Upstream bug has details, this is a blocker for GNOME 2.26 release.
Correction, this problem also exists for firefox-3 Hence, the problem is in gnome-overlay
Correction again.
Just to confirm, xulrunner-1.9.1_beta3 fails to compile against pango-1.24.0. (It seems to run ok against it when compiled against an earlier version, probably just means I haven't hit the bug yet)... 5:)
(In reply to comment #3) > Just to confirm, xulrunner-1.9.1_beta3 fails to compile against pango-1.24.0. > (It seems to run ok against it when compiled against an earlier version, > probably just means I haven't hit the bug yet)... 5:) > Try to scroll while an image is loading, ff should crash :p
Ah, I just got the new package.mask and rebuilt the old pango. If you'd like me to test it and make sure it still fails, let me know... 5;)
This patch from the upstream bugzilla seems to work for me: https://bugzilla.mozilla.org/attachment.cgi?id=369648
(In reply to comment #6) > This patch from the upstream bugzilla seems to work for me: > https://bugzilla.mozilla.org/attachment.cgi?id=369648 > Could you please fix the URL you removed?
So, the patch from comment 6 can simply be dropped into the files/3.1_beta3 directory as 145-pango.patch and it'll apply cleanly and rebuild (assuming you don't run into bug 238687). I haven't spotted any further crashes, but I haven't had it recompiled for very long yet...
(In reply to comment #8) > So, the patch from comment 6 can simply be dropped into the files/3.1_beta3 > directory as 145-pango.patch and it'll apply cleanly and rebuild (assuming you > don't run into bug 238687). I haven't spotted any further crashes, but I > haven't had it recompiled for very long yet... > The reason I'm abstaining for now is because upstream is not yet done with reviewing the patch, and since this problem only affects people using *both* overlays, I don't want to impose this patch (and it's potential problems) on people using only the mozilla overlay. I'll add it once upstream ACKs the patch.
No probs, I'm still experiencing the odd crash, so I dunno whether the patch worked properly or not. I'm trying to track down the problem (it appears to be on loading image-heavy pages, like facebook), but so far I haven't gotten too far. Revdep-rebuild doesn't suggest anything needs rebuilding and I haven't seen it on my other machine so I'm not sure what's causing it...
(In reply to comment #10) > No probs, I'm still experiencing the odd crash, so I dunno whether the patch > worked properly or not. I'm trying to track down the problem (it appears to be > on loading image-heavy pages, like facebook), but so far I haven't gotten too > far. Revdep-rebuild doesn't suggest anything needs rebuilding and I haven't > seen it on my other machine so I'm not sure what's causing it... > Can you give an example of such a page (I don't use facebook =))? There were no problems compiling firefox-3.1b3 against pango-1.24.0
ooops, my fault, doesn't compile.
It's more likely my crashes are related to the nvidia drivers, or some other problem, rather than pango directly...
So finally, firefox-3.1b3 + pango-1.24.1 (modified ebuild, commented patch section) builds and works ok (at least it doesn't crash on flickr and stuff).
Ok, so the final patch on the upstream bug (https://bugzilla.mozilla.org/attachment.cgi?id=372739) seems to have had approval to go into the final version of xulrunner-1.9.1. I've tested it out, and get compilation errors without it, but dropping the patch in as 140-pango-1.23.0-fix.patch seems to work fine (and seems relatively solid from a runtime point of view). Is there any chance we could get this added (now that it's had review+ and superreview+ from upstream)?
Does this work with pango-1.22 as well? Or do we need conditional patching or something?
I don't know, I've moved onto xulrunner-1.9.1_beta4-r1 and it seems to be fine without the patch (I'm assuming it made it into the code itself). I think it'd be fine without, but really it might just be worth leaving it, since xulrunner-1.9.1_beta3 is never going to hit the main tree...
(In reply to comment #16) > Does this work with pango-1.22 as well? Or do we need conditional patching or > something? > The patch was applied upstream and is already in the overlay ... the pango mask can be removed.
(In reply to comment #18) > The patch was applied upstream and is already in the overlay ... the pango mask > can be removed. > Okay, I forgot to close this bug -- FIXED now.