Re: Problems on AMD64 with Glib module

[Ross McFarland <rwmcfa1 neces com>]

On Sat, 2004-01-10 at 09:33, Jacek Konieczny wrote:
> On Sat, Jan 10, 2004 at 02:58:24AM -0500, muppet wrote:
> > On Friday, January 9, 2004, at 02:09 PM, Jacek Konieczny wrote:
> > > $ perl -I blib/lib -I blib/arch t/a.t
> > > 1..8
> > > [snip output with every g_log message doubled]
> > > # Looks like you planned 8 tests but ran 7 extra.
> >
> > turns out this is a 64-bit bug in glib's g_bit_nth_msf.  yosh's fix 
> > should be in the next glib release; torsten and i whipped up a 
> > workaround for the Glib (perl) test suite that will skip the test when 
> > you're using the broken libraries.  irc can be a wonderful thing.
>
> It seems someone here in PLD have already found the problem and fixed it
> for me:
>
> (14:39) <+ABW-1> qboosh * SPECS/glib2.spec:
> (14:39) <+ABW-1> - added g_bit-64bit patch from glib CVS (for
> g_bit_{l,m}sf on 64-bit archs);
> (14:39) <+ABW-1>   fixes g_log() duplicating messages and thus perl-Glib
> tests on amd64
> (14:39) <+ABW-1> - release 3

we've found quite a few bugs in gtk+, with our perl regression tests,
which is pretty cool.

at some point later today i hope to get some time to look into some of
the other issues on the s.f. compile farm. (we might even have rpms for
suse 64bit, hehe) 

later,
-rm