[11:39] <fabbione> WO HOOO
[11:39] <fabbione> another gcc-4.0 regression on sparc
[11:40] <fabbione> ../../ghc/compiler/ghc-inplace -H16m -O -cpp -fffi -Iinclude  -ignore-package X11 -O -Rghc-timing -fgenerics  -package base -fgenerics -split-objs -hisuf p_hi -hcsuf p_hc -osuf p_o -prof   -c Graphics/X11/Xlib/Misc.hs -o Graphics/X11/Xlib/Misc.p_o  -ohi Graphics/X11/Xlib/Misc.p_hi
[11:40] <fabbione> gcc: Internal error: Segmentation fault (program as)
[11:40] <fabbione> Please submit a full bug report.
[11:40] <fabbione> See <URL:http://gcc.gnu.org/bugs.html> for instructions.
[11:40] <fabbione> For Debian GNU/Linux specific bug reporting instructions, see
.
[11:40] <fabbione> <<ghc: 296405008 bytes, 62 GCs, 4390005/8691260 avg/max bytes residency (4 samples), 20M in use, 0.01 INIT (0.00 elapsed), 17.04 MUT (84.43 elapsed), 4.34 GC (4.36 elapsed) :ghc>>
[11:43] <doko> isn't that as?
[11:43] <fabbione> dunno.. you are the expert :)
[11:47] <doko> you are the sparc expert :)
[11:48] <fabbione> well that's general toolchain stuff :P
[11:48] <fabbione> seriously.. what should i look for?
[11:50] <fabbione> if it's really as
[11:50] <doko> does ghc allow you to run gcc with -v ? then you should at least see the as command line, and then determine, if it fails in as or cc1
[11:50] <fabbione> then it's binutils fuckage
[11:51] <fabbione> i have NO idea
[11:51] <fabbione> i can try to do that
[11:51] <fabbione> it fails at 1.6M in the log that's pretty deep in the build
[11:51] <fabbione> i will try that asap. thanks
[11:52] <doko> didn't debian experience some problems with ghc lately?
[11:52] <fabbione> dunno
[11:52] <fabbione> they did experience glibc problems on sparc
[11:52] <fabbione> that's for sure
[11:53] <fabbione> it didn't even start to build in debian
[11:53] <fabbione> B-D issues
[12:51] <fabbione> doko: is the new gcc-4.0 good to build or should i wait a bit?
[01:00] <doko> fabbione: I tested on i386 and amd64, so maybe you want to wait until these succeed on the buildd
[01:00] <fabbione> ok
[04:37] <doko> infinity, lamont, lamont__: gcc-opt ping
[04:39] <fabbione> doko: anything that needs to be tuned on sparc?
[04:39] <fabbione> otherwise i am taking off to go to school
[04:40] <doko> fabbione: gcc-4.0 did succeed
[04:41] <doko> gcj-4.0 as well
[04:41] <fabbione> yup i saw that
[04:41] <doko> gcc-3.4 as well
[04:41] <fabbione> they will build "soon"
[04:42] <fabbione> i am also scheduling oo2
[04:42] <fabbione> but i have to go now
[04:42] <fabbione> ttyl
[08:06] <doko> fabbione: new oo2 upload needed
[08:09] <fabbione> doko: ok
[08:46] <fabbione> doko: when do you plan to upload oo2?
[08:46] <doko> uploading ...
[08:47] <fabbione> oh cool
[08:47] <fabbione> thanks
[08:48] <doko> the build did succeed on powerpc and amd64, there must be some problem with parallel make, which I didn't see with my testing
[08:48] <fabbione> oh ok
[12:00] <jbailey> doko: Around?
[12:00] <jbailey> I'm curious if you think it's safe to make the biarch amd64 package "libc6-i686" instead of "libc6-i386"
[12:00] <jbailey> Since I'm fairly sure that amd64 can support all the i686 instructions including CMOV.