Bug 452045

Created attachment 309803 [details]
RSSOwl debug output

Hm, I see dead.package in rssowl/F9 and devel. What's up?

I'm using the file from the rssowl homepage since there is no package for fedora
right now:
http://downloads.sourceforge.net/rssowl/rssowl-2.0-M8a.linux.gtk.x86.zip

Then report this bug at rssowl upstream. We can't fix it when the package is not
in fedora.

I'm not reporting a bug in RSSOwl but in xulrunner.

The bug is only is clearly and reproducably triggered by the new fedora
xulrunner package and the debug log shows the segfault occuring in libxul.so
which is part of that xulrunner package.

Other people are reporting a similar problem for other apps too on fedora-devel
(Miro is one of the affected applications apparently).

Hi Dennis,

There are some nightlies for RSSOwl here:

http://rssowl.pluginbuilder.org/latest/

They use a newer Eclipse RCP build that _may_ solve the problem for you. I am
using Fedora 9 with a RSSOwl without problems.

Having said that, there was a bug filed in Eclipse bugzilla about some issues
with the browser with the latest version too:

https://bugs.eclipse.org/bugs/show_bug.cgi?id=236724

Throwing the bug over the fence to Eclipse guys, because hopefully they could
have more clue about what's going on.

Reporter, would you please try some of these nightlies whether it works better
for you?

Yep, the nightly starts up fine with the final version of the xulrunner package.

FWIW, I verified that using the Fedora version of Eclipse with the Fedora
version of Firefox 3 (both using the most recent updates as of this comment)
works properly but that using the F9 Firefox 3 (really xulrunner) does not work
with the Eclipse 3.3.2 Classic installed from eclipse.org, i.e. it breaks with
the same signature in the debug log filed by the reporter.  So this is clearly
not a Fedora bug per se but I'd hazard a guess that the xulrunner Fedora
maintainer could easily debug and report upstream to either Eclipse or Mozilla
how to resolve the upstream issue.  Meanwhile I'll count my blessing that Fedora
has their collective act together on the Eclipse/Firefox/Java chain ... for now. :-)

-pmr

I probably should have added that to see the actual breakage is easy:

1) create a new workspace, say ~/452045/workspace.
2) download eclipse classic from the eclipse.org download site.
2) start the downloaded Eclipse and select ~/452045/workspace.

Wham, a jvm crash is reported showing a xulrunner sigsegv.  I would hazard a
guess that anyone astute enough to connect gdb and the offending xulrunner
(which leaves me out) will be able to provide useful information to the upstream
maintainers for either Eclipse or xulrunner.  I'd guess Eclipse since the F9
eclipse/xulrunner pair seems pretty content.  It almost feels as if there is a
static binding rather than dynamic but I'm way over my head at this point so
I'll stop right here.

-pmr

So is this the same as bug #452929?

It looks like (and is confirmed by Paul's comment #10) this only affects
upstream builds, right?

This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle.
Changing version to '10'.

More information and reason for this action is here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

This isn't a problem with current builds of RSSOwl 2.0 anymore => closing