[06:15] <fabbione> morning
[11:24] <fabbione> doko: ping?
[11:30] <doko> fabbione: pong
[11:33] <fabbione> doko: how do i use gcc-snapshot?
[11:33] <fabbione> should i just prefix my path with /usr/lib/gcc-snapshot/bin ?
[11:36] <doko> yes, that should do it. And the kernel build doesn't need any shared libs, so you don't need to set LD_LIBRARY_PATH
[11:36] <fabbione> ok
[11:37] <fabbione> i am trying to reproduce the ice on ppc
[11:48] <fabbione> doko: good news or bad news?
[11:48] <fabbione> which one do you want first?
[11:48] <doko> you can reproduce it, but you get a preprocessed source file?
[11:50] <fabbione> doko: no.. after unionfs upgrade from this morning i cannot reproduce it anymore...
[11:50] <fabbione> atleast on breezy-64 chroot
[11:50] <fabbione> i am testing in the normal one
[11:53] <doko> hmm, nice, or not nice? ;)
[12:01] <fabbione> well it's nice because we can ship the driver
[12:01] <fabbione> it's not nice because we don't know if it is a code change that fixed it or something in recent gcc
[12:01] <doko> you did test with 3.4 or 4.0?
[12:01] <fabbione> 3.4
[12:01] <fabbione> but i did do a normal test without using gcc-snapshot 
[12:01] <fabbione> so i guess it's the code in unionfs to be changed and doesn't trigger the ice anymore
[08:54] <jbailey> doko: Around>
[09:03] <doko> jb-support: yes :-)
[09:06] <jbailey> *lol*
[09:06] <jbailey> Bah, you just want me to be nice to you ;)
[09:06] <jbailey> I think I figured out how to morgue worked enough to find an old copy of gcc-3.4 anyway.
[09:07] <jbailey> Yup, Just made a 64 bit binary that works, woohoo
[09:09] <jbailey> Should I call the package libc6-amd64 ?
[09:09] <jbailey> And the matching libc6-dev-amd64
[09:12] <doko> yes, sounds ok. although I hate the naming ... but it's consistent with the other ones
[09:12] <doko> btw, the debian gcc-3.4 versions are still biarch, so these should work as well
[09:12] <jbailey> Cool.
[09:13] <doko> jbailey: will you sort it out tonight with infinity or lamont?
[09:13] <jbailey> doko: I'm going to try to put together a glibc and gcc-3.4 package now that they can use.
[09:17] <doko> ok, I uploaded a new 3.4 today, which FTBFS on i386. I'll send you a patch later today
[09:29] <jbailey> Cool, thanks.
[09:29] <jbailey> I'll also post the biarch glibc on people as soon as I have it.
[09:31] <doko> ok, then maybe we can merge the changes into one gcc-3.4 upload, which laminity can use
[09:31] <jbailey> *G*!
[09:36] <doko> hmm, when do you plan the glibc uploads/builds? still Monday?
[09:43] <jbailey> doko: Subject to the availability of Lamont or Adam, that sounds right.
[09:43] <jbailey> Unless one of them was bored this weekend.
[09:44] <jbailey> doko: What I'd like to do is hand them .deb's and sources and say "Install these debs.  Rebuild them with their matching sources.  Stuff the new binaries into the archive"
[09:44] <jbailey> That way it can be done at their convenience.
[09:44] <jbailey> lamont: Assuming I've got the steps right, yes?
[09:47] <doko> jabiley: yes, that sounds fine. do you prepare both biarch builds (i386 and amd64)?
[09:56] <lamont> jbailey: works for me.
[09:56] <lamont> I'm off to a birthday party at Renn Faire all day tomorrow, but we should be able to schedule something in...
[09:57] <lamont> is this just i386 we're doing funky bootstrapping stuff on, or is it more than just the one?
[09:57] <lamont> back in a fdew
[10:03] <jbailey> lamont: Just i386 for this round.  
[10:03] <jbailey> The i386-libs has a whole pile of crack in it that I'm not sure what the right answer to it is.
[10:03] <jbailey> Or even how much people on amd64 actually care about producing i386 binaries yet.
[10:08] <doko> jbailey: grub needs it
[10:09] <lamont> jbailey: yeah, what doko said.  grub is a 32-bit executable on amd64
[10:10] <jbailey> Ah, interesting.
[10:13] <jbailey> Ouch:
[10:13] <jbailey> checking build system type... i486-pc-linux-gnu
[10:13] <jbailey> checking host system type... powerpc64-unknown-linux-gnu
[10:17] <lamont> dude - that's serious multiarch
[10:17] <fabbione> ahaha
[10:19] <lamont> ccache sucks
[10:19] <lamont> cache hit                            296
[10:19] <lamont> cache miss                         12401
[10:19] <lamont> that or kernel option hacking sucks
[10:19] <fabbione> the latter
[10:19] <lamont> the latter makes the former suck
[10:22] <fabbione> i agree
[10:22] <fabbione> that's why i tend to build with -j200
[10:22] <fabbione> it scrolls faster and it looks that is compiling much more
[10:31] <fabbione> jbailey: to your knowledge.. how difficult it is to setup a cross compiler?
[10:32] <doko> fabbione: get Dan Kegel's crosstools
[10:33] <jbailey> fabbione: I don't have any troubles, but Dan's crosstools are supposed to be quite nice.
[10:33] <fabbione> doko: is it packaged?
[10:33] <doko> no, and the current cross build support in binutils/gcc-4.0 is a bit broken
[10:35] <jbailey> doko: Do you think Lamont would kill us if we built cross compilers for every arch on every arch?
[10:36] <doko> jbailey: so the build deps are: amd64-libs-dev [i386] , i386-build-deps [amd64]  ?
[10:36] <doko> jbailey: he doesn't have that many machines ...
[10:36] <fabbione> actually i am only curious to try to build sparc and m68k on i386
[10:38] <fabbione> but given that's broken i don't think i am going to spend time on it this night
[10:39] <doko> you start feeling bored during your vacations from the first day? :-P
[10:39] <jbailey> it's not even his first day yet.
[10:39] <jbailey> Like a Heroin addict... =)
[10:39] <fabbione> ahaha
[10:39] <jbailey> MUST HACK
[10:39] <fabbione> unfortunatly my evening has been ruined
[10:40] <fabbione> so better to hack than do nothing
[10:40] <jbailey>  /msg fabbione I told you bringing out the goat would be a mistake.
[10:40] <fabbione> ahahhaaha