/srv/irclogs.ubuntu.com/2005/06/01/#ubuntu-toolchain.txt

jbaileydoko: Sent off another lkh, should fix both mysql and libgii12:16
dokoso libgii was a lkh thing?12:17
jbaileyyeah, the upstream included a define that only exists in 2.4 kernels and earlier.12:17
jbaileyGiven that the case is a noop, I'm surprised to see it in mysql, but ah well.  It's not supposed to be in the 2.6 header.12:18
lamontdoko: done12:21
lamontKamion: yeah - in the report, they show up... I think those particular ones probably want to get removed from the archive12:22
dokolamont: yes, too early, the 2nd pike upload didn't build yet12:25
dokoenchant did build, so abiword can be retried12:32
lamontkicked12:41
lamontdoko: you are waiting for the cron.daily run after the builds finish before saying anything, yes?12:41
dokohmm, I don't do idle waiting ... so from time to time I refresh your web page. that's ok12:42
lamontgii now d-w lkh12:43
lamontjbailey: thanks12:43
lamontwas it mysql-dfsg, or the universe package?12:43
dokothe former12:43
lamontonly have a failure log on ia64 - d-w'd12:44
dokoyes, the other archs did succeed12:44
lamontkewl.12:44
lamontanything else before I run off again for a while?12:44
jbaileyNot from me, I think.12:45
lamontback in 2-3 or so then12:48
lamontcloser to 2 than 3, I think12:48
dokoKamion: could you check out, why unixodbc's build-deps cannot be fulfilled on powerpc?01:15
Kamiondoko: it's hard for me to check anything in breezy on my own system because of that bash bug :P01:16
dokoouch!01:17
Kamionbut http://people.ubuntu.com/~cjwatson/testing/breezy_probs.html agrees01:17
dokoyes, first thing I do tomorrow morning01:17
Kamionthanks :)01:17
Kamionit looks like gnome-libs is out of sync on powerpc01:18
dokoyes, I read this list, but even the build logs aren't very helpful01:18
Kamionlibgnome32 |   1.4.2-19 |        breezy | powerpc01:18
Kamionlibgnome32 |   1.4.2-20 |        breezy | amd64, hppa, i386, ia64, sparc01:18
Kamionand gnome-libs FTBFS/powerpc01:18
Kamionlooks like a missing imlib include?01:19
dokoworse: libpng error: Call to NULL read function01:22
dokohmm, libpng was the cause for xorg build failures on powerpc as well01:23
=== ajmitch_ [~ajmitch@port162-41.ubs.maxnet.co.nz] has joined #ubuntu-toolchain
jbaileydaniels: Hmm, on the upgrade to the latest X bits, dexconf or whatever it's called now doesn't appear to write the mouse stuff to the xorg.conf file.01:50
dokojbailey, libgii and mysql-dfsg did fail02:00
jbaileychroots weren't updated with the new lkh02:00
dokolamont: after libpng did build and install on powerpc, please build gnome-libs on powerpc02:00
=== svenl_ [~luther@AStrasbourg-251-1-59-4.w82-126.abo.wanadoo.fr] has joined #ubuntu-toolchain
lamont../sigcperl/convert.h:71: error: explicit qualification in declaration of `std::string SigCPerl::get_string(SV*)'02:03
lamontbad sigcperl02:03
lamont../include/linux/i2o-dev.h:51: error: variable or field '__user' declared void02:03
lamontbad raidutils02:03
lamont../llapi/vector2d.h:120: error: 'cerr' was not declared in this scope02:04
lamontbad panorama02:04
lamontpension board meeting - back in about an hour02:17
danielsjbailey: yeah, the mouse stuff is screwed02:31
danielsjbailey: that's -21 material02:31
dokodaniels: congrats! -20 did build on all archs ;-)02:33
danielsfabbione: eh, branden isn't doing any of the xorg work02:33
danielsfabbione: gravity and I are doing xorg monolithic stuff ... he's so far caught one mistake and done a bit of syncing from Debian02:34
danielsfabbione: joshtriplett and I are working on the modular tree and we're hoping to have the same packages02:34
danielsdoko: yeah, I noticed ... finally02:35
=== zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain
zulgday02:36
dokoand xbase-clients seems to be built correctly02:37
dokodaniels: could you have a look at xscreensaver? the changes were necessary when I built the package, but they do fail on the buildd.02:38
danielsok02:39
=== desrt [~desrt@kopesetik.desrt.ca] has joined #ubuntu-toolchain
=== zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain
jbaileydaniels: That and the path being not where startx is located now. =)03:09
danielsjbailey: serves you right for not having /usr/X11R6/bin in your path :P03:11
danielsi think i'm just going to dump everything in /usr/bin with the next upload03:11
danielsbeing that /usr/bin/X11 is a REALLY BAD IDEA03:11
jbaileyI thought the plan was to move everything to /usr/bin ages ago.03:12
danielsright, but moving it within the monolith is PAIN03:18
danielsso I'm just making all the new modules use it03:18
danielsi'm seriously considering moving everything and xbase-clients and xutils with dh_install though03:19
Kamiondoesn't that break some third-party code using imake?03:22
Kamion(due to stuff not being in ProjectRoot where they expect it)03:23
danielsthat's the tricky part I have to work out03:31
jbaileyHmm, libgii seems to be an X casualty this time.04:14
daniels\o/04:15
danielsdoes it need to build soon?04:15
danielsif it doesn't need to build in the next few days, just wait for it to start building again04:16
jbaileyIt's an rdepends for anything libggi based that needs to make the g++ transition.04:16
jbaileyI don't know how importnat that makes it.04:16
danielsmain or universe?04:18
jbaileymain04:18
danielshum04:18
jbaileyGiven that I'm heading to bed soon, I don't actually care that much for any value that you might consider to mean "today"04:19
danielsheh :)04:20
jbaileydoko punted it to me since the first failure was a linux-kernel-headers one, lamont punted it back when it failed to build the package right because of a missing binary.04:20
danielsheh :)04:22
lamontwas someone fixing gstreamer0.8?  fabbione ?04:34
jbaileyIt wasn't punted to me.04:41
=== lamont just remembers someone muttering about it several hours ago...
lamontcothreads.c:654: error: invalid lvalue in assignment04:41
lamontwas sparc ftbfs, iirc... --> fabbione 04:41
=== warthylog [~warthylog@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain
=== Topic for #ubuntu-toolchain: GNU Compiler Collection, Glibc, Binutils, Linux-kernel-headers | GLIBC Todo: hppa, sparc NPTL, i386 biarch, C++ ABI change: 33/55 library packages in the archives
=== Topic (#ubuntu-toolchain): set by doko at Thu May 19 00:40:28 2005
fabbionedaniels: xorg -20 is FTBFS on sparc06:22
fabbioneConnDis.c:38:23: error: X11/Xauth.h: No such file or directory06:22
fabbionearound 1.5MB in the build log06:22
danielsuhm, sure you've got libxau-dev installed?06:24
danielsbecause that should be in /usr/include/X11/Xauth.h06:24
daniels(and a recent libxau-dev?  doko uploaded a fixed -2)06:24
fabbionechecking...06:25
fabbionewell libxau-dev should be a build-dep06:26
danielsit is06:26
fabbioneUnpacking libxau-dev (from .../libxau-dev_6.8.2-16_sparc.deb) ...06:26
danielsthat's why it worked on amd64/i386/powerpc06:26
danielshmmmm.  if it doesn't work with that, I guess I'll have to tighten the build-dep. :\06:27
fabbionei think it needs the new splitted one, doesn't it?06:27
danielsin the meantime, build libxau 1:0.1.2-2 and make sure that gets used06:27
fabbioneyeah ok06:27
danielsyeah, so I'll tighten it up for -2106:27
fabbionebut please tight the dependencies for the next upload06:27
danielswill do06:27
fabbioneit's not too much of a problem for me06:28
fabbionebut a new breezy re-build-strap would fail06:28
=== Kamion_ [~cjwatson@host81-153-126-219.range81-153.btcentralplus.com] has joined #ubuntu-toolchain
=== desrt [~desrt@kopesetik.desrt.ca] has joined #ubuntu-toolchain
fabbionedoko: please pull http://people.ubuntu.com/~fabbione/binary-gcc.mk into gcc-3.4 debian/rules.d/07:36
fabbioneit fixes biarch install (dh_link) target07:37
fabbionethere are a couple of \ missing ;)07:37
=== cartman [foobar@cartman.developer.konversation] has joined #ubuntu-toolchain
=== netjoined: irc.freenode.net -> clarke.freenode.net
=== Kamion_ [~cjwatson@host81-153-126-219.range81-153.btcentralplus.com] has joined #ubuntu-toolchain
=== svenl [~luther@AStrasbourg-251-1-59-4.w82-126.abo.wanadoo.fr] has joined #ubuntu-toolchain
=== ajmitch [~ajmitch@port162-41.ubs.maxnet.co.nz] has joined #ubuntu-toolchain
=== \sh [~shermann@server3.servereyes.de] has joined #ubuntu-toolchain
=== elmo [~james@83-216-141-215.jamest298.adsl.metronet.co.uk] has joined #ubuntu-toolchain
=== lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-toolchain
=== fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain
=== Riddell [jr@80.1.73.118] has joined #ubuntu-toolchain
=== lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-toolchain
=== chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain
fabbionedoes anybody mind if i spin up concordia a bit?11:10
=== fabbione spins it
dokogood morning11:46
\shhey doko11:48
ajmitchhi doko 11:48
=== doko [~doko___@dsl-084-059-040-255.arcor-ip.net] has joined #ubuntu-toolchain
jbaileydaniels: When I added /usr/X11R6/bin back to my path, it seems that /usr/X11R6/lib/X11/xinit/xserverrc referes to a non-existant /usr/bin/X11/X02:04
danielsjbailey: ROCK02:22
Kamionkilling /usr/bin/X11/ this way seems pretty painful, considering that it seems that current Debian policy was written with an eye to *using* /usr/bin/X11/ as part of the migration path away from /usr/X11R6/ ...02:24
=== jbailey still wishes everything we in /opt/PACKAGE/{bin,lib} and stow'd or path'd into place. But ah well.
fabbioneKamion: that part of the policy was written by Branden iirc and he said: "If you kill *X11* we will change policy.. i wrote it :D"02:25
Kamionyeah, I know Branden wrote it, but using /usr/bin/X11/ to help the transition rather than hindered it seemed like a design feature to me at the time02:26
Kamionand still does, fwiw02:26
Kamions/hindered/hinder/02:26
jbaileyWell, policy is a reflection of current best practice.  If X changes what it does, by definition best practice has changed.02:27
KamionI was kind of hoping that eventually we'd have /usr/bin/X11 -> /usr/bin02:28
Kamionso as not to screw over people with hardcoded paths in local config02:29
Kamionor, y'know, things like ssh02:29
Kamionwhich hardcodes /usr/bin/X11/xauth because that was the path that was supposed to keep working02:30
fabbioneKamion: yup... make sense02:30
danielseh, /usr/bin/X11 isn't going away02:31
danielsthe plan is to move everything to /usr/bin and make /usr/bin/X11 a symlink to /usr/bin02:31
danielsbut there are only so many hours in my day, y'know?02:31
Kamionoh, ok, that's exactly the opposite impression to the one I'd got02:32
danielsdude, I don't hate /usr/bin/X11 any more than I hate /usr/X11R602:32
Kamionyou told me in #1685 that /usr/bin/X11/xauth was bad02:32
danielsi told you in #1685 that a hardcoded path in general was probably bad, yes02:33
danielsi have no intention of killing it off completely02:33
Kamionah02:33
danielsi just don't want to put binaries in there02:33
KamionI hadn't realised that, ok02:33
danielsbbiab02:33
=== desrt gives daniels a big hug
=== cartman is now known as namtrac
dokolamont: 4.0 did sucessfully build on hppa/unstable03:55
lamontdoko: which kernel?03:56
lamontor do you mean on the buildd?03:57
dokoLinux pampa 2.6.8.1-pa11 #1 SMP Fri Sep 10 10:40:02 MDT 2004 parisc64 GNU/Linux03:57
dokono, my machine03:57
lamontwow.  that's with the test suite off?  because 2.6.8.1 has lots of issues with the test suite... 03:57
dokono, on03:57
lamonthrm.. maybe it was everything else that it had issues with...03:58
dokoI'll restart it on breezy03:58
=== lamont just did a massive give-back & dep-wait flush on breezy *4, btw
lamontand off to run to town. sigh03:59
dokowait, kdelibs for i386?04:00
dokolamont: ^^^04:00
lamontdaniels needs to upload -21 and quit claiming to build xlibmesa-gl*-dev, then they'll not be in the archive, and kdelibs will build04:02
lamontalternatively, reversing the order of the build-deps will fix it, but they need to be in the other order for debian...04:03
=== namtrac is now known as cartman
dokohmm, talking with riddell04:06
Riddellwhat's up?04:07
dokolamont: no, kdelibs doesn't deal with xlibmesa-gl*-dev, the last qt upload did fix that, and kdelibs should be retried04:13
danielslamont: ?!?!??!?!?!??!05:47
danielslamont: nothing ever happened to xlibmesa-gl-dev05:47
=== daniels looks at debian/control, sees nothing about xlibmesa-glu-dev than Conflicts/Replaces/Provides from libglu-dev-xorg, goes to bed.
danielslamont: by the way, I could stop building xlibmesa-gl-dev, but seriously it wouldn't do you any good05:52
jbaileydaniels: Another upload then sleep? =)05:53
danielsno.  just sleep.05:54
=== chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain
\shxorg is against gcc4 now? doesn't look like06:36
jbaileyfabbione: You there?06:42
fabbionejbailey: i am very close to go and cook dinner...06:42
fabbionei am hungry to deaht06:42
fabbionedeath06:42
jbaileyfabbione: No worries.  Just I did some research on Xen this morning.  I can babble about it another time.06:43
fabbionejbailey: ok.. btw i solved a big problem with latency06:43
fabbionefor some reasons my eth0 on the gw was like adding 800ms to the next eth hop :)06:43
fabbioneifdown / ifup solved :))06:44
jbaileyErr. duplexing?06:44
fabbionenope06:44
fabbioneit started all of a sudden06:44
fabbioneit's connected to a proper swtich06:44
fabbione100Mb FD06:44
fabbioneit's the first time it does something like that06:44
fabbioneanyway dinner...06:44
jbaileyenjoy!06:45
fabbionei will pass by later06:45
fabbioneso you can tell me about xen06:45
fabbionethanks06:46
jbailey20 packets transmitted, 15 received, 25% packet loss, time 19189ms06:46
jbaileyrtt min/avg/max/mdev = 754.573/1149.822/1433.117/166.396 ms, pipe 206:46
fabbionedoko: btw.. did you pull in the fix for gcc-3.4?06:46
jbaileyThat's to your router for me. =)06:46
fabbionejbailey: can you check where you actually lose packets?06:46
fabbionemrt/traceroute...06:46
jbaileyWeird, I dind't have traceroute installed.06:47
dokofabbione: no, jbailey will do the next 3.4 upload06:47
jbaileyfabbione: In which case, the dh_link \'s are already in my tree. =)06:48
fabbionedoko: ok06:49
fabbionejbailey: thanks :)06:49
=== fabbione goes in the kitchen
fabbionedoko: btw.. it did build fine this time :)06:49
fabbioneno probs with livtest06:49
dokoheh, nice06:50
fabbionejbailey: so.. want to tell me that stuff about xen?07:09
jbaileyfabbione: It appears that Xen may run under NPTL these days, but really suckily.07:09
jbaileyI haven't got an answer yet on my questions, but there's a patch to glibc that should be it generally be better.07:10
fabbionejbailey: ah ok.. right now i am slightly more worried of xen not even being able to mount my root partition :)07:10
fabbionei need to do another test next weekend07:10
fabbionegot 2.0.6 ready on my server07:10
fabbionewith 2.4.3007:10
fabbione(can't run 2.6 there yet)07:10
jbaileyfabbione: Right.  I've been hesitant about Xen if we'd wind up having to drop support for it.07:11
fabbionejbailey: i know smurfix and Mith are looking at other options too07:11
fabbionelike L4+afterburner07:11
fabbionethat appears to be much more portable and way less intrusive than xen07:11
jbaileyHah, it would be funny if L4 were useful in the Linux world.07:12
fabbionejbailey: well they run linux at the moment07:12
fabbioneon top of debian07:12
fabbionethat's their devel platform07:12
jbaileyOh?  Interesting.07:12
fabbionethat's also why Benno was at UdU07:12
jbaileyI only know L4 from my Hurd exposure.07:12
fabbioneMith and I were at the University there to see a demo07:13
jbaileyHmm, wish I had knowm about it07:13
fabbioneand it was afterburner + xen/l407:13
fabbioneit was nice.. 07:13
jbaileyAh well. =)07:13
fabbionei mean.. it worked :)07:13
fabbionebut after that they disappeared07:13
fabbionethe userland wasn't ready yet07:14
fabbioneso that's probably why07:14
jbailey'l4 afterburner' is apparently not a useful thing to google for.07:14
fabbionejbailey: afterburner is not very well known at the moment07:14
fabbionethe code has been kept private for a while07:15
fabbionebut they were sorting the last license bits to go public07:15
jbaileyCan you tell me where I could look to find info on it?07:15
fabbioneMith has all the URL's07:15
fabbionei don't have them handy.. sorry07:15
jbaileyTx, I'll poke him later.07:16
jbaileyIt's not critical, I was just looking at the Breezy tasks this morning and figured I'd look into the glibc on Xen issues a bit.07:16
fabbioneyeah but i suggest to wait to push patches around07:17
fabbionespecially if we are not sure we can commit to support it07:17
jbaileyWell, my i386 asm isn't good enough to understand the implications of the patch.07:19
jbaileySo I would need help anyway.07:19
fabbioneehhe neither is mine07:22
jbaileyhey doko.. =)07:22
dokohi07:23
jbaileySprechen ze i386-asm? =)07:23
\shs/ze/Sie/07:27
\sh;)07:27
jbailey\sh: Thanks.  My german is limited to asking if something is vegan.  Ordering by number off a menu.  Asking for Redwine or apple juice.  Looking at someone pleadingly and saying "Haut ban hof?" and telling them that I just shit my pants.07:29
\shhahaha :)07:29
\shjbailey: but in the deepest german slang you say: Sprechen'se <insert your fav language here>07:30
\shjbailey: so u were very close to become an originating NRW inhabitant :)07:31
dokojbailey, no assembler ...07:34
\shlda $0a07:36
\shsta $d02007:36
\shthe last bit of assembler i remember ;)07:36
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain
=== doko [~doko___@dsl-084-059-038-196.arcor-ip.net] has joined #ubuntu-toolchain
=== cartman [foobar@cartman.developer.konversation] has joined #ubuntu-toolchain
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain
dokofabbione, lamont: is 3.4.4 built on hppa and sparc?09:34
desrtdoko; with a few hickups, it looks like linux-2.6.12-rc5 is buildable with my gcc-4 cross-compiler10:01
jbaileyWow, I didn't know the kernel was gcc-4 safe.10:02
desrtjbailey: only very recently10:03
desrtand not entirely... some drivers are still broken10:04
dokodesrt: nice10:05
desrtthe ubuntu kernel config is the ideal stress-test platform :)10:08
jbaileyI can't see Fabio diving at the chance to run an experimental kernel in breezy.10:10
desrt2.6.12 should be soon10:16
lamontdoko: 3.4.4 of what?10:16
lamont gcc-3.4_3.4.4-0ubuntu3 fails on hppa10:17
dokoahh, ok10:20
lamont ../include/linux/byteorder/swab.h:131: error: parse error before '__fswab16'10:23
lamontis that a lkh thing, I wonder???10:23
lamontconfigure: Using /var/lib/gstreamer/0.8 as registry cache dir10:25
lamontconfigure: WARNING: Sissy ! By asking to not build the tests known to fail, you hereby waive your right to customer support.  If you do not agree with this EULA, please press Ctrl-C before the next line is printed.  By allowing the next line to be printed, you expressly acknowledge your acceptance of this EULA.10:25
lamontchecking for valgrind > 2.1... checking for sigaction... yes10:25
=== lamont laughs
lamontbad gstreamer0.810:25
lamontdoko: actually, gcc-{3.3,3.4,4.0} are all ftbfs in their last hppa attempt10:26
lamont(all retrying now, just for giggles)10:26
lamontanyway, off again for a while10:27
jbaileylamont: Possibly.  Why are they using linux/byteorder instead of the glibc functions for that?10:29
jbaileybyteswap.h is a far better choice, I think.10:30
lamontdunno... palo_1.8 on hppa, fwiw10:31
=== lamont must leave. supposed to be 5 min away 2 min ago.
lamontone more massive giveback executed10:33
jbaileyI'll take a look at palo in my chroot on bdale's machine.10:33
jbaileyMight submit a patch for palo isntead, though.10:33
lamontkewlness10:34
lamontthat'd be fine10:34
lamont(remember it was written by a kernel hacker.... so the line may be a bit more blurred than it should...)10:34
=== lamont flees
desrtdesrt@gorecki:~$ uname -a; cat /etc/issue11:43
desrtLinux gorecki.cas.mcmaster.ca 2.6.12-rc5 #1 SMP Wed May 25 17:34:15 EDT 2005 ppc64 GNU/Linux11:43
desrtUbuntu 5.10 "Breezy Badger" Development Branch \n \l11:43
dokoheh ;)11:43
desrtthere's this file in /proc that when you write to it it causes a kernel oops... other than that everything seems ok11:44
desrt(/etc/init.d/networking writes to this file)11:44
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain

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