounting issue. Unfortunately I have not had time to investigate the relevant code in depth.
So, do you think it likely that it was caused by
http://svn.sourceforge.net/viewvc/inkscape/inkscape/trunk/src/event-context....
and related changes? At least, it looks plausible time-wise, as this change was shortly after 0.44 and people report that 0.44 was ok and crashes started soon thereafter.
I will investigate if this is really the case. Scislac, please review the color cursors code for obvious errors. If this is indeed the culprit, and if you don't have a good idea of how to fix it, I'm afraid we will have to roll it back. As you remember, on the first attempt the color cursors also caused problems on win2k (white rects), and now these crashes... it's really not worth it to suffer all this trouble for what is basically superficial eye candy.
I will look into it. I do remember the win2k issues, and those got resolved... hopefully we can get this one resolved too. I definitely agree that it isn't worth it for something superficial.
-Josh
My reading of the above ( http://svn.sourceforge.net/viewvc/inkscape/inkscape/trunk/src/event-context.... suggests that you are looking at changes occurring between v11889 (may 21) and v12394 (jun 19). However, please note that the official version 0.44.1-1.win32 (sep 9), brought out three months later, DOES NOT have this problem. Why does that version work and nightly versions to date do not?
regards, Erik