Bug#660487: xulrunner-dev: instant segfault in startup (related to jemalloc)
Luke Kenneth Casson Leighton
lkcl at lkcl.net
Mon Feb 20 06:54:54 UTC 2012
On Mon, Feb 20, 2012 at 6:46 AM, Mike Hommey <mh at glandium.org> wrote:
>> but that's not the problem: the problem is generically to *all* -dev
>> libraries. hmm... let me raise this somewhere on a debian list, but
>> essentially what i need to do is to create a package
>> python-hulahop-that-compiles-and-links-to-xulrunner-9
>
> And that would have absolutely no usefulness in debian
beh? :)
>> why?
>>
>> because of "pkg-config --variable=sdkdir libxul".
>>
>> that line says it all: where does the libxul .pc file come from?
>> from the -dev package. what that *should* be is this:
>>
>> pkg-config --exact-version 9.0 --variable=sdkdir libxul
>
> like pkg-config --variable=sdkdir libxul = 9.0 ?
> tough that would break with 9.0.1.
sorry i meant --exact-version=9.0.1
>> but... but... that doesn't work either, does it, because _if_
>> xulrunner-dev is installed and it's xulrunner-10.0-dev, then we're
>> f****d, because that file libxul.pc is exclusively-named.
>>
>>
>> > The good news for you is that
>> > xulrunner 10 will stay there for a while and /might/ be what is released
>> > in wheezy.
>>
>> argh - actually that's _bad_ news, for exactly the reasons above.
>> not only can it not be compiled (against xulrunner-9) but also it's
>> xulrunner-10 which is severely borked.
>>
>> hmm... i wonder if xulrunner 11 is similarly borked?
>
> pyxpcom doesn't build with xulrunner 11.
it doesn't? argh! the fun continues :)
ok. so that means waiting for todd to sort it out, or take a look at
that as _well_. *sigh*.
so until then, basically, patches have to be tracked down that
stabilise xulrunner 10 (in order to do a xulrunner-10.0.2-NN or
whatever).
what i find odd is that the memory corruption which results in the
segfault on FocusWindow are triggered by, it looks like, macosx and
python-hulahop, but *not* firefox (under debian/linux amd64).
that's not to say that something *else* isn't getting corrupted...
l.
More information about the pkg-mozilla-maintainers
mailing list