=== rtcm_ [n=jman@81.84.150.197] has joined #ubuntu-toolchain [01:08] infinity: looks, like I messed up the archive ... did upload both gcc-4.0_4.0.2-4ubuntu3 and gcj-4.0_4.0.2-4ubuntu3, but didn't see the dependency libgcj6: gcc-4.0-base (= 4.0.2-ubuntu3), which did break with the ubuntu4 upload. now ubuntu5 doesn't build anymore ... [01:25] doko : Yeah, we've noticed this. [01:34] infinity: anyway, just uploaded another package to build without gettext [02:17] doko: another _what_ package? === lamont notest that gcj6 is FTBFS on hppa (using the old stuff) [02:30] doko: gcj-4.0 still build-depends debhelper, which is uninstallable... [02:31] or do we just downgrade the chroots, I wonder... [06:24] morning [06:24] doko: why does gjc run the gcc testsuite? [06:24] gcj [06:25] make[3] : Entering directory `/build/sparcbuildd/gcj-4.0-4.0.2/build/gcc' [06:41] doko: ubuntu6 is still FTBFS [06:43] doko : Meh, I'm going to jave to bootstrap this to unbreak it, aren't I? === infinity puts that on his TODO for later tonight. [06:51] doko: please stop uploading now and give us the time to bootstrap it [06:51] otherwise katie is NOT happy about old binaries with new sources around [06:51] infinity: afaict it's debhelper that can't be installed [06:52] checking how to unroll the loop [06:53] Unrolling it is simple enough. [06:53] infinity: yeah it gets down to gettext libgcj6 again [06:53] nothing fancy if you have the old binaries somewhere [06:53] Yup, which is easy enough to fix. [06:54] Easy enough if ou don't have the old binaries, too. [06:54] dpkg has --force flags for a reason. :) [06:56] yeah [06:58] there we go === infinity goes to fix things on the DC arches. === fabbione looks at gcj build log and he can't understand why on earth it's building the all gcc to ship only java [07:04] this is like [07:04] cp gcc... gcj [07:04] vi debian/rules [07:04] upload [07:18] Oh, cock. [07:19] what? [07:19] E: Package libc6-dev-i386 has no installation candidate [07:19] eh? [07:24] Oh, never mind, that's just apt being retarded. [07:48] dpkg-gencontrol: error: package gcc-4.0-hppa64 not in control info [07:48] dh_gencontrol: command returned error code 65280 === lamont-away beats doko with gcj-4.0 [07:49] lamont-away: i won't be surprised if it will FTBFS on all arches at the end of the build [07:50] lamont-away: there is ubuntu6 out anyway [07:50] you might want to give it a shot [07:50] did anyone start uploading new packages for the transition, or is that still waiting? [07:50] that was ubuntu6 [07:50] gcj-4.0_4.0.2-4ubuntu6_20051122-1842 [07:50] i think nobody is uploading giving that everything is FTBFS [07:50] and it will be until this gcj thing is fixed [08:02] Well, gcj is being built right now in the DC. [08:02] So you guys better sort your issues. :) [08:04] infinity: let see if it builds at DC :) [08:04] it might as well be FTBFS there too [08:13] I have faith. === chmj [n=chmj@wbs-146-184-221.telkomadsl.co.za] has joined #ubuntu-toolchain [10:45] Alright, gcj-4.0 has built on all the DC arches. [10:46] infinity: i hate you :) [10:51] Meh, it doesn't matter anyway, doko messed up that upload too. [10:51] ah [10:51] how? [10:51] libgcj6: Depends: gcc-4.0-base (>= 4.0.2-4ubuntu6) but 4.0.2-4ubuntu3 is to be installed === fabbione sighs [11:32] hrm, I shouldn't do uploads in the early morning ... [12:08] lamont-away: what is this thing about gcc-4.0-hppa64? [12:36] doko : Thanks for the upload, I'll bootstrap it in 5 mins. [12:37] infinity: hmm, it should build on it's own? === fabbione restarts the build AGAIN [12:38] debhelper -> po-debconf -> gettext -> libgcj6 [12:38] doko: is there any reason for gcj to build gcc and even run the testsuite? [12:38] Anyhow, seedin chroots and doing manual builds right now. [12:42] fabbione: you have to build gcc to build gcj, yes, I'll disable the testsuite for all but libjava [12:43] doko: can't we use external gcc? [12:43] given that we build it with testsuite and stuff [12:43] Okay, building everywhere. [12:44] doko : Please don't upload again until these binaries are all installed. [12:44] fabbione: but then you'll miss the bootstrap comparision for gcj [12:44] infinity: yes ... [12:44] (including SCC please) [12:45] doko: hmm ok.. [12:50] ubuntu7 building here [12:51] let's hope it's not FTBFS [01:00] Riddell: looks like you can't use 4.0 for KDE on hppa yet. See PR21123 [01:01] fabbione: how far did the previous gcj-4.0 versions build? [01:02] about 3.6MB of log [01:02] probably more [01:02] let me check [01:02] 4708 -rw-r--r-- 1 sparcbuildd sparcbuildd 4808189 Nov 23 06:32 gcj-4.0_4.0.2-4ubuntu5_20051122-2142 [01:02] but then i did stop it [01:02] because of the new versions coming in [01:02] and gcj is not exactly ccache friendly [01:03] It would be if it used the system compiler instead of bootstrapping its own. :/ [01:03] doko : I hope that's something you intend to fix before you push these changes to Debian... [01:03] it appears it can't be done [01:04] Don't see why not. [01:04] fabbione doko: can't we use external gcc? [01:04] doko fabbione: but then you'll miss the bootstrap comparision for gcj [01:05] infinity: so ccache does understand java? [01:06] Where there's a will, there's a way. [01:06] doko : Oh, no, but isn't gcj itself in C? [01:06] think about it ... [01:06] Or does it build with gcj? [01:06] yes, you'll save about 10 source files [01:07] it's libjava which does consume the time [01:07] Well, yes, libjava is huge and painful. [01:07] java doesn't know anything about "preprocessed source", so ccache is not possible [01:07] But you're bootstrapping a full GCC as well here. [01:07] no, not full, just java [01:07] Surely that can be done away with. [01:07] fabbione doko fabbione: but then you'll miss the bootstrap comparision for gcj [01:08] doko : I see a full n-stage xgcc build going on here... [01:09] With some makefile fiddling, I don't see why one couldn't start a bootstrap of gcj using the system C compiler, s'all I'm saying. [01:10] infinity: you can do that, but that's not longer a "bootstrap" [01:10] (The eventual goal being to split out most of the esoteric compilers to their own self-contained source packages) [01:10] doko : Sure, it's a gcj bootstrap, just not a gcc bootstrap. [01:10] no that's crap, and I won't do that [01:10] doko: yeah, I saw the debian guys had problems [01:10] doko: all other arches OK for gcc 4? [01:10] Riddell: looks so [01:11] doko: am I ok to start uploading? [01:11] Riddell : Nothing will build until gcj is in the archive anyway. [01:11] bah, who uses gcj anyway :) === infinity needs to do a mass give-back when that happens. [01:11] Riddell: gettext [01:12] java sucks [01:12] it should die [01:13] infinity: a global give back is fine.. assuming ubuntu7 is ok === fabbione hides from doko === chmj_ [n=chmj@wbs-146-161-75.telkomadsl.co.za] has joined #ubuntu-toolchain === rtcm [n=jman@81.84.150.197] has joined #ubuntu-toolchain === elmo [n=james@83-216-156-21.jamest747.adsl.metronet.co.uk] has joined #ubuntu-toolchain === rtcm_ [n=jman@81.84.150.197] has joined #ubuntu-toolchain [03:51] doko: that's the gcj-4.0 build failure on hppa === lamont-away patiently waits for -4ubuntu7 [03:52] http://buildd.mmjgroup.com/buildLogs/g/gcj-4.0/4.0.2-4ubuntu6/gcj-4.0_4.0.2-4ubuntu6_20051122-1842-hppa-failed.gz [03:53] dh_testroot [03:53] : # provide as and ld links [03:53] dh_link -p gcc-4.0-hppa64 \... [03:53] now that's amusing... for gcj-4.0 :-) [03:54] wow. -4ubuntu7 started building 2.5 hours ago. that's cool [03:55] 2/3 of the way to where it died last time [03:55] infinity: btw, thanks for dealing with all the DC architectures === elmo [n=james@83-216-156-21.jamest747.adsl.metronet.co.uk] has joined #ubuntu-toolchain [04:06] lamont-away: it will fail again [04:06] doko: I see. [04:06] eta for -7? [04:06] er, -8? [04:06] I should just kill the build of -ubuntu7 then? [04:10] lamont-away: infinity did tell me to wait with a new upload until the packages are in the archive [04:11] doko : They're in. [04:15] infinity, fabbione: sparc as well? [04:17] Negative. [04:17] When sparc has gcj binaries uploaded, you'll see them here: [04:17] http://ports.ubuntu.com/ubuntu-ports/pool/main/g/gcj-4.0/ [04:18] I suspect it may have failed to build.. Or fabio's sparc is even slower than I thought. [04:18] lamont-away : What's the URL for ports logs again? [04:18] sparc has not sent a -4ubuntu7 log yet [04:18] buildd.mmjgroup.com/buildLogs/... [04:18] buildLogs/g/gcj-4.0 [04:19] Build needed 05:31:26, 878096k disk space [04:19] that was to the failure point on -4ubuntu6 [04:19] Oh, ouch. [04:19] Kay, it may be a while, then. === lamont-away considers jumping up and down screaming that he "wants -4ubuntu8" [04:20] fabbione: you admitting to being around? [04:20] I think he took off over an hour ago. [04:21] hrm.. actually, that -6 log is 5:31 to the point where he killed it [04:22] and was working on stamps/05-build-stamp at that point [04:22] doko: how confident are you that -4ubuntu7 will build on sparc? [04:23] Modulo goofy packaging errors like hppa, it should work fine. [04:24] lamont-away: I don't see anything special on sparc, it's comparable to i386 and powerpc [04:25] any reason for me to _not_ kill the hppa build of -7? [04:26] no === fabbione yawns [04:29] sparc is building [04:31] doko: and none of the transition uploads have happened yet, correct? [04:33] lamont-away: correct [04:34] fabbione: how big is your log file? [04:35] 2.3M on ubuntu7 [04:35] almost as long as my penis [04:35] it's a long way to go [04:35] what's the problem? [04:36] hppa is ftbfs [04:36] so we're waiting for sparc to upload and unblock before uploading -4ubuntu8 [04:36] do you already have a fix? [04:36] I expect doko knows the fix [04:36] doko: ? [04:36] (it's trying to package something it didn't build...) [04:37] well if you want to upload ubunt8 go ahead [04:37] E: Couldn't find package libgtk+2.0-directfb-dev [04:37] that'd be why cdebconf keeps retrying every cron.daily [04:37] it just means if ubuntu7 does build i can use it from the local cache [04:37] if it doesn't sparc is screwed and we will ubuntu9 [04:37] just go for it [04:38] fabbione: sounds OK to me... doko? [04:38] infinity: any objections? [04:38] None here. I already got all my binaries in. :) [04:38] meh [04:38] i am not sure how long it takes to build.. and i am heading out to have dinner with my wife [04:39] fabbione: you killed -6 at 5.5 hours [04:39] so it's unlikely i will look at the buildd for the next 12 hours or more [04:39] ah, cool. [04:39] -5 went much deeper in the build [04:39] doko: ready for that upload... [04:40] 2142->0632 on that one [04:40] about 9 hours [04:40] yeah and 4.8M of log [04:40] before sbuild received SIGINT -- shutting down [04:40] gcc-4 takes about 19 hours to build here [04:40] so you can make your calculation [04:41] but again just go ahead [04:41] i will sort it out later on [04:41] gotta run and fast [04:41] i am late :) [04:41] cya later or tomorrow [04:44] fabbione, lamont-away: yes, I have a fix [04:44] doko: may as well upload - fabbione will use -4ubuntu7 locally if it succeeds [04:48] and if you upload in the next 11 minutes or so, it'll make the next cron.daily [04:51] lamont-away: done [05:23] err [05:23] I've got a bunch of c2a libraries in new [05:23] should I be not processing them? [05:24] elmo: yes, should be safe, all buildd's are running the new gcc packages [05:25] doko: kthx [05:37] doko: hppa is running the old gcc [05:37] since that's what has to be installed to build gcj [05:37] gcj-4.0 build started [05:37] lamont-away: ok, but you can shut it down/upgrade after the gcj-4.0 build [05:38] doko: yes [05:38] given a .deb, how can I tell whether it is pre-transition or post-transition? [05:46] it depend on one of the libraries, which we are going to rename [05:58] ok [05:58] is krb4 one such animal, I wonder [05:58] ?> [06:00] no [06:25] elmo: please sync boost, overwriting Ubuntu changes [06:48] boost_1.33.0-5.. [06:48] ..failed updating 2 targets... [06:48] ...skipped 2 targets... [06:48] ...updated 1487 targets... [06:48] make: *** [build-stamp] Error 1 [06:48] doko: for extra credit, make it compile on hppa [06:50] interesting... I wonder who uploaded it, since all Ihave is a build failure [08:30] echo timestamp > javax/xml/datatype.stamp [08:30] make[4] : *** [javax/xml/datatype.stamp] Segmentation fault [08:30] ew [08:35] lamont-away: didn't see that in unstable [08:35] yeah, I've pushed it to the top of the hill again [09:33] ok ubuntu7 did build [09:33] so i guess i can just build ubunut8 and get along with it === doko_ [n=doko@dslb-084-059-072-254.pools.arcor-ip.net] has joined #ubuntu-toolchain === ubuntulog [n=ubuntulo@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain === Topic for #ubuntu-toolchain: Ubuntu Toolchain - The GNU Compiler Collection, Binutils, glibc, and moon-buggy. === Topic (#ubuntu-toolchain): set by infinity at Wed Nov 9 18:03:55 2005 === lamont-away is still known as lamont-away [10:37] fabbione: you could even just drop ubuntu7 in your local cache and let it go from there [10:37] and 8 will "just happen" [10:38] lamont-away: that's what i did basically [10:38] but i am prioritizing ubuntu8 to unsuck the other buildd [10:38] that's waiting for 7/8 [10:38] but 7 will never hit archive [10:38] so... [10:38] 8 it is [11:01] Riddell: ping [11:01] doko: hi [11:05] doko: hmm? [11:10] Riddell: ? [11:14] doko: you pinged? [11:25] ahh, yes, arts upload? [11:30] Riddell: ^^^ [11:30] doko: arts and kdelibs uplo [11:30] doko: arts and kdelibs uploaded [11:32] Riddell: thanks [11:39] elmo: please process the changed allocator packages in NEW, should be nearly all libs in main [11:47] ahh, crap, these are already in, but in universe ... [11:48] which are? [11:49] mdz: please can you promote: libgdome2-cpp-smart0c2a libid3-3.8.3c2a libwpd-stream8c2a libwpd8c2a, or do you need inclusion reports? [11:49] Riddell: ^^^ [11:50] also libarts1c2a and kdelibs4c2a [11:51] also libakode2 and libakode-dev which are about to replace akode and akode-dev [11:51] Riddell: they are not yet in the anastacia output [11:51] ok, just fore-warning :) [11:55] doko: 2nd try is past the first failure [11:58] doko: the usual rules apply; if they're just package renames and not new code, they don't need reports [11:58] doko: however, I can't do any promotions right now; elmo needs to do some maintenance on katie [12:01] Riddell: amarok-gstreamer shouldn't depend on mp3 stuff [12:01] mdz: I'll investigate [12:02] thanks