/srv/irclogs.ubuntu.com/2006/10/26/#ubuntu-toolchain.txt

=== doko__ [n=doko@dslb-088-073-093-228.pools.arcor-ip.net] has joined #ubuntu-toolchain
=== anibal [i=anibal@ns1.mssinc.biz] has joined #ubuntu-toolchain
dokofabbione: gcc-4.1 did sucessfully build. please install on feisty01:10
=== rodarvus [n=rodarvus@ubuntu/member/rodarvus] has joined #ubuntu-toolchain
fabbionedoko: i need to ask Spads to do it. i dont have root there05:09
fabbionejbailey: ping?05:17
jbaileyfabbione: yoyosup?05:17
fabbioneyo dude05:18
fabbioneso it's confirmed.. glibc builds hang on Niagara. i am trying to build on a normal sparc now.05:18
fabbioneppc is rebuilding now (without the Zombie processes)05:18
fabbioneia64 is ahead of everything.. rebuilding glibc for the second time after glibc -> gcc05:18
jbaileyIs the T1000 also a Niagara?05:18
fabbioneyes05:18
fabbioneT1000 and T2000 are Niagara05:18
jbailey'kay.  We have one coming to Montral, apparently.05:19
fabbionenice05:19
fabbionebut i am pretty sure it's something Niagara related05:19
jbaileyI had actually wanted a regular sparc so that we didn't have the noise problems, but ah well.05:19
fabbionedid your build with BUILD_CC finished?05:19
jbaileyBUILD_CXX, lemme check.05:20
jbaileyYup, lesse the errors.05:20
jbaileyTesting ASCIItst-tables.sh: 259: tst-table.sh: not found05:21
jbaileyWeird, that's different than last time.05:21
fabbionehmm ok05:21
fabbionei will start soon a test build on a ppc32 kernel05:21
fabbionejust to isolate05:21
jbaileymake[3] : *** [/home/jbailey/i386/glibc/glibc-2.5/build-tree/i386-libc/posix/tst-05:22
jbaileyrxspencer.out]  Error 13905:22
jbaileyThat's a segfault.05:22
fabbioneinteresting05:22
jbaileyOh, right.  I hadbeen looking at i686 before.05:22
jbaileyi486 has problems, but those might exist today.05:22
jbaileyi686 is now correct for those C++ tests.05:23
jbaileyamd64 biarch, I didn't specify -m64 in there.05:24
jbaileyAside from that looks good.05:24
jbaileyso mostly just i486 needs some love there.05:24
jbaileyLemme commit that fix.05:24
fabbionethanks05:24
jbaileyDone, thanks.05:33
jbaileyfabbione: Need anything else before I crawl off to bed?05:35
fabbionethis fix is only for the test suite right?05:36
jbaileyYup, reduce some false negatives.05:36
fabbioneok thanks05:39
fabbioneno more sir.. good night sir!05:39
jbaileyLuvly.   demain!05:47
=== Starting logfile irclogs/ubuntu-toolchain.log
=== ubuntulog [i=ubuntulo@ubuntu/bot/ubuntulog] has joined #ubuntu-toolchain
fabbionelamont, infinity: ia64 toolchain is good.07:17
fabbioneglibc -> gcc -> glibc and everything still hold up07:17
fabbionewe only need the kernel headers fixes now07:17
fabbionebut that's for arch all07:18
fabbionedoko: i requested gcc install on ronne's amd64 chroot and opening of an i386 chroot+glibc install.07:18
fabbioneppc/sparc are a bit doomed at the moment07:19
fabbioneppc glibc build can cause a DoS07:19
fabbione(WOWOWOW)07:19
fabbionesparc just need a bit of extra love07:19
fabbione(test suite crappage)07:19
dokofabbione: did you request a chroot on davis as well?09:53
fabbioneyes09:53
fabbionelibc6_2.5-0ubuntu1_sparc.deb                                                                           100% 4520KB   4.4MB/s   00:00    09:54
fabbionethere we have sparc09:55
fabbionewe miss ppc still09:55
fabbionethere is an issue in the testsuite that hangs but i think i have it now09:55
fabbionedoko: asked for a chroot on faure too with gcc build-dep10:19
fabbioneglibc fails to build on my Niagara. I know why but i don't have time to fix it right away10:19
fabbionedoko: gcc-4.1 FTBFS on sparc with glibc-2.5 11:33
fabbionesee /msg11:33
fabbionedoko: anyway chroot on faure has been requested11:38
dokoahh, this was the txxx?11:38
fabbioneyes11:39
fabbionebut in this case should make no difference11:39
fabbioneit was bootstrapping11:39
fabbione./xgcc -B./ -B/usr/sparc-linux-gnu/bin/ -isystem /usr/sparc-linux-gnu/include -isystem /usr/sparc-linux-gnu/sys-include -L/usr/src/sparc/gcc-4.1-4.1.1/build/gcc/../ld -O2  -O2 -g -O2  -DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -fno-stack-protector  -I. -I. -I../../src/gcc -I../../src/gcc/. -I11:39
fabbione../../src/gcc/../include -I../../src/gcc/../libcpp/include  -DL_muldc3 -fvisibility=hidden -DHIDE_EXPORTS -c ../../src/gcc/libgcc2.c -o libgcc/./_muldc3.o11:39
fabbione(that was the command)11:39
fabbionedoko: trying to rebuild with linux32 ...11:49
fabbionejust for the fun11:49
fabbionedoko not a diff with linux3212:06
dokofabbione: is this failure in the 64bit part? i.e. does building with WITHOUT_LANG=biarch work?12:07
fabbionedunno what part is that.. it's just a few lines above..12:08
fabbionei can test yes.. but i am heading for food and sleep soon12:08
fabbioneok build started12:09
fabbionei will let you know12:09
fabbioneexport WITHOUT_LANG=biarch12:12
fabbionedoko.. same error12:12
fabbionethat looks like the 32bit part of it12:12
fabbioneanyway i am off for real now12:12
fabbionewe will wait faure to be ready12:12
=== jbailey [n=jbailey@montreal.canonical.com] has joined #ubuntu-toolchain
jbaileyfabbione: g'm02:47
dokohi jbailey 02:54
jbaileyHeya Matthias02:54
dokodid drepper release 2.6?03:02
jbaileyEh?  No.  It's not expected for another 6 months.03:03
jbaileyThey're releases with FC.03:03
jbaileys/They're/They/03:04
jbaileys/releases/release/03:46
jbailey*clearly* awake.03:46
jbaileydoko: Another advantage to using the HJ Lu snapshots is getting the PT_GNU_HASH stuff.03:46
jbaileyIs the concern basically that the packages are too bleeding edge?03:47
dokoAFAIK elmo did not want them, because they are not tested on the minor debian archs and did break lot of things in the past.03:48
=== doko invites elmo
jbaileyI suspect the sysadmins are either keeping the harddrives cool with firehoses or are out drinking at the moment ;)03:50
dokobah03:55
=== mlpug [n=user@a84-231-238-186.elisa-laajakaista.fi] has joined #ubuntu-toolchain
fabbionemorning jb04:46
fabbionejbailey: ping?04:49
jbaileyfabbione: PONGEROO!04:50
fabbionejbailey: hey dude.. so sparc glibc are go.. gcc FTBFS04:50
jbaileyWhat's the FTBFS?04:50
fabbioneppc can't build glibc and the build can DoS the machine04:50
fabbionei already passed it on to doko04:50
jbaileyInteresting.  I wonder if that's a kernel issue.04:51
fabbione../../src/gcc/libgcc2.c:1629: error: size of array 'compile_type_assert' is negative04:51
fabbionemake[5] : *** [libgcc/./_muldc3.o]  Error 104:51
fabbionemake[5] : Leaving directory `/usr/src/sparc/gcc-4.1-4.1.1/build/gcc'04:51
fabbionemake[4] : *** [stmp-multilib]  Error 204:51
fabbionemake[4] : Leaving directory `/usr/src/sparc/gcc-4.1-4.1.1/build/gcc'04:51
jbaileyI'm running latest edgy on mine, and I didn't have a failure when I last tried it.04:51
fabbioneppc is running ppc64 kernels. i need to try on my ppc3204:51
jbaileyI can try a build of your packages, though.04:51
fabbionethat's tomorrows job04:51
jbaileyRight, so is mine.04:51
fabbionebrb son is crying04:51
jbaileyI'll try a build at home over lunch.04:51
fabbionere04:52
fabbioneyes please04:52
fabbionejbailey: ia64 is gold.. 04:54
fabbionei am waiting for am64 gcc to be installed on ronne and feisty-libc i386 chroot created + glibc04:54
fabbioneso for now it looks decent04:54
fabbionecould be better tho04:54
fabbionei also found out part of the problem with glibc build hanging04:54
fabbione(on sparc)04:54
fabbionethere is a commit that works around it04:55
fabbionejbailey: also.. please don't use davis to build for now.. each build requires a davis reboot to cleanup04:59
fabbionemight be just a kernel thingy.. but well04:59
jbaileyUnkillable processes are always broken kernel, no?05:02
jbaileyDo we have the updated kernels?05:02
jbaileyBut in general, I have a decent ppc box at home, so I don't tend to use davis.05:02
fabbionewell i don't see necessarely how it can be a kernel fault but i am too tired to look into it now05:08
fabbioneanyway i will test here tomorrow morning with ppc3205:08
fabbione(since i can reboot locally)05:08
jbaileySure, or I'll test later too.  Whatever.05:11
jbaileyUnless you'd rather do it?05:11
jbaileyI'm surprised at your sudden interest in glibc. =)05:11
fabbionewe should test both05:16
fabbionei am on ppc32.,. you ppc64 kernel05:16
fabbioneguess why we should do it05:16
jbaileyI could test on ppc32, but my Pegasos box is quite slow.05:17
fabbionejbailey: bah i decided i want to be the only one that got to upload all possible source packages > 10MB or so :)05:17
fabbionenah my G4 is fast05:17
fabbionedon't worry05:17
fabbionewe can just share the load05:17
jbaileySounds lovely. =)05:17
fabbionei will only need to convince doko to let me upload gcc and OOo :)05:17
fabbioneonce those 2 are done, i can go in pension :)05:17
jbaileyDude, this is Ubuntu.  No BML. =)05:17
jbaileyBesides.05:17
jbaileydoko: Eh, Matthias.05:17
jbaileydoko: Fabio wants to take over OpenOffice.  Any objections?  =)05:18
=== jbailey hides.
fabbionejbailey: f**k y*u05:18
fabbionewith love05:18
fabbioneFabio05:18
jbailey*lol*05:19
jbaileyHmm.  if it's likely to hang my ppc box, I should probably do that when I get home.05:19
fabbioneok. i am off for the evening05:19
jbaileyOtherwise I lose access to a bunch of things.05:19
fabbioneme needs relax a lot. kthxbye05:19
jbaileyg'n F!  Happy drinking!05:19
fabbionejbailey: it doesn't kill your machine05:19
fabbioneit leaves Zl processes there05:19
jbaileyfabbione: Oh, I can still reboot it remotely?05:19
jbaileyEasy then, I'll do a build in a sec.05:19
fabbioneand it needs a really hard reboot05:20
jbaileyI'll also take a look at the i386 failures on davis if I can.05:20
jbailey /sbin/reboot isn't enough?05:20
fabbioneif you issue reboot from console you are screwed05:20
fabbionereboot -n -f probably might do it05:20
jbaileyMeh05:20
fabbionebut the machine doesn't die till it's up05:20
fabbionethat's really funny05:20
fabbionei386 failures on davis <- that doesn't work.. on ronne? :P05:20
jbaileyronne rather, right.05:20
jbaileyDo we have new kernel from Ben yet?05:21
fabbioneeheh eok05:21
fabbionethe kernel is in git, but don't install linux-libc-dev or you are doomed. that hasn't been fixed yet AFAIK05:21
fabbioneworth to try tho05:21
fabbionei am happy if we can start glibc -> gcc -> glibc05:21
jbaileyDoes he keep that on kernel.org ?05:21
fabbioneyeps05:21
jbaileyCool.05:21
fabbionehttps://wiki.ubuntu.com/KernelGitGuide05:21
jbaileyPerfect, thanks.05:22
fabbionenp05:22
=== fabbione heads offline
jbaileyg'n!05:23
dokojbailey: are you working on faure or davis?05:33
jbaileyNeither.05:33
dokook, trying to build the current gcc-4.1 there on edgy first ...05:34
DvalinLinux cm-84.210.59.173.chello.no 2.6.19-rc3.1mdv #1 PREEMPT Thu Oct 26 00:17:35 CEST 2006 sparc64 TI UltraSparc IIe (Hummingbird) unknown GNU/Linux05:41
Dvalinreverting to old gcc did actually do the trick :)05:42
Dvalindoko: I had same issues with sb100 booting your guys kernel for sparc too.. maybe the same one?05:42
=== mdz [n=mdz@217.205.109.249] has joined #ubuntu-toolchain
mdzfeisty exists in launchpad now; presumably infinity has some work to do before anything useful can be done with it07:23
mdzbut that bit is done07:23
jbaileymdz: Sweet, thanks.07:39
dokonice07:49
dokojbailey: the GCC build failures are bootstrapping problems with long double 128; anything that needs to be enabled in glibc?07:52
Dvalindoko: I have some issues that might be similiar..?07:58
Dvalin./xgcc -B./ -B/usr/sparc-mandriva-linux-gnu/bin/ -isystem /usr/sparc-mandriva-linux-gnu/include -isystem /usr/sparc-mandriva-linux-gnu/sys-include -L/home/peroyvind/RPM/BUILD/gcc-4.1.2/obj-sparc-mandriva-linux-gnu/gcc/../ld -O2 -O2 -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -m32 -mtune=ultrasparc  -DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -I. -I. -I../../gcc -I.07:58
Dvalinooops07:58
Dvalin/usr/include/bits/types.h:147: error: size of array ?__val? is too large07:58
Dvalina lot of similar errors..07:58
Dvalinmanaged to bootstrap, but then you see what happened afterwards.. (applied your sparc_biarch patch to gcc)07:59
jbaileydoko: I think glibc has everything enabled by default.07:59
jbaileydoko: Is there a build log I can look at to understand better?08:00
dokojbailey: overwritten08:04
lamontdoko: remind me why a binary defining rpath is a bad thing...08:10
jbaileylamont: It's not possible to override an rpath with a LD_LIBRARY_PATH, iirc.08:10
lamontis that all>08:10
lamont?08:10
dokoI don't know anything more08:11
lamontok08:11
lamonthrm... lintian warning: W: xxx: postrm-has-useless-call-to-ldconfig08:12
lamontcat postrm08:12
lamont#!/bin/sh08:12
lamontset -e08:12
lamont# Automatically added by dh_makeshlibs08:12
lamontif [ "$1" = "remove" ] ; then08:12
lamont        ldconfig08:12
lamontfi08:12
lamont# End automatically added section08:12
lamontand there's a shlib in the package...08:12
lamontotoh, it's not in /usr/lib08:12
lamontis that a bogus warning, or a valid one, I wonder?08:12
jbaileyIf it's not in /usr/lib, ldconfig won't see it.08:12
jbaileyWell.08:13
lamont /usr/lib/foo08:13
jbailey /lib, /usr/lib, or a patch specified in /etc/ld.so.conf or /etc/ld.so.conf.d08:13
lamont /usr/lib/foo/libfoo.so.2.5.208:13
lamontso why did dh_makeshlibs add the call, I wonder08:13
jbaileyDunno.  Unusual to have a versioned lib in a subdirectory, though.08:14
lamontyeah, well...  this is unusual code08:18
fabbionemdz: i am pretty sure infinity needs to create the chroots for it. In any case we can't upload glibc just yet. PPC has issues and it will kill buildds08:44
fabbionemdz: i386/amd64/sparc/ia64 looks good for glibc. gcc needs some love on sparc apparently (unknown ppc)08:45
fabbionedoko: Nick did the chroots on ronne and davis. but note that davis doesn't have glibc-2.5 quite yet08:54
fabbionechecking for sparc too08:55
fabbioneno it's not done yet08:57
fabbioneanyway movie time08:57
dokofabbione: have fun09:03
fabbionedoko: thanks dude09:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!