=== doko_ [n=doko@dslb-084-059-064-213.pools.arcor-ip.net]  has joined #ubuntu-toolchain
[07:39] <fabbione> doko: http://people.ubuntu.com/~fabbione/gcc-4.1.log.bz2
[07:39] <fabbione> we didn't get that far at all
[07:46] <doko_> fabbione: it needs some love to _build_ biarch. it should build fine, if you
[07:46] <doko_> disable biarch (debian/rules.defs:583).
[07:47] <fabbione> doko: well what's the point of not building biarch?
[07:47] <fabbione> we need biarch for the kernel
[07:47] <fabbione> and quite a bunch of stuff
[07:50] <fabbione> doko: so you plan to make 4.1 the new default compiler for dapper?
[07:50] <fabbione> or are we going to stick with 4.0 and start preparing 4.1?
[07:50] <fabbione> that's sort of important given that dapper will open soon
[07:50] <fabbione> and we do have a set of SCC release out there
[07:52] <doko> fabbione: we'l have to rebuild the archive first, and decide then, not the other way around.
[07:53] <fabbione> #ifdef SPARC_BI_ARCH
[07:53] <fabbione> #undef CPP_ARCH32_SPEC
[07:53] <fabbione> #define CPP_ARCH32_SPEC "%{mlong-double-128:-D__LONG_DOUBLE_128__}"
[07:53] <fabbione> #endif
[07:54] <fabbione> this is in gcc/src/sparc/linux64.h
[07:54] <fabbione> as far as i can tell that one should just go away
[07:54] <doko> that looks ok for the 32bit case
[07:55] <fabbione> yes exactly
[07:55] <fabbione> the same lines are both for 32 and 64 bit
[07:55] <fabbione> i can see that __LONG_DOUBLE_128 is used later on
[07:56] <fabbione> i suspect that the mlong-double-128 bit should disappear
[07:56] <fabbione> let see.. it's building the normal one from scratch
[07:57] <fabbione> i have seen too many kick the rebuild and it works cases
[07:58] <fabbione> doko: the rebuild archive bit is scary.. specially because sparc doesn't have the horse power yet to do that
[07:59] <fabbione> doko: when building gcc.. and get these errors, can i safely do the modification and ./debian/rules build?
[07:59] <fabbione> or is it better to create the patch and rebuild from scratch?
[08:03] <doko> fabbione: no, you need to rerun the configury ... shortest way: rm stamps/05-build-stamp; make -C build
[08:04] <doko> sorry, make -C build bootstrap
[08:06] <fabbione> i think that error shows up way before that stamp
[08:07] <fabbione> yeah there is no 05
[08:09] <doko> ok, plese remove the configure stamp
[08:09] <fabbione> ok
[08:11] <doko> hmm, maybe we want to set need_64bit_hwint=yes as well for the biarch case (config.gcc)
[08:12] <doko> fabbione: what was the name of the ia64 machine in the dc?
[08:16] <doko> fabbione; ^^^
[08:26] <fabbione> halley
[09:32] <fabbione> doko: i can't get out of that error
[09:32] <fabbione> i managed to get rid of the -128 thing
[09:32] <fabbione> but not the -64
[09:33] <fabbione> ../../src/gcc/crtstuff.c:1: error: -mlong-double-64 not allowed with -m64
[09:33] <fabbione> and i did remove all possible references of that in gcc/config/sparc
[09:33] <fabbione> so it's either autogenerated somewhere
[09:33] <fabbione> or hidden in some .$fuckdir that i can't find
[09:34] <doko> did you set need_64bit_hwint=yes ?
[09:36] <fabbione> no
[09:36] <fabbione> i am testing one bit at a time
[09:37] <fabbione> otherwise we can't isolate it properly...
=== fabbione cleans the tree and tries again with that
[09:37] <doko> btw, did you get some feedback if OOo is actually working on sparc?
[09:37] <fabbione> doko: i did test it personally and i even told you :)
[09:37] <fabbione> that's faster than amd64 :)
[09:38] <doko> ahh, I remember :-)
[09:39] <fabbione> now.. fix gcc-4.1 dude..
[09:40] <doko> fabbione: so you build kernels and glibc using 4.0 on sparc?
[09:40] <fabbione> doko: we also need to fix gcc-4.0
[09:40] <fabbione> no
[09:40] <doko> ahh
[09:40] <fabbione> we don't do that even on main arches
[09:40] <doko> I know
[09:42] <fabbione> meh can you give me the line in config.gcc where i am supposed to stick that need_64bit thingy?
[09:43] <fabbione> doko: we also want to look at this failure http://bld-3.mmjgroup.com/buildLogs/k/klibc/1.0.14-1ubuntu2/klibc_1.0.14-1ubuntu2_20050915-0618-sparc-failed.gz
[09:43] <fabbione> that happens only with gcc-4.0
[09:44] <doko> fabbione: about 307, same as for sparc64
[09:45] <fabbione> doko: ok.. there is also another reference at around 2004
[09:46] <fabbione> sorry 2047
[09:46] <fabbione> ok let's try with only this change
[09:46] <fabbione> and see what happens
[10:13] <fabbione> doko: same story with that modification of config.gcc
[10:19] <doko> :(
[10:19] <fabbione> doko: time to fire up your sparc and install breezy?
[10:21] <doko> fabbione: have to get it going for the standard architectures first, currently wading through test results.
[10:22] <fabbione> doko: you can still do it in parallel :)
=== elmo [n=james@83-216-156-21.jamest747.adsl.metronet.co.uk]  has joined #ubuntu-toolchain
[11:50] <fabbione> doko: so ok.. i am going to try to disable biarch to test build
[11:59] <doko> fabbione: thanks
[11:59] <fabbione> doko: no problem, but please try to get biarch working :)
[12:48] <fabbione> doko: it fails even disabling biarch
[12:48] <fabbione> at a different point
=== fabbione collects the log
[12:48] <doko> yep, please send me it
[12:49] <fabbione> yup
[12:51] <fabbione> doko: do you prefer email or web?
[12:53] <doko> web is ok
[12:53] <doko> fabbione: ^
[12:53] <fabbione> ok
[12:54] <fabbione> doko: people/~fabbione
[12:55] <fabbione> there are 2 gcc-4.1-* log files
[12:55] <fabbione> with and without biarch defined in debian/rules.defs
[12:57] <doko> well, yes, that patch could be more robust
[12:59] <fabbione> don't assume i know all the patches applied to gcc..
[12:59] <fabbione> i just build gcc on your commands :)
=== lamont__ [n=lamont@15.238.7.17]  has joined #ubuntu-toolchain
[05:23] <doko> fabbione: updated, same location
[05:31] <fabbione> doko: ok.. taking it now
[05:35] <infinity> doko : Mail me when you have (what you believe to be) finalised sources for the gcc-4.1 testbuild, I'll bootstrap it on the buildds, and we'll see if we can get a breezy-gcc-4.1 building before UBZ.
[05:45] <fabbione> doko: i will let you know soon how it goes.. it's building 
[05:45] <fabbione> still didn't pass that point tho
[05:45] <doko> fabbione: it's currently in stage3 in my build
[05:46] <fabbione> on sparc?
[05:46] <doko> yes
[05:46] <fabbione> ah cool
[05:47] <fabbione> did you actually bootstrapped a breezy chroot ? or are you building on debian?
[05:48] <doko> breezy
[05:49] <fabbione> ohhh finally :)
[05:49] <doko> stopped the sid build after it completed stage1
[05:49] <fabbione> did it bootstrap properly?
[05:49] <fabbione> it should have...
[05:49] <doko> don't know, did it two months ago
[05:49] <fabbione> ahh
[05:49] <fabbione> ehehe
[05:50] <fabbione> ok thanks dude.. bbl
[10:01] <lamont__> configure: error: 
[10:01] <lamont__> The following requested languages could not be built: ada
[10:01] <lamont__> Recognised languages are: c,ada,c++,fortran,java,objc,obj-c++,treelang
=== lamont__ kicks doko
[10:01] <doko> lamont__: I'm innocent ;-p
[10:02] <lamont__> checking for gnatbind... gnatbind
[10:02] <lamont__> checking whether compiler driver understands Ada... no
[10:03] <doko> gnat-3.4 or gnat-4.0?
[10:04] <doko> lamont__: what did you use for bootstrapping?
=== lamont__ tries the simple thing first
[10:04] <lamont__> breezy
[10:05] <lamont__> and sbuild
[10:05] <lamont__> well, actually, dpkg-buildpackage
[10:05] <lamont__> so gcc-4.0, I expect
[10:07] <doko> strange, anything more in the changelog?
[10:07] <doko> s/changelog/config.log/
[10:09] <lamont__> iz my bad
[10:09] <lamont__> configure didn't like the 'could not open /CurrentlyBuilding'
[10:09] <doko> ;.)
[10:11] <doko> lamont__: but I should warn you ...
[10:11] <doko> xgcc: ../libmath/.libs/libmath.a: No such file or directory
[10:11] <doko> xgcc: ../libmath/.libs/libmath.a: No such file or directory
[10:11] <doko> make[7] : *** [libstdc++.la]  Error 1
[10:12] <doko> I don't know why ...
[10:12] <doko> works even on sparc
=== lamont__ will look
[10:13] <fabbione> doko: s/even//g
[10:13] <fabbione> doko: sparc > hppa ;) ask lamont on the final breezy buildd run :P
=== fabbione hides from lamont
[10:13] <lamont__> fabbione: that's all kde's fault
[10:14] <fabbione> ehehe
[10:14] <lamont__> well, actually, it's gcc-4.0's fault
[10:14] <lamont__> since it liked to ICE for lots of kde packages
[10:14] <doko> well, remembers me on hppa's messages on kernel errors: "your system ate a sparc."
[10:14] <fabbione> ahha
[10:16] <lamont__> "      _______________________________ \n"
[10:16] <lamont__> "     < Your System ate a SPARC! Gah! >\n"
[10:16] <lamont__> "      ------------------------------- \n"
[10:16] <lamont__> "             \\   ^__^\n"
[10:16] <lamont__> "              \\  (xx)\\_______\n"
[10:16] <lamont__> "                 (__)\\       )\\/\\\n"
[10:16] <lamont__> "                  U  ||----w |\n"
[10:16] <lamont__> "                     ||     ||\n");
[10:16] <lamont__> iz in die_if_kernel()
[10:16] <lamont__> and that's from upstream. :-
[10:16] <lamont__> )
[10:17] <fabbione> ahahha
[10:17] <doko> lamont__, maybe fabbione has difficulties laughing about it ;-P
[10:17] <fabbione> why?
[10:17] <lamont__> arch/parisc/kernel/traps.c
[10:18] <doko> we should help him ;)
[10:18] <lamont__> machine eats sparc, dies.
[10:18] <lamont__> film at 11
[10:36] <doko> fabbione: sparc build is in libjava ...
[10:37] <fabbione> doko: it's building here
[10:37] <fabbione> stage1/xgcc -Bstage1/ -B/usr/sparc-linux-gnu/bin/ -c   -O2 -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wold-style-definition -Wmissing-format-attribute     -DHAVE_CONFIG_H -I. -Ifortran -I../../src/gcc -I../../src/gcc/fortran -I../../src/gcc/../include -I../../src/gcc/../libcpp/include     ../../src/gcc/fortran/match.c -o fortran/match.o
[10:37] <fabbione> i guess i am far behind you
[10:42] <doko> fabbione: we'll have two independent results, I think, my hppa build is failing for some other reasons
[10:42] <fabbione> doko: yes.. i prefer to have 2 of them
[10:42] <fabbione> doko: given how delicate is gcc