[05:33] <fabbione> doko: i get the same error on gcc/pcc with  WITHOUT_LANG=biarch debuild -uc -us -b
[07:11] <doko> fabbione: don't have a G4 to check ...
[07:11] <fabbione> doko: ok.
[07:20] <fabbione> doko: i am about to start the rebootstrap on sparc. do you want to receive the build logs or not?
[08:03] <doko> fabbione: just the failed ones please
[08:13] <fabbione> doko: i can't filter from here
[08:14] <fabbione> i will need to forward it manually
[08:28] <doko> fabbione: we should get these filter scripts from lucas (the guy who did make the cluster build)
[08:30] <fabbione> doko: sbuild doesn't filter.. it just sends
[08:32] <infinity> It's trivial to filter is with procmail.
[08:32] <infinity> Given the uniform subject lines.
[12:42] <Keybuk> do we have a toolchain for feisty yet?
[12:42] <fabbione> no
[12:43] <fabbione> it should be by the end of the week
[12:45] <infinity> We have several, and some of them almost work, we don't have one in the archive yet.
[02:24] <jbailey> doko: g'day!
[02:24] <doko> jbailey: good morning!
[02:24] <jbailey> doko: I put the binutils stuff in bzr based on the current packaging.  One patch merged upstream (i386/amd64 biarch).
[02:24] <jbailey> Build of glibc shows no regressions.
[02:25] <jbailey> (ppc)
[02:25] <jbailey> and ppc64.
[02:25] <jbailey> I have the .orig.tar.gz in my homedir on chinstrap if you're interested.
[02:25] <jbailey> glibc correctly detects the hashing stuff and enables it.
[02:25] <doko> jbailey: ok, will have a look
[02:41] <fabbione> jbailey: hashing is detected fine on sparc too
[02:42] <jbailey> fabbione: Luvly.
[02:43] <jbailey> I have a feeling i'mgoing to need to dust off my G4 for some build tests.
[02:43] <jbailey> the ppc64 glibc has no significant errors.
[02:43] <jbailey> the ppc32 one does.  I suspect it's all compat bugs, though.
[02:43] <jbailey> I still want to look over the i486 one.
[02:43] <jbailey> i686 and amd64 looked fine.
[02:44] <jbailey> I'd be surprised if ia64 were anything other than fine, but I should build that too.
[02:44] <jbailey> fabbione: I haven't committed my patch stuff yet.
[02:44] <jbailey> I'll try to get it done by end of today, I think I was too aggressive in my pruning and want to review it again.
[02:45] <fabbione> jbailey: what patch? glibc or binutils?
[02:46] <jbailey> glibc.
[02:46] <jbailey> binutils is probably final subject to a review from doko.
[02:46] <fabbione> jbailey: this binutils can build hppa right?
[02:47] <jbailey> fabbione: Right.  glibc itself is missing a patch, but the binutils is fine.
[02:47] <fabbione> jbailey: ok. i am going trough a bootstrap sequence.. i will stop before glibc and wait for you
[02:47] <jbailey> fabbione: on hppa?
[02:48] <fabbione> yes
[02:48] <jbailey> I nudged adam a bit about the hppa kernel last night.  Dunno if he found time for it or not.
[02:48] <fabbione> i am building a kernel from git now. just to get the headers and get to binutils -> glibc -> gcc
[02:49] <fabbione> i am pretty sure the kernel does build
[02:49] <fabbione> but it might not run.. tough
 jbailey: binutils looks ok on sparc.. builded glibc on top and gcc on top of glibc... no errors up till now
 it would have fallen apart a while ago
[05:31] <jbailey>  anyway i am off for a bit
 up until now, like you just got an error, or like no errors so far? =)
 time to enjoy wife :)
[05:31] <jbailey> doko: ^^ 
[05:31] <jbailey> We were in -kernel, instead of -toolchain, sorry.
[05:31] <fabbione> he is in kernel too
[05:32] <jbailey> *lol*
[05:32] <fabbione> ok i am off now
[06:05] <fabbione> jbailey: libc6-dev in dapper Depends: linux-kernel-headers
[06:05] <fabbione> and linux-libc-dev doesn't provide it afaict
[06:06] <fabbione> so once we build hppa kernel, we can't build glibc
[06:06] <fabbione> i guess we will need to force that one in
[06:07] <fabbione> actually no
[06:08] <fabbione> binutils doesn't need new linux-libc-dev to build
[06:08] <fabbione> but glibc does..
[06:08] <fabbione> just postponing the issue..
[06:09] <doko> fabbione: did see it, nice.
[06:09] <fabbione> doko: gcc build seems to going fine now on sparc
[06:09] <fabbione> even on Niagara
[06:09] <fabbione> doko: once gcc is done i will rebuild glibc and open the gates for testing
[06:09] <fabbione> NOTE that i will not be using the rebuilt debs
[06:10] <fabbione> we are only going to test pkg foo built on edgy+feisty toolchain
[06:10] <fabbione> and not feisty rebuilt on feisty
[06:10] <fabbione> that will happen automatically while we import stuff
[06:10] <fabbione> jbailey: i need the glibc patch for hppa somewhere.. even offbandwith so i can test build and test
[06:11] <doko> fabbione, jbailey: for long double 64->128 change: we'll need to rebuild all packages (at least on powerpc and sparc) which uses long double to be safe; it's not that easy to detect thoese packages easily
[06:11] <fabbione> doko: well i will try to rebuild all of main BEFORE we open the gate
[06:11] <doko> not sure, if we want to open the gates for libraries and binary-indep stuff first, then for other packages
[06:11] <fabbione> do you expect just plain FTBFS or runtime errors?
[06:12] <fabbione> we can't really filter manually.. it would be hell of a work to let one package at a time in
[06:12] <doko> fabbione: runtime: libc and libstdc++ dohave support for both 64 and 128 bit long double, but other libraries don't have it.
[06:13] <fabbione> oh christ
[06:13] <fabbione> doko: we do have a local mirror on people.. what about a nice apt-get source $allofmain and a lot of grep magic?
[06:13] <fabbione> that would help me to prioritize pkgs that use that 128bit thingy
[06:13] <fabbione> and test them
[06:14] <doko> fabbione: you would have to unpack tarballs inside packages, scan for typedefs and defines, and probably still miss some
[06:15] <fabbione> doko: well that's doable.. 
[06:15] <fabbione> missing some can happen
[06:15] <fabbione> but we can still catch a few for testing-?=
[06:15] <doko> yes, probably
[06:17] <fabbione> doko: well want to start a script now so i can actually get to test by tomorrow or something?
[06:18] <doko> yes, probably write one tonight. but you can start testing as well
[06:18] <jbailey> doko: I think in general it would be nice to go through and look at any package that hasn't changed ina release and rebuilt it for good measure anyway.
[06:19] <fabbione> doko: i can't start testing without rebuilding first
[06:19] <fabbione> jbailey: statically i haven't see ONE release in which all packages in main have been the same
[06:19] <fabbione> only one _all.deb in one release IIRC
[10:48] <jbailey> doko: awake?