| jbailey | lamont: *poke* | 01:28 |
|---|---|---|
| doko | jbailey: I cannot reproduce fabbione's build failures of 3.4 on sparc, so the toolchain seems to be in shape for all archs except hppa ... | 01:39 |
| jbailey | doko: Great. I'm just about to do the hppa work. | 01:40 |
| jbailey | doko: Did you and lamont touch the -0ubuntu4 glibc that I put on chinstrap to enable biarch amd64? | 01:41 |
| jbailey | Oo, looks like goto might have already done a bunch of the porting work. | 01:44 |
| === jbailey builds what Debian has on hppa to see. | ||
| doko | jbailey: no | 01:45 |
| doko | gcc cannot do biarch for hppa | 01:45 |
| jbailey | doko: 'kay. The last point was the libgcc1 depends on amd64-libs, and I'm conflicting against it. | 01:45 |
| jbailey | No, I meant the biarch i386/amd64 stuff | 01:45 |
| doko | haven't tried yet, I'm currently working on the C++ ABI transition | 01:47 |
| jbailey | It works fine. I need to know what you want to do about the amd64-libs confict. I'd like to just conflict against it, and ask for it to be removed from breezy. | 01:47 |
| jbailey | But I don't want to cause you more problems enabling the biarch compiler. | 01:48 |
| jbailey | Or I can drop this all until the C++ transition is done. | 01:48 |
| doko | yes, the latter would be nice | 01:48 |
| jbailey | 'kay. I will get hppa in shape today hopefully and get a glibc uploaded. | 01:49 |
| jbailey | I have ppc64 ready to go and amd64 ready to go as soon as you have time for either of those. | 01:49 |
| doko | yes, let's decouple these, although svenl is poking about ppc biarch | 01:51 |
| jbailey | No prob decoupliung them, but there's no reason not to do them in the same day. | 01:52 |
| jbailey | Given that we have to coordinate them at the time, it might be easiest if one of us handled them, or if we locked down a time for you and I to sit down at the same time and slug through them when we're ready. | 01:52 |
| doko | feel free to update the gcc packages :-) | 01:52 |
| jbailey | Cool, I can do that. | 01:53 |
| jbailey | I'll do the c++ transition with you first though. | 01:53 |
| doko | jbailey: looks like it's finished for the libraries in main, although the motus may need some support converting the remaining universe libraries | 01:55 |
| jbailey | WOw, that's incredible. | 01:57 |
| jbailey | Did you sleep? =) | 01:57 |
| doko | a bit | 01:59 |
| jbailey | When are you bumpding gcc-defaults? | 02:01 |
| doko | anyway, I'll sleep *now* | 02:02 |
| doko | when lamont and elmo are awake and syncs & uploads for same packages are frozen, maybe Mo/Tu | 02:02 |
| jbailey | Cool, good sleeps doko. | 02:04 |
| === Riddell_ [jr@muse.19inch.net] has joined #ubuntu-toolchain | ||
| lamont | jbailey: note that the new glibc is still waiting for a usable libgcc1 | 02:09 |
| jbailey | lamont: Right, Sounds like we'll wait until later this week when gcc-defaults bumps, and I'll pick a time to sit down with you and Just Do It from all sides, including gcc. | 02:11 |
| === doko [~doko___@dsl-082-082-197-172.arcor-ip.net] has joined #ubuntu-toolchain | ||
| === infinity stares at doko. | ||
| jbailey | infinity: It's not nice to stare at someone when they're sleeping. | 04:35 |
| infinity | When did we go from "help us out on Monday/Tuesday" to "finished for the libraries in main"? | 04:35 |
| infinity | Oh well. I'd feel more guilty, if I hadn't been moving all weekend. | 04:38 |
| === jbailey fires up an hppa glibc build with Carlos' patches. | ||
| jbailey | Off to watch a show while this builds. | 05:51 |
| svenl | hi jbailey | 07:01 |
| svenl | jbailey: did you not want to send me some new biarch glibc packages ? | 07:01 |
| jbailey | svenl: I posted them, didn't I? | 07:11 |
| jbailey | Nope, I didn't. | 07:11 |
| jbailey | svenl: Sending, I'll let you know when it's done. | 07:12 |
| jbailey | Hmm. I think I need to spend part of tomorrow moving glibc bits into baz or something. | 07:16 |
| jbailey | svenl: http://testhaus.cns.utoronto.ca/~jbailey/ppc64-nptl/ | 07:22 |
| jbailey | svenl: Bed time now. I'll be back in around 6 hours. | 07:23 |
| svenl | jbailey: ok, thanks. | 07:32 |
| svenl | I want the source packages also :) | 07:35 |
| === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
| === chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain | ||
| === Riddell [jr@muse.19inch.net] has joined #ubuntu-toolchain | ||
| doko | morning all | 11:05 |
| chmj | morning doko | 11:07 |
| doko | hi chmj | 11:08 |
| doko | infinity: there are more things to do ... | 11:08 |
| Riddell | doko: what's the status of the toolchain transition? has any qt/kde stuff been uploaded or is that tomorrow? | 11:48 |
| doko | Riddell: qt and kdelibs are prepared, amu did a qt 3.3.4 merge | 11:50 |
| doko | Riddell: please search the BTS for gcc-4.0 and prepare patches for the kde specific things. I don't know, if amu did work on these | 11:50 |
| Riddell | doko: how can I change over to g++ 4 to test things? | 11:51 |
| doko | https://www.ubuntulinux.org/wiki/CxxLibraryList | 11:52 |
| doko | see the sources line at the top | 11:52 |
| Riddell | thanks | 11:54 |
| chmj | doko, is the g++4.0 app test build been done ? | 11:55 |
| doko | chmj: which one? | 11:56 |
| chmj | I mean,all apps. there is surpose to be an automatic rebuild right ? | 11:57 |
| chmj | so we can see which ones fail? | 11:58 |
| doko | chmj: yes, this one is prepared as well. Riddell, do you want to do this for the KDE apps? | 11:59 |
| Riddell | doko: sounds like a good idea, what would I have to do? | 12:00 |
| doko | Riddell: make a chroot, add my test builds to it and rebuild all of KDE in main ;) | 12:01 |
| Riddell | doko: I'll give it a shot | 12:02 |
| chmj | doko, what do you mean prepared? | 12:06 |
| doko | chmj: sources ready for upload | 12:07 |
| chmj | doko: sweet | 12:08 |
| doko | svenl: ocaml FTBFS with 4.0 | 12:23 |
| svenl | doko: mmm. | 12:58 |
| svenl | doko: on ubuntu ? | 12:58 |
| svenl | doko: i will have a look later on ppc. | 12:58 |
| svenl | doko: or maybe you do have a log ? | 12:58 |
| svenl | doko: what version of ocaml is this anyway ? | 12:58 |
| doko | svenl: should fail in unstable/experimental as well. | 12:59 |
| svenl | doko: well. | 01:00 |
| svenl | doko: i prefer working on the ubuntu/breezy machine, i have killed the hoary install by installing random glibc/gcc upgrades anyway. | 01:00 |
| svenl | doko: i will contact upstream if we have something serious and get a patch. | 01:03 |
| fabbione | re | 01:15 |
| fabbione | doko: as i said.. right in time :) | 01:15 |
| === doko [~doko___@dsl-082-082-212-246.arcor-ip.net] has joined #ubuntu-toolchain | ||
| svenl | doko: if i have both gcc-4.0 and gcc-3.4 installed, which one will it default to ? | 01:40 |
| jbailey | svenl: That's up to gcc-defaults | 01:41 |
| svenl | jbailey: hi. | 01:41 |
| jbailey | svenl: Check to see what /usr/bin/gcc points to. | 01:41 |
| svenl | jbailey: so how do i make sure gcc-4.0 is used ? CC=gcc-4.0 ? | 01:41 |
| jbailey | Yup | 01:41 |
| jbailey | Although in Breezy, that's the default now. | 01:41 |
| jbailey | CXX is still gcc-3.4, though | 01:41 |
| jbailey | err. | 01:41 |
| jbailey | 3.3 | 01:41 |
| svenl | gcc-3.3 | 01:41 |
| doko | ocaml? no, there's something like -cc gcc-4.0 in debian/rules | 01:41 |
| doko | svenl: deb http://people.ubuntu.com/~doko/GCC-4.0/i386 ./ | 01:42 |
| svenl | launch started. | 01:42 |
| svenl | doko: what should i do with that ? I have some serious doubt it will be usefull on my powerbook. | 01:43 |
| doko | deb http://people.ubuntu.com/~doko/GCC-4.0/powerpc ./ | 01:43 |
| svenl | doko: not the official gcc-4.0 in breezy ? | 01:44 |
| doko | you did want to change the defaults, didn't you? | 01:44 |
| svenl | nope, its ok, i use the -cc gcc-4.0 trick, should do just fine. | 01:45 |
| svenl | ok, build launched, will tell you how it goes. Is using jbailey's latest glibc. | 01:45 |
| jbailey | svenl: the ppc64 snap that I gave you? | 01:46 |
| svenl | doko: i would really like to build a biarch compiler myself though. | 01:46 |
| svenl | jbailey: sure. | 01:46 |
| jbailey | Cool. | 01:46 |
| svenl | jbailey: i killed my hoary install anyway, upgraded to breezy, and installed your snaps. | 01:46 |
| doko | why do you try to compile ocaml on powerpc? | 01:46 |
| jbailey | svenl: Remember to pin glibc. | 01:46 |
| doko | it works | 01:46 |
| svenl | doko: do you think i will have more chances in the gcc-4.0 ? | 01:46 |
| jbailey | svenl: There's a newer one than that in the archive that doesn't have ppc64 support. | 01:46 |
| svenl | doko: because i have powerpc machines. | 01:47 |
| svenl | jbailey: mmm. | 01:47 |
| svenl | jbailey: how do i pin glibc ? | 01:47 |
| jbailey | svenl: Something in some apt configuration. In practice, I just don't run apt-get upgrade after. | 01:48 |
| svenl | jbailey: :) | 01:48 |
| svenl | jbailey: maybe better would be to always rebuild newest ppc64 enabled glibc for each official version, so you don't have this problem. | 01:48 |
| jbailey | svenl: I was thinking earlier that I might move the glibc packaging into bzr so that it's easy for me to do updates of the biarch i386/amd64 and ppc/ppc64 glibcs that aren't in the archive yet. | 01:49 |
| svenl | jbailey: when will they be in the archive ? | 01:49 |
| svenl | doko: do you mean the gcc-4.0 on powerpc will not manifest the ocaml bug ? | 01:50 |
| jbailey | svenl: amd64 support will go in the archive after the c++ transition is finished (I don't want to interfere with that) | 01:50 |
| jbailey | svenl: ppc64 will go in the archive as soon as we figure out how to make the biarch compiler build on a 32 bit kernel. | 01:50 |
| jbailey | (subject to the same limitation as amd64 beyond that) | 01:51 |
| doko | svenl: http://people.ubuntu.com/~lamont/buildLogs/o/ocaml/3.08.3-3/ | 01:51 |
| svenl | doko: oh. | 01:56 |
| svenl | doko: well. | 01:56 |
| svenl | doko: i guess gcc-4.0 is buggy on i386. | 01:56 |
| svenl | bng_ia32.c: In function 'bng_ia32_mult_add_digit': | 01:57 |
| svenl | bng_ia32.c:111: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' | 01:57 |
| svenl | well. | 01:57 |
| svenl | doko: i don't deal with register starved broken archs :) | 01:57 |
| svenl | doko: i can't really help you there, i have one amd64 with a pure64 install and only ppc boxes apart from that, and one m68k. | 01:59 |
| doko | svenl: install a i386 chroot on your amd64 ;-P | 02:00 |
| jbailey | svenl: Going to start the Ubuntu m68k port? =) | 02:00 |
| svenl | doko: i would write to http://pauillac.inria.fr/bin/caml-bugs | 02:00 |
| svenl | jbailey: nope. | 02:00 |
| === jbailey would love to see m68k and arm as the first buntu targets. | ||
| svenl | jbailey: i lost my apus card, so only an mmu-less 68020 in it. | 02:01 |
| svenl | oh, but i have an amiga 4000 waiting for me in oberursel. | 02:01 |
| svenl | will try to get it next week. | 02:01 |
| svenl | doko: i could do that. | 02:01 |
| svenl | doko: not today though. | 02:04 |
| svenl | doko: what is your email address ? | 02:06 |
| fabbione | hey doko | 02:06 |
| fabbione | hi svenl | 02:06 |
| svenl | doko@ubuntu.com ? | 02:06 |
| doko | hi fabbione | 02:06 |
| doko | svenl: yes | 02:07 |
| svenl | doko: does gcc 4.0 eat up more registers than 3.4 used to do ? | 02:09 |
| svenl | doko: i just filled a bug upstream, CCed you on it even. | 02:10 |
| doko | svenl: thanks | 02:10 |
| doko | <doko> fabbione: you can find the gcc-3.4 build for sparc in my home on chinstrap. did build properly in a fresh breeeezy chroot | 02:56 |
| jbailey | fabbione: YEs, and regular sparc ideally, but still need to figure out whether we continue to care about pre sparc9 | 02:57 |
| fabbione | i did complete the build manually | 02:58 |
| jbailey | (For those that are confused, consider this the equivalent of renaming a file in CVS) | 02:58 |
| fabbione | jbailey: pre v9 are 32 bits, right? | 02:58 |
| fabbione | if so just kill them | 02:58 |
| jbailey | fabbione: 32 bits only. | 02:58 |
| fabbione | i don't care | 02:58 |
| jbailey | All the classic sparc joy. | 02:58 |
| fabbione | perfect.. KILL THEM ALL | 02:58 |
| jbailey | 'kay. Then in that case, we have opt packages for sparcv9 and sparcv9b. | 02:59 |
| jbailey | Should the sparcv9 one go away? | 02:59 |
| fabbione | whatever fits you better is ok with me :) | 02:59 |
| fabbione | see.. i am very simple | 02:59 |
| fabbione | doko: what kind of chroot did you creat? | 03:00 |
| doko | breezy | 03:01 |
| fabbione | doko: try to bootstrap a buildd chroot :) | 03:01 |
| doko | why? | 03:01 |
| fabbione | debbootstrap --variant=buildd | 03:01 |
| fabbione | it's a different set of base packages? | 03:01 |
| fabbione | doko: at what time are we going to start? | 03:02 |
| doko | sorry, why is it different to deboostrap breezy and install the build-deps? | 03:03 |
| === lamont wanders off for a while | ||
| fabbione | doko: because breezy has plenty of things that are not installed on a buildd | 03:03 |
| doko | fabbione: elmo around, lamont around, xorg packages updated | 03:03 |
| doko | what about sparc-utils? | 03:04 |
| fabbione | also sparc-utils | 03:04 |
| fabbione | you need to install that manually on the buildd... i forgot to add it to the list when i did debootstrap ubuntu34 | 03:04 |
| doko | why is sparc-utils not in the buildd? | 03:04 |
| fabbione | because i forgot to add it to the list when i did debootstrap ubuntu34 | 03:05 |
| jbailey | afk a sec. | 03:05 |
| fabbione | doko: when are we going to start the transition? | 03:06 |
| fabbione | (also.. elmo is not here ;)) | 03:07 |
| doko | fabbione: elmo needs to freeze the import of C++ packages, lamont needs to prepare the buildd's, daniels needs to upload xorg compiled with 4.0 | 03:08 |
| fabbione | is it going to happen today? | 03:08 |
| doko | I have my doubts now ... | 03:10 |
| fabbione | well fuck i come back from holidays for this transition | 03:10 |
| === fabbione goes for a smoke | ||
| fabbione | checking whether the C++ compiler (g++-4.0 -O -DDEBIAN ) works... no | 03:25 |
| fabbione | configure: error: installation or configuration problem: C++ compiler cannot create executables. | 03:25 |
| fabbione | hmmm | 03:25 |
| doko | which package? | 03:27 |
| fabbione | mozilla | 03:27 |
| fabbione | same error on all arches | 03:27 |
| doko | yes, firefox works, mozilla maybe needs an update | 03:30 |
| jbailey | fabbione: Well, now my client won't talk to CAnonical's imap server. | 03:46 |
| jbailey | *cRy* | 03:46 |
| fabbione | jbailey: get a serious client... | 03:47 |
| jbailey | fabbione: I use evo because that's what I have to support. It makes sense for me to know it very well. | 03:47 |
| === jbailey turns on various debugs. | ||
| fabbione | jbailey: well.. i use thunderbird.. it's crap.. it crashes, but i can open N imap mboxes without any problem :) | 03:49 |
| jbailey | fabbione: This was all working on my laptop on Friday. | 03:49 |
| jbailey | No idea why it deteriorates on my main box. Main difference is that the laptop is i386 and this is ppc. | 03:49 |
| jbailey | fabbione: Oh ouch. This client is going through and doing a STATUS on each file in buildLogs. =( | 04:21 |
| fabbione | amen | 04:21 |
| fabbione | no wonder it takes ages | 04:21 |
| fabbione | well i am not too worried about bw usage... | 04:21 |
| fabbione | it will just be slow for you | 04:21 |
| fabbione | jbailey: if you prefer i can even setup a forward to you, but that means getting all the logs from the buildd... | 04:29 |
| jbailey | I'm trying to beg for better behaviour in #evolution | 04:31 |
| fabbione | doko: wouldn't be a good idea to start bitching elmo, lamont and daniels? | 04:32 |
| fabbione | otherwise you can forget to start the transition :) | 04:33 |
| doko | :( | 04:40 |
| fabbione | should we revisit all the steps, one by one just to be 200% sure before we start? | 04:56 |
| lamont | doko: "prepare the buildds" == ?? | 05:08 |
| fabbione | lamont: i think he means installing the new gcc-defaults | 05:10 |
| lamont | fabbione: ok | 05:11 |
| === lamont needs to run to the post office this morning sometime, but that's only 15 min from home | ||
| fabbione | lamont: we are still waiting for elmo, so i think it's safe to go right away :) | 05:12 |
| fabbione | lamont: do you think it's worth to let the new gcc-defaults in so that it can be built? | 05:12 |
| fabbione | but not installing it in the chroots? | 05:12 |
| fabbione | that would probably save sometime for the process | 05:12 |
| lamont | the chroots will install it at the start of a build if it's a dependency. Likewise, around 0215 DCT (Data Center Time), the chroots are automatically upgraded | 05:13 |
| fabbione | ok than we should wait i guess | 05:15 |
| fabbione | but nothing build-dep on gcc-defaults | 05:15 |
| fabbione | it's pulled in as part of build-essential iirc | 05:15 |
| lamont | libtool: build-depends gcj which comes from gcc-defaults. | 05:23 |
| fabbione | hmm right | 05:25 |
| fabbione | yeah it does here too | 05:26 |
| fabbione | ops | 05:26 |
| lamont | yeah - that's currently uninstallable, hence the ease of finding it... :-) | 05:28 |
| lamont | if I have 40 minutes, I'm going to run away and visit the postoffice. | 05:30 |
| fabbione | lamont: probably much more than that :) | 05:31 |
| lamont | right. back in a bit then | 05:31 |
| fabbione | oky | 05:31 |
| jbailey | lamont: Around? | 06:21 |
| lamont | yo | 06:21 |
| jbailey | lamont: Are you brave enough to step me through reading hppa assembler, or should I wait until Carlos is around? =) | 06:22 |
| lamont | hppa assembly is my frien d | 06:22 |
| jbailey | lamont: Lovely, might be something for the evening hacking session then. | 06:23 |
| lamont | woot | 06:23 |
| jbailey | lamont: The story so far is basically: | 06:23 |
| lamont | still no elmo? | 06:23 |
| jbailey | 10: 00000000 0 NOTYPE GLOBAL DEFAULT UND memcmp | 06:23 |
| jbailey | 11: 00000000 764 FUNC GLOBAL HIDDEN 1 __GI_memcmp | 06:23 |
| fabbione | apparently no | 06:23 |
| jbailey | lamont: And the similar file on ppc has it defined. So I think I need to see what's up. =) | 06:23 |
| lamont | hrm.. that's not assembly... that's .o format. :-) | 06:25 |
| jbailey | Right, but I want to see what it's feeding to the assembler. | 06:25 |
| jbailey | For the symbol to not exist at all. | 06:25 |
| lamont | ok. note that gas and I are not always good friends, but the actual machine code is one of my oldest friends... | 06:27 |
| jbailey | 'kay. =) | 06:28 |
| === doko [~doko___@dsl-082-082-213-004.arcor-ip.net] has joined #ubuntu-toolchain | ||
| fabbione | still no elmo? | 07:04 |
| fabbione | lamont: still around? | 07:20 |
| fabbione | bah eek.. brb | 07:20 |
| lamont | heh | 07:20 |
| === jbailey can smell the pizza in the oven. | ||
| fabbione | lamont: we don't know where elmo is, do we? | 07:28 |
| === lamont bets on .uk. | ||
| lamont | (that'd be a "no") | 07:29 |
| fabbione | ahaha | 07:29 |
| lamont | mind you, I could be wrong... :-) | 07:29 |
| fabbione | sorry... phone call | 07:30 |
| fabbione | re | 07:33 |
| fabbione | lamont: i need to go offline soon | 07:33 |
| fabbione | and without elmo we cannot do the transition | 07:33 |
| fabbione | lamont: can you check if you still have access to the sparcbuildd | 07:33 |
| fabbione | and stop it when needed | 07:33 |
| lamont | have I mentioned that I hate transitions that require manual steps on all the buildd's? | 07:35 |
| lamont | esp since hppa isn't able to do the walk at the same time as the rest? | 07:35 |
| fabbione | lamont: yes and i am with you | 07:35 |
| fabbione | since sparc is the slowest | 07:35 |
| === lamont glares at doko/jbailey, just for good measure | ||
| lamont | fabbione: just vultus5 | 07:36 |
| lamont | ? | 07:36 |
| lamont | br | 07:37 |
| lamont | b | 07:37 |
| fabbione | lamont: yes.. there is only vultus5 | 07:37 |
| lamont | coolness | 07:37 |
| lamont | back in a couple | 07:37 |
| === jbailey runs off for lunch, ecaping the glares from lamont. | ||
| doko | leave me alone ... | 07:37 |
| doko | jbailey: stop! | 07:37 |
| fabbione | later guys | 07:37 |
| jbailey | doko: Eh? | 07:37 |
| fabbione | doko: try to summon me.. if i won't respond in a decent time lamont will stop the buildd | 07:37 |
| fabbione | and we will look at it tomorrow... | 07:38 |
| fabbione | but this really suck | 07:38 |
| doko | jbailey: please have a look at the latest directfb build failure on amd64. | 07:38 |
| fabbione | and sparc.... | 07:38 |
| doko | conflicting types in {sys,asm}/types.h ? | 07:38 |
| fabbione | bbl | 07:38 |
| jbailey | doko: Does it have to be before lunch? The timer just rang. | 07:38 |
| doko | fabbione: have fun | 07:38 |
| doko | no | 07:39 |
| jbailey | Lovely. =) | 07:39 |
| === jbailey goes for lunch. =) | ||
| === lamont is informed that lunchtime approaches | ||
| === lamont makes sure that jbailey has his cell, on the off chance that elmo surfaces before he gets back from lunch... | ||
| === lamont goes to lunch with the wife, back in a while - holler if/when I'm needed | ||
| jbailey | Enjoy. | 08:42 |
| jbailey | doko: Around? | 09:20 |
| doko | jbailey: yes | 09:25 |
| jbailey | Was just following glibc on hppa and noticed that upstream actually released the new binutils. Do you want me to add that to my queue of things to test? | 09:32 |
| jbailey | doko: I'll look at directfb on ppc first, since I have one of those here. | 09:35 |
| doko | jbailey: yes, these are already packaged and tested. one regression on ia64, just waiting for elmo's ok | 09:40 |
| jbailey | Nice. Did you see http://sources.redhat.com/ml/binutils/2005-05/msg00360.html ? | 09:40 |
| doko | interesting | 09:44 |
| jbailey | directfb is a cast-as-lvalue failure, I'm trying to figure out what the current best practice is supposed to be. | 10:28 |
| jbailey | data8 = (unsigned char *)data16 = (void*)0; | 10:29 |
| jbailey | Isn't allowed anymore. | 10:29 |
| doko | jbailey, did you get the .22 version from the archive | 10:45 |
| doko | jbailey: ^^^ | 10:45 |
| doko | current dirctfb version should be 0.9.22-0ubuntu1 | 10:56 |
| jbailey | Ah, nope. apt-get source had fetched me .20-0ubuntu1 | 10:58 |
| === jbailey fetches. | ||
| jbailey | err .20-5 | 10:59 |
| === 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 | ||
| === Topic (#ubuntu-toolchain): set by jbailey at Wed May 11 01:20:59 2005 | ||
| === fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain | ||
| fabbione | morning | 06:28 |
| fabbione | damn my server died again | 06:28 |
| fabbione | how is it going guys? | 06:28 |
| === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
| svenl | jbailey: hi. | 08:17 |
| svenl | Mmm, gcc 4.0 is better, since it builds libgcc1 package, but not the 64bit stuff. | 08:30 |
| svenl | need to find how gcc decide wheter to build lib64gcc1 or not, doesn't seem obvious. | 08:45 |
| === chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain | ||
| chmj | morning | 09:27 |
| === Seveas [~seveas@ksl403-uva-131.wireless.uva.nl] has joined #ubuntu-toolchain | ||
| === doko [~doko___@dsl-082-082-205-075.arcor-ip.net] has joined #ubuntu-toolchain | ||
| fabbione | hey doko | 10:19 |
| chmj | hey doko | 10:19 |
| doko | hi fabbione | 10:20 |
| fabbione | doko: so what is the status for the transition? | 10:21 |
| doko | fabbione: let me wade through my mail first ... | 10:26 |
| fabbione | ok | 10:26 |
| === Riddell [~jr@jasmine.wyrdweb.com] has joined #ubuntu-toolchain | ||
| === cartman [foobar@cartman.developer.konversation] has joined #ubuntu-toolchain | ||
| svenl | doko: hi. | 11:29 |
| svenl | doko: i built gcc-4 biarch. | 11:29 |
| svenl | doko: it is better, since there is a non-empty libgcc1, but still no 64bit libs. | 11:29 |
| svenl | doko: i have searched some, but am at a loss on where in the gcc makefiles the building of the 64bit stuff is defined. Do you have any hint on that ? The debian stuff tries to build the 64bit packages but fails. | 11:30 |
| svenl | doko: also, i was told to use --added-target=powerpc64-linux when building by hand. I see no trace of this in the debian files. | 11:32 |
| svenl | --enable-targets=powerpc64-linux that is. | 11:33 |
| === Riddell_ [jr@muse.19inch.net] has joined #ubuntu-toolchain | ||
| === Seveas [~seveas@ksl403-uva-131.wireless.uva.nl] has joined #ubuntu-toolchain | ||
| === daniels [~daniels@amnesiac.heapspace.net] has joined #ubuntu-toolchain | ||
| daniels | represent | 02:08 |
| === mvo [~egon@ip181.135.1511I-CUD12K-01.ish.de] has joined #ubuntu-toolchain | ||
| doko | fabbione, daniels, lamont, jbailey, elmo: ping | 02:12 |
| lamont | Running /build/buildd/gcc-4.0-4.0.0/src/gcc/testsuite/gcc.c-torture/compile/compile.exp ... | 02:13 |
| lamont | make[1] : *** [stamps/06-check-stamp] Terminated | 02:13 |
| lamont | make: *** [check] Terminated | 02:13 |
| lamont | Build killed with signal 15 after 150 minutes of inactivity | 02:13 |
| === lamont grumbles at doko | ||
| lamont | morning doko | 02:13 |
| daniels | doko: wassup | 02:13 |
| === elmo [~james@83-216-141-215.jamest298.adsl.metronet.co.uk] has joined #ubuntu-toolchain | ||
| lamont | morning elmo | 02:14 |
| doko | time plan for tomorrow ... | 02:14 |
| === lamont is about to get dragged off for breakfast for 20-30 minutes | ||
| doko | 1) freezing the imports/syncs 2) upgrading the buildd's 3) freezing the archive for C++ application uploads 4) library uploads | 02:16 |
| doko | 1) and 2) make only sense, if things for 4) are ready, xorg is outstanding. for 1) and 2) we need lamont and elmo | 02:17 |
| daniels | xorg is going to be fun | 02:17 |
| daniels | we need to get 3 arch: all NEWs in, then xorg | 02:17 |
| fabbione | doko: why not today? | 02:18 |
| doko | fabbione: daniels prepares the final xorg source upload | 02:18 |
| fabbione | daniels: i understood it was ready.... | 02:19 |
| daniels | i've been working with joshtriplett so we have the same stuff between debian and ubuntu | 02:19 |
| daniels | and i want to try some more upgrade tests | 02:19 |
| daniels | in any case, it's 2220 here and I'm fucking tired, so I'd love if I could do it tomorrow | 02:19 |
| daniels | but if it needs to be done tonight, sure | 02:20 |
| lamont | doko: I'd also be interested in hearing how this will work for the buildd's that don't even have a gcc-4.0 right now. | 02:20 |
| lamont | (hppa) | 02:20 |
| doko | lamont: is hppa ok? | 02:20 |
| lamont | hppa has the build failure pasted aboge | 02:20 |
| lamont | above, even | 02:20 |
| doko | ohh, they don't? | 02:20 |
| lamont | no. still don't | 02:20 |
| lamont | back in about 20 minutes. sorry | 02:21 |
| doko | lamont: disable the testsuite :-( | 02:21 |
| fabbione | well the sooner the better | 02:21 |
| fabbione | daniels: what is missing from your packages? | 02:22 |
| doko | hmm, ok. so let's sort out hppa later | 02:22 |
| fabbione | only upgrade tests? | 02:22 |
| daniels | fabbione: yeah, and I need to reversion it too | 02:24 |
| fabbione | daniels: but are you going to upload the monolithic tree right? | 02:24 |
| fabbione | reversion it? | 02:24 |
| daniels | the monolithic tree ... plus modular packages of xc/include/*.h and lib/xtrans/* | 02:25 |
| fabbione | wouldn't be wise to just do the C++ transition with the monolithic and split everything else later? | 02:25 |
| daniels | unfortunately I only did the C++ transition later | 02:26 |
| daniels | and it would take more time to split out all the patches with gcc4 fixes etc and fix all the offsets than it would to test all this | 02:26 |
| fabbione | is there any option to upload only libglu from the splitted? | 02:26 |
| daniels | hmm? | 02:27 |
| fabbione | iirc that's the only library that has to do the C++ transition | 02:27 |
| fabbione | xorg 6.8.2-$foo | 02:27 |
| fabbione | -> libgluwhateversoname-$foo | 02:27 |
| daniels | yeah, it's libglu1-xorg now | 02:27 |
| daniels | from xlibmesa-glu | 02:28 |
| fabbione | if you upload libgluwhateversoname-$bar | 02:28 |
| fabbione | where $bar is > $foo | 02:28 |
| fabbione | that's all you need to start with | 02:28 |
| fabbione | but can libglu1-xorg build abe uploaded that way? | 02:28 |
| daniels | from the modular tree? nope | 02:28 |
| fabbione | if so there is no need to get the entire splitted tree in | 02:28 |
| daniels | mesa is going to be the hardest part of this lot | 02:28 |
| fabbione | doko: how much can we do without X? | 02:29 |
| doko | not much, the whole KDE depends on it. and other libs as well. | 02:30 |
| fabbione | KDE is an application :) | 02:31 |
| fabbione | what about the other libs? | 02:31 |
| fabbione | how many of them? | 02:31 |
| daniels | it's ok, we can do libglu | 02:31 |
| daniels | so, i'll try to get x-common, x11proto-core-dev and xtrans-dev in tonight | 02:32 |
| daniels | and get packages of xorg up on chinstrap for testing and building on powerpc | 02:32 |
| daniels | and then upload that tomorrow morning? | 02:32 |
| doko | fabbione: it doesn't make sense to have the archive in an unusable state for too long. this just adds 24 hours | 02:32 |
| doko | daniels: sounds ok. do the packages need new love? i.e. is elmo awake tomorrow morning? | 02:33 |
| daniels | the first 3 packages need NEW love | 02:33 |
| fabbione | they will mostlikely need elmo's love for NEW | 02:33 |
| fabbione | all the sources are new | 02:33 |
| daniels | libglu1-xorg will need NEW love for the binary package too tho | 02:33 |
| daniels | would preseeding be useful here? | 02:34 |
| fabbione | that's sure thing | 02:34 |
| fabbione | elmo will still need to do the manual uni->main | 02:34 |
| daniels | ahr | 02:34 |
| doko | same for all these new library binary packages ... | 02:34 |
| daniels | elmo: will you be around for the next couple of hours? | 02:35 |
| elmo | mostly, yes | 02:35 |
| doko | elmo: how long does it take for a package to enter main from universe? | 02:36 |
| daniels | ok, so if I upload x-common, x11proto-core-dev and xtrans-dev, would you be able to new them? | 02:36 |
| elmo | daniels: yes | 02:37 |
| daniels | phat, thanks | 02:37 |
| jbailey | doko: I'm here now, sorry about the lag. | 02:39 |
| doko | jbailey: elmo delegated the decision about binutils in breezy to us | 02:40 |
| jbailey | doko: Lovely, I'll go over the list again of what changed. | 02:40 |
| elmo | (FWIW, I think it's a no brainer; if Debian wasn't frozen, I'd be uploading it) | 02:41 |
| doko | there's one regression in the testsuite on ia64 | 02:41 |
| fabbione | is there somebody that still cares about ia64? | 02:42 |
| doko | <doko> drow: one binutils 2.16 ld testcase did regress on ia64, compared to 2.15 | 02:42 |
| doko | <drow> ? | 02:42 |
| doko | <doko> +Running /home/doko/binutils/binutils-2.16/ld/testsuite/ld-bootstrap/bootstrap.exp ... | 02:42 |
| doko | <doko> PASS: bootstrap | 02:42 |
| doko | <doko> PASS: bootstrap with strip | 02:42 |
| doko | <doko> -PASS: bootstrap with --static | 02:42 |
| doko | <doko> +FAIL: bootstrap with --static | 02:42 |
| doko | <drow> wird | 02:42 |
| doko | fabbione: lamont? | 02:42 |
| daniels | i bet the latest hj lu binutils would fix that | 02:42 |
| daniels | let's use that | 02:42 |
| elmo | HA FUCKING HA | 02:43 |
| doko | :-))) | 02:43 |
| jbailey | mmm.. crack. | 02:43 |
| fabbione | ahahah | 02:43 |
| jbailey | elmo: The biggest concern I have is yet another variable in all the toolchain changes. | 02:44 |
| doko | jbailey: what the ppc64 patch applied to the 2.16 branch? | 02:44 |
| elmo | [I'm off to get some breakfa^Wlunch, bbiab] | 02:44 |
| jbailey | doko: Required for glibc to build acc. to modra. | 02:44 |
| === lamont back | ||
| === doko decides to go for lunch if elmo is back ;) | ||
| lamont | fabbione: I still care about ia64. moreso now. | 02:46 |
| fabbione | lamont: ehehe ok :) | 02:46 |
| lamont | more to the point, I can't evangelize Ubuntu with my team when it doesn't even run on the platform in question...... | 02:58 |
| fabbione | lamont: make sense | 03:00 |
| jbailey | lamont: It generally works... ;) | 03:00 |
| jbailey | lamont: Thinking of which. Do you know efi at all? =) | 03:01 |
| lamont | I know of it. I expect to know it well sometime this year... | 03:01 |
| jbailey | <voice who="Mr. Burns">Eexcellent</voice> | 03:02 |
| fabbione | ahah | 03:02 |
| jbailey | The grub folks have decided that they'd rather not use libefi because of copyright assignment hassles. | 03:03 |
| jbailey | So I may have questions. | 03:03 |
| fabbione | daniels: do you have i386 binaries somewhere for xorg? | 03:03 |
| fabbione | daniels: i have a machine or 2 i can trash testing an upgrade | 03:03 |
| daniels | fabbione: only amd64 at the moment ... i'm building up an i386 chroot because my laptop just ran out of space mid-build | 03:04 |
| fabbione | ENOAMD64 | 03:04 |
| fabbione | and i guess you can't build on concordia.. right? | 03:04 |
| jbailey | daniels: Are you looking for testers? I can do ppc if you want (and can build it quite fast) | 03:05 |
| fabbione | doko, jbailey: did anybody fixed the ppc64 chroot on davis? | 03:07 |
| daniels | fabbione: concordia doesn't have a chrot | 03:07 |
| daniels | although, erk | 03:07 |
| fabbione | daniels: yes it does... | 03:07 |
| daniels | packages will take a fucking long time on my system | 03:07 |
| daniels | since I can't strip on an XFS /home | 03:07 |
| daniels | fabbione: an i386 chroot? | 03:07 |
| fabbione | yeps | 03:07 |
| lamont | daniels: breezy-i386 isn't there??? | 03:07 |
| jbailey | fabbione: I didn't know there was one. | 03:08 |
| fabbione | linux32 dchroot -c breezy-i386 | 03:08 |
| fabbione | jbailey: it's where your ppc64 libc packages are installed... | 03:08 |
| jbailey | fabbione: Oh? I didn't know that they had been installed anywhere. | 03:09 |
| daniels | ah shit, need build-deps first | 03:09 |
| daniels | elmo: ping | 03:09 |
| jbailey | fabbione: Thinking of which, I'd like to upload a glibc that wakes up hppa. | 03:12 |
| jbailey | fabbione: No sense wasting cycles building it on sparc, though. You'll get your love later in the week. | 03:12 |
| === Seveas [~seveas@dyn127.roaming.few.vu.nl] has joined #ubuntu-toolchain | ||
| lamont | doko: you have a gcc-4.0 upload planned anytime soon? | 03:14 |
| doko | lamont: no | 03:14 |
| jbailey | lamont: I do after the transition gets underway. | 03:14 |
| doko | besides the one I did one minute ago | 03:14 |
| lamont | ok. if turning off the testsuite for hppa fixes it, I'll have jbailey include that in his upload | 03:14 |
| jbailey | Maybe Friday? For getting biarch working on i386/amd64 again. | 03:14 |
| jbailey | Ouch. | 03:15 |
| jbailey | lamont: BTW, I've asked Carlos whether there's suckage-reduction for hppa in binutils 2.16 that we really need. | 03:15 |
| fabbione | jbailey: ok. i will kill it as soon as it arrives here | 03:16 |
| fabbione | no actually i can't | 03:16 |
| fabbione | because that would make locales uninstallable | 03:17 |
| fabbione | and some packages build-dep on it | 03:17 |
| jbailey | Oh right, bugger. | 03:17 |
| fabbione | no big deal.. | 03:17 |
| fabbione | it's ccached :) | 03:17 |
| jbailey | 'k | 03:18 |
| jbailey | I'd do a build locally on my u5, but I haven't wired it up yet. | 03:19 |
| fabbione | jbailey. don't worry :) | 03:19 |
| lamont | fabbione: the other trick is to just keep the right locales locally... | 03:20 |
| fabbione | lamont: yeah taht would work too, but it's not a big deal | 03:20 |
| fabbione | libgc is faster and ccachable | 03:20 |
| lamont | yeah | 03:21 |
| fabbione | the real issue is gcc... | 03:21 |
| fabbione | it takes ages | 03:21 |
| fabbione | and it's not ccache friendly | 03:21 |
| lamont | doko: 2) upgrading the buildd's | 03:22 |
| lamont | is that just to get the new gcc-defaults stuff there? | 03:22 |
| lamont | well, and g++-4.0 | 03:22 |
| doko | lamont: yes | 03:22 |
| fabbione | lamont: should we review debootstrap in the meantime? | 03:22 |
| lamont | ok. We'll also want to make sure that debootstrap gets some love sometime this week | 03:22 |
| doko | oops, yes, g++-4.0 is not in main? | 03:22 |
| === jbailey reads this as "last chance for a bathroom break" =) | ||
| lamont | doko: is in main | 03:23 |
| daniels | elmo: any chance of getting some thpethul chroots, or telling everyone else to fuck the fuck off out of them? | 03:23 |
| === fabbione goes for more coffee before we start | ||
| lamont | I more meant that gcc-defaults would drag in g++-4.0 | 03:23 |
| daniels | xorg build now kind of requires /usr/{lib,include,share}/X11 to be directories | 03:24 |
| lamont | daniels: WTH is it looking at the real directories, instead of it's copies? | 03:25 |
| daniels | lamont: hm? | 03:25 |
| lamont | and what are they if not directories, anyway? | 03:26 |
| lamont | <daniels> xorg build now kind of requires /usr/{lib,include,share}/X11 to be directories | 03:26 |
| daniels | right | 03:26 |
| daniels | they used to point to /usr/X11R6/... | 03:26 |
| daniels | which is an utter anachronism | 03:26 |
| daniels | and the transition is sort of starting now | 03:26 |
| lamont | ah, right | 03:26 |
| jbailey | daniels: I wonder if that is *supposed* to make me feel old... ;) | 03:27 |
| lamont | jbailey: nah - you _are_ old | 03:27 |
| daniels | jbailey: r7 coming yo' way | 03:27 |
| jbailey | daniels: Do you mean in general, or do you want a test build/use cycle on a ppc? | 03:29 |
| daniels | jbailey: in general | 03:29 |
| jbailey | Ah. =) | 03:29 |
| doko | lamont: do you want the testsuite disabled on hppa? then maybe you have to cancel the current build | 03:30 |
| === cartman [foobar@cartman.developer.konversation] has left #ubuntu-toolchain ["Ich] | ||
| lamont | doko: I'll live | 03:30 |
| doko | who's in your way? | 03:30 |
| lamont | doko: running a testbuild with the suite disabled now, we'll see how well it does. Once I have that, then I really don't mind if it's out of date for a day or 6... or do I need something in the last upload? | 03:30 |
| lamont | freshening the local mirror (for source) now. | 03:31 |
| fabbione | ahh almost nothing is better than dark choccolate + biscuits + coffee | 03:38 |
| === chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain | ||
| jbailey | How does that far side comic go? The ideal life of the archaeologist, a beautiful woman in one arm, and the focilised skull of a homo habilis in the other... | 03:40 |
| chmj | O.O | 03:41 |
| lamont | doko: no need to kill any builds, since hppa hasn't managed to _start_ autobuilding of breezy yet. | 03:43 |
| lamont | (still trying to get a gcc-4.0, you see... and then there were the glibc issues, that jbailey fixed yesterday) | 03:43 |
| jbailey | lamont: Ah, are you running on that glibc now? | 03:44 |
| lamont | yes | 03:44 |
| lamont | well, the buildd chroot is. | 03:44 |
| jbailey | That'll certainly exercise it. =) | 03:44 |
| lamont | yeah. built glibc, doxygen, l-k-h, and now chunking on gcc-4.0. | 03:45 |
| lamont | of course, maybe it's glibc's fault that gcc-4.0 hangs in the test suite. | 03:45 |
| lamont | nah.. | 03:45 |
| doko | lamont, if you get the buildd to set an env var WITHOUT_CHECK=yes ... | 03:45 |
| jbailey | lamont: I did a summary of the testsuite changes from the previous glibc. If you're interested I can /msg them to you. | 03:47 |
| lamont | doko: actually, I just created -0ubuntu2hppa1, with "check_no_cpus := hppa # arm m68k" | 03:47 |
| daniels | fabbione: do you have a chroot you want to fuck shit up on? | 03:47 |
| fabbione | daniels: sure | 03:47 |
| lamont | jbailey: actually, email would be even better | 03:47 |
| daniels | grab the packages from p.u.c/~daniels/newx/ | 03:47 |
| daniels | build and install x-common, x11proto-core and xtrans, in that order | 03:47 |
| daniels | you should have /usr/{lib,include,share}/X11 as a directory | 03:47 |
| daniels | and I'll throw a xorg source package (and later binaries) up as soon as it's finished building here | 03:48 |
| jbailey | lamont: lamont@u.c? | 03:48 |
| lamont | daniels: the packages take care of transitioning those links->dirs, yes? | 03:48 |
| lamont | jbailey: sure | 03:48 |
| daniels | lamont: yah | 03:48 |
| daniels | x-common does, and x11proto-core-dev is a replacement for the old x-dev | 03:48 |
| daniels | it should dist-upgrade cleanly; i'm just building a chroot to test that hypothesis now | 03:49 |
| jbailey | lamont: cym | 03:49 |
| lamont | tnx | 03:49 |
| fabbione | daniels: bootstrapping the chroot now | 03:49 |
| daniels | cool | 03:49 |
| fabbione | Setting up x-common (1.0) ... | 03:58 |
| fabbione | ok what should i check after this install | 03:59 |
| fabbione | it's a completely empty package? | 03:59 |
| fabbione | except the copyright) | 04:00 |
| daniels | oh, fucking shit | 04:00 |
| daniels | yeah, I just noticed that myself | 04:00 |
| daniels | had the real package on my laptop, sigh | 04:01 |
| daniels | ok, new one uploaded -- sorry | 04:01 |
| fabbione | x-dev is empty? | 04:03 |
| fabbione | daniels: i builded all the 3 sources.. and installed them | 04:07 |
| daniels | yep, just depends on x11proto-core-dev | 04:07 |
| fabbione | other than x-dev is empty.. the others look ok | 04:07 |
| fabbione | want to give me libglu? | 04:07 |
| daniels | that's no accident :) | 04:07 |
| daniels | still in the xorg source package; i can't run debuild -S until this build has finished, for obvious reasons | 04:07 |
| fabbione | i understood that libglu was already splitted... | 04:08 |
| daniels | er, no | 04:09 |
| daniels | it changed its package name, that's it | 04:10 |
| daniels | mesa is going to be one of the hardest things to split, because it's so deeply tied in with the x server at the moment | 04:10 |
| fabbione | ok so what do you want me to build now? | 04:10 |
| daniels | i'm in dh_builddeb of xorg, so I'll give you that next | 04:11 |
| fabbione | ok | 04:11 |
| daniels | since it needs elmo to install build-deps in the concordia chroot | 04:11 |
| fabbione | i guess now xorg build-deps on these 3 new packages.. right? | 04:12 |
| daniels | it build-deps on x11proto-core-dev and xtrans-dev | 04:12 |
| daniels | which in turn build-dep on x-common | 04:12 |
| fabbione | right | 04:13 |
| fabbione | do you want to uplaod the source somewhere? | 04:13 |
| fabbione | so i can start downloading it? | 04:13 |
| fabbione | at least the orig | 04:13 |
| daniels | yep, just waiting for dh_builddep to finish | 04:13 |
| daniels | the orig is just the same | 04:13 |
| fabbione | ah ok | 04:13 |
| daniels | uploading that over my DSL would take about an hour | 04:13 |
| daniels | there we go, finished building | 04:13 |
| elmo | remoo | 04:15 |
| daniels | hello sunshine | 04:15 |
| fabbione | doko: do yo have a gcc-defaults ready? | 04:17 |
| fabbione | doko: if so can you handle the i386 version to me please? | 04:17 |
| daniels | fabbione: xorg sources up | 04:17 |
| fabbione | downloading now | 04:18 |
| fabbione | daniels: i just need the new gcc-defaults to be sure that we are building with g++4.0 | 04:18 |
| daniels | sure | 04:19 |
| === fabbione summons doko | ||
| === fabbione hits doko with a get-here-bat | ||
| === fabbione power ups the sodomotron and inserts doko's coordinates in the system | ||
| === lamont wonders if overfiend knows fabbione has the somdomotron | ||
| fabbione | lamont: i have the EU version :) | 04:27 |
| fabbione | that runs on 220v/50hz | 04:27 |
| jbailey | Does it have gears? | 04:27 |
| fabbione | jbailey: do you happen to have gcc-defaults? | 04:27 |
| fabbione | the new one? | 04:27 |
| jbailey | Is that different that what's at deb http://people.ubuntu.com/~doko/GCC-4.0/powerpc ./ ? | 04:28 |
| === lamont is more partial to the colostomizer | ||
| fabbione | jbailey: thanks.. it seems to be the correct one | 04:29 |
| jbailey | "We call this gun the fecalator. One look at it and the target shits his or her self." | 04:31 |
| fabbione | daniels: building x now | 04:32 |
| daniels | cool | 04:32 |
| === doko [~doko___@dsl-084-059-063-208.arcor-ip.net] has joined #ubuntu-toolchain | ||
| fabbione | hmmm | 04:33 |
| fabbione | this is not going to work | 04:34 |
| doko | ? | 04:35 |
| fabbione | daniels: the problem seems to be installing xorg build-deps | 04:36 |
| daniels | fabbione: hm? | 04:36 |
| fabbione | daniels: xorg build-deps on some of its own packages | 04:36 |
| fabbione | that depends on xorg-common | 04:37 |
| fabbione | that fails to install due to x-common | 04:37 |
| daniels | oh, cock | 04:37 |
| daniels | cock, cock, cock, cock, cock, cock, cock | 04:37 |
| fabbione | xorg-common preinst error: /usr/include/X11 exists and is not a symbolic link; | 04:37 |
| fabbione | this package cannot be installed until this directory is removed | 04:37 |
| daniels | right | 04:37 |
| daniels | which is fixed in -11's xorg-common | 04:37 |
| fabbione | cock | 04:38 |
| fabbione | because that can't be built | 04:38 |
| daniels | correct, for ten points | 04:38 |
| fabbione | so we need a transitional package | 04:38 |
| fabbione | or | 04:38 |
| fabbione | make x-common Provides: Conflict: Replaces: xorg-common ? | 04:39 |
| daniels | hmmm | 04:39 |
| fabbione | tho i am not 100% sure that's enough | 04:39 |
| daniels | what's even in xorg-common? | 04:39 |
| fabbione | a bunch of files iirc | 04:39 |
| daniels | shit, mainly conffiles | 04:39 |
| daniels | moving them is way too hard to consider right now | 04:39 |
| doko | fabbione, jbailey, lamont: I'm not at the TBM tonight, but will read my backlog later | 04:40 |
| daniels | i suppose I could upload -10.1 with the relaxed check, get that built everywhere, and then we can do the rest | 04:40 |
| fabbione | i am not sure i will be at TBM either | 04:40 |
| daniels | but MY GOD THAT'S NASTY | 04:40 |
| jbailey | I'll be there. Anything you want mentioned specifically? | 04:40 |
| fabbione | daniels: i don't think there is any other way | 04:40 |
| daniels | fabbione: le sigh | 04:40 |
| lamont | TBM? | 04:41 |
| fabbione | daniels: let's think for a minute or 2 | 04:41 |
| daniels | lamont: tech board | 04:41 |
| lamont | tech board | 04:41 |
| lamont | doh | 04:41 |
| fabbione | Tech Board Mee | 04:41 |
| jbailey | lamont: Took me a sec too. =) | 04:41 |
| lamont | yeah - kept thinking of Martin | 04:41 |
| jbailey | As much fun as visiting MArtin would be. =) | 04:41 |
| fabbione | daniels: i think that we can make it easier, but we need to be extremely syncronized | 04:41 |
| daniels | fabbione: how so, though? | 04:42 |
| daniels | we'll still need to get -11 through the buildds somehow | 04:42 |
| fabbione | daniels: let's say we upload x-common that PRC xorg-common, we upload the new xorg and we reupload the new x-common | 04:42 |
| daniels | hm | 04:42 |
| daniels | i have an idea | 04:42 |
| fabbione | without PRC | 04:42 |
| daniels | maybe x-common could pre-depend on xorg-common | 04:42 |
| daniels | so the apt run is unpack/configure xorg-common, unpack/configure x-common, then do the rest | 04:42 |
| fabbione | that's even more scary :) | 04:43 |
| daniels | so xorg-common's postinst would get run when the symlinks were still present | 04:43 |
| fabbione | let me try | 04:43 |
| daniels | then when -11 hits the archive, we could remove it | 04:43 |
| fabbione | Predepends: xorg-common | 04:43 |
| fabbione | or was it Pre-Depends? | 04:43 |
| daniels | Pre-Depends | 04:44 |
| fabbione | just a sec... | 04:46 |
| === fabbione performs the big purge of death manually | ||
| daniels | heh :) | 04:46 |
| fabbione | well that seems to work | 04:47 |
| === daniels beams. | ||
| fabbione | at least installing only x-common and xorg-common | 04:47 |
| fabbione | now let me see if i can install all the build-deps | 04:47 |
| daniels | yep | 04:47 |
| === lamont mumbles bad things about uploading packages just to get past the buildd | ||
| daniels | lamont: i'm sorry, I'm a bad man | 04:48 |
| fabbione | lamont: yes i agree.. the best would be a transitional xorg | 04:48 |
| === lamont just feels sorry for the alpha/mips/m68k ports next month, when they have to recreate the whole thing as part of a bootstrap | ||
| lamont | then again, if the new one is just plain ftbfs, that's actually not _too_ bad. | 04:49 |
| daniels | lamont: the whole xorg thing? | 04:49 |
| doko | lamont: you still have to time to mumble before hppa is in shape again? ;-) | 04:49 |
| lamont | it's when it builds, wrong, that it's really evill | 04:49 |
| lamont | doko: gcc-4.0 takes _forever_ to build | 04:49 |
| fabbione | nah it's building wrong.. | 04:49 |
| lamont | of course, it'd go faster if I wasnt' building a test kernel, too. | 04:50 |
| fabbione | lamont: 17 hours on sparc :) | 04:50 |
| lamont | fabbione: sparc sucks | 04:50 |
| fabbione | daniels: xorg building now | 04:50 |
| lamont | gcc-4.0: 04:02:12 (2 entries, sigma 01:53:34) | 04:50 |
| fabbione | lamont: sorry.. but i can't find a single _hppa.deb on ports.... | 04:50 |
| === doko is watching, if sparc or hppa wins | ||
| fabbione | doko: you better STFU :P | 04:51 |
| fabbione | doko: or do i need to remind you that thanks to a missing build-dep on gcc-4 sparc didn't make hoary? | 04:51 |
| fabbione | :) | 04:51 |
| lamont | fabbione: dunno if my key is there yet, or if I'm a muppet | 04:51 |
| fabbione | daniels: well.. it SEEMS to work | 04:51 |
| fabbione | xorg is building | 04:52 |
| daniels | oh dear :) | 04:52 |
| daniels | x-common 0.99 it is, then | 04:52 |
| fabbione | well i am just a bit scared of the Pre-Depends to be hounest, but apt-get should do the right thing | 04:52 |
| daniels | it's scary, but less scary than uploading a whole new xorg imo | 04:53 |
| fabbione | daniels: s/scary/timeconsuming | 04:53 |
| fabbione | doko: i think we are ready with X | 04:53 |
| fabbione | doko: your call now :) | 04:53 |
| fabbione | daniels: xorg-common is arch: all | 04:54 |
| fabbione | so once it's builded on i386, we can basically upload x-common 1.0 | 04:54 |
| daniels | fabbione: sure, but everyone get a whole new set of binary packages | 04:54 |
| daniels | unless we binary-NMUed xorg-common with source changes | 04:54 |
| daniels | but that would be REALLY REALLY BAD | 04:55 |
| doko | fabbione, lamont: assume we can rebuild all C++ libraries on all architectures for the release archs, I'd like to continue building the C++ apps, but if sparc and hppa didn't finsish with the libs at this time, you have to make sure that no C++ app is built before. can you manage this for your buildds? | 04:55 |
| fabbione | doko: not sure... | 04:55 |
| fabbione | lamont: do we have a way to filter the output from wanna-build -d breezy --take ? | 04:55 |
| doko | fabbione: I'm away in one hour ..., let's start this at 22:00 UTC, if X is ready | 04:55 |
| daniels | i can upload x-common 0.99 now if we're confident with it | 04:56 |
| daniels | then x11proto-core-dev and xtrans-dev, then xorg | 04:56 |
| fabbione | daniels: if you need to go to sleep, please make all the sources available on people | 04:56 |
| daniels | sure | 04:56 |
| fabbione | daniels: i would rather prefer to get them in at the right time | 04:56 |
| daniels | i'm good for a couple more hours now | 04:56 |
| fabbione | daniels: apparently doko is going away | 04:57 |
| daniels | so we can at least get the prerequisites in, elmo can give them some NEW loving, and we can throw xorg in when I wake up in about 7 hours | 04:57 |
| lamont | doko: if you have a list of the library and app source packages, then sure. no problem. | 04:57 |
| daniels | fabbione: have you got an X build going? | 04:57 |
| fabbione | daniels: yes | 04:57 |
| === lamont hopes to have a breezy hppa buildd up sometime soonish | ||
| doko | lamont: ok, I'll update these when we start | 04:57 |
| lamont | elmo: do the lib packages all wind up sorted ahead of the rest of the world? | 04:57 |
| === lamont doubts that | ||
| doko | lamont: yes, why not? | 04:58 |
| daniels | fabbione: cool | 04:58 |
| fabbione | daniels: prepare x-common 0.99 on people with the Pre-Depends | 04:58 |
| fabbione | daniels: if you are asleep i will upload the packages for you in order | 04:59 |
| daniels | sure | 04:59 |
| fabbione | just sign them all | 04:59 |
| fabbione | so i can just lftp from there | 04:59 |
| lamont | doko: because it would be too convenient. :-) | 04:59 |
| daniels | ok | 04:59 |
| fabbione | lamont: can't we just add a filter to wanna-build? | 04:59 |
| daniels | i'll put all the sources on people, and signed changes files in my homedir on chinstrap | 04:59 |
| lamont | fabbione: so the plan is 2 xorg uploads? | 04:59 |
| fabbione | lamont: no. one. | 04:59 |
| doko | lamont: let's see, at least on i386 they all build, and I did check half of it for amd64 as well. | 04:59 |
| daniels | lamont: one xorg upload, two x-common uploads | 05:00 |
| fabbione | lamont: x-common -> xorg -> x-common | 05:00 |
| lamont | fabbione: I suspect we could, but the source package names are not necessarily aligned well with whether or not they happen to deliver a binary lib package. | 05:00 |
| fabbione | lamont: i was hoping for doko to give us the list of sources :) | 05:00 |
| lamont | doko: was talking about wanna-build, not compiles | 05:00 |
| fabbione | doko: you have a list of sources, don't you? | 05:01 |
| lamont | (specifically doubting that the answer to my question to elmo was 'yes') | 05:01 |
| doko | https://www.ubuntulinux.org/wiki/CxxLibraryList | 05:01 |
| doko | http://www.ubuntulinux.org/wiki/CxxApplicationList | 05:01 |
| lamont | fabbione: so add CxxApplicationList to @no_auto_build | 05:02 |
| lamont | until the libs are all built | 05:02 |
| doko | I'm updating these when we freeze | 05:02 |
| fabbione | doko: does the App lists includes universe? | 05:02 |
| doko | yes | 05:02 |
| daniels | fabbione: x-common 0.99 and 1.0 both on p.u.c/~daniels/newx/, signed changes files for everything on chinstrap:~daniels | 05:02 |
| fabbione | doko: also the library list? | 05:02 |
| fabbione | daniels: rocking. | 05:03 |
| fabbione | daniels: good night kid | 05:03 |
| doko | fabbione: maybe I should split this one into two | 05:03 |
| daniels | eh, I'm still good for a couple of hours yet | 05:03 |
| fabbione | keep the mobil phone on :) just in case ;) | 05:03 |
| daniels | any reason why I shouldn't upload x-common 0.99 now? | 05:03 |
| doko | fabbione: read! | 05:03 |
| daniels | fabbione: yeah, it'll be next to my bed | 05:03 |
| doko | daniels: cool! | 05:03 |
| fabbione | doko: ETOOLAZY :P | 05:03 |
| doko | so who wakes up daniels first? ;-) | 05:04 |
| fabbione | daniels: because it will change the symlinks to dir on installed systems, where Xorg is not ready? | 05:04 |
| daniels | ack, wait, I think I have broken Pre-Depends, lemme check | 05:05 |
| daniels | fabbione: ah, but it won't get installed | 05:05 |
| daniels | because nothing will depend on it yet | 05:05 |
| doko | daniels: what's the name and version of the glu-dev package? | 05:06 |
| daniels | doko: libglu-dev-xorg, 6.8.2-11 | 05:06 |
| daniels | currently it's xlibmesa-glu-dev 6.8.2-10 | 05:06 |
| fabbione | daniels: better to wait... | 05:07 |
| doko | ok, just for the wiki and thighended build deps. so xlibmesa-glu-dev (>= 6.8.2-11) should be still fine? | 05:07 |
| fabbione | lamont: where do i define the @no_auto_build ? | 05:07 |
| daniels | doko: libglu-dev-xorg (>= 6.8.2-11) | 05:08 |
| lamont | buildd.conf | 05:08 |
| lamont | # list of packages which shouldn't be picked up by buildd | 05:08 |
| lamont | @no_auto_build = qw(); | 05:08 |
| lamont | @weak_no_auto_build is the list of packages to build when there is _nothing_ else to do | 05:08 |
| fabbione | yup.. found it | 05:09 |
| daniels | fabbione: don't worry about it now, but I've just updated my xorg sources on p.u.c and chinstrap to rename libglu1-xorg-dbg to libglu1-dbg-xorg and libglu-xorg-dev to libglu-dev-xorg. no build changes. | 05:13 |
| fabbione | doko: before you go away... | 05:14 |
| fabbione | the gcc-defaults and build-essential on people.u.c.~doko/GCC-4.0/source are the ones that need to hit the archive? | 05:15 |
| fabbione | daniels: did you also rename all the install/links/etc files? | 05:15 |
| doko | yes, nobody did complain about these | 05:15 |
| daniels | fabbione: yep | 05:15 |
| fabbione | doko: ok, is there anything we can start to do while you are away? | 05:16 |
| lamont | no cxx libs in multiverse? | 05:16 |
| doko | lamont: yes, only three in the list | 05:16 |
| lamont | ah, and multiverse sorted before universe. sigh | 05:16 |
| === lamont tends to sort in ogre-model order | ||
| doko | fabbione: if you want to start, upload the library packages for main (except kdelibs) from chinstrap:~doko/cxxsrc | 05:17 |
| lamont | doko: those are signed and all? | 05:17 |
| lamont | and note that the buildds aren't updated yet... | 05:18 |
| fabbione | doko: don't we need to switch gcc-defaults first? | 05:18 |
| lamont | are we ready for that? | 05:18 |
| doko | but xorg should be installed on the buildd's before | 05:18 |
| fabbione | no i don't think we are | 05:18 |
| doko | fabbione: sure | 05:18 |
| fabbione | ok stop | 05:18 |
| lamont | WHY??? <doko> but xorg should be installed on the buildd's before | 05:18 |
| fabbione | let's start again the list of things that needs to be done and in what order | 05:18 |
| lamont | fabbione: ++ | 05:18 |
| fabbione | because i think we are skipping some steps here | 05:18 |
| doko | lamont: because I did not updated the build-deps for libraries, which depend on libglu-dev-xorg (>= 6.8.2-11) | 05:19 |
| fabbione | 1) we need to stop the syncs from debian of all the libs and app from the 2 lists on the wiki | 05:19 |
| doko | fabbione: elmo did prepare this, and stop the import of NEW sources | 05:19 |
| fabbione | elmo: is that in place already? | 05:20 |
| elmo | yes | 05:20 |
| fabbione | ok | 05:20 |
| elmo | well err, no | 05:20 |
| fabbione | ok :) | 05:20 |
| elmo | I only did the first list no the wiki | 05:20 |
| elmo | s/no/on/ | 05:20 |
| elmo | 12:12 <elmo> https://www.ubuntulinux.org/wiki/CxxLibraryList | 05:20 |
| elmo | ^-- that one | 05:20 |
| fabbione | doko: do we need to stop syncing the applications too i guess... | 05:21 |
| doko | yes, the current one has to be replaced by the second one, if we start building the libs | 05:21 |
| fabbione | doko: i am not sure i understand what you mean by second one.... | 05:22 |
| fabbione | daniels: xorg is FTBFS here | 05:22 |
| fabbione | Xrandr.c:32:36: error: X11/extensions/Xrender.h: No such file or directory | 05:22 |
| doko | the one at http://www.ubuntulinux.org/wiki/CxxApplicationList | 05:23 |
| doko | containing all packages with a dependency on libstdc++5 | 05:23 |
| daniels | fabbione: hrm, doesn't happen here | 05:24 |
| daniels | fabbione: can you check whether Xrender.h is in /usr/include/X11/extensions or /usr/X11R6/include/X11/extensions please? | 05:24 |
| fabbione | daniels: it's in /usr/X11R6/include/X11/extensions | 05:25 |
| daniels | ok, I'll sort that out | 05:25 |
| fabbione | doko: ok i understand now | 05:26 |
| fabbione | so the next step would be to update the buildd's with new gcc-defaults and build-essential + the CxxApp list to @no_auto_build | 05:28 |
| fabbione | once we have done that... | 05:29 |
| fabbione | we need to upload x-common x-x11proto-core and xtrans | 05:29 |
| fabbione | xorg | 05:29 |
| fabbione | new x-common | 05:29 |
| fabbione | and after that we are ready to go with doko's libs... | 05:30 |
| fabbione | once all the libs are built, we can release apps... | 05:31 |
| fabbione | doko: ? | 05:31 |
| fabbione | did we miss anything? | 05:31 |
| jbailey | Which parts the rest of us do. =) | 05:32 |
| jbailey | Looking through libs, looks like almost everything but kde is done. | 05:32 |
| lamont | jbailey: once we start building libs, aggressively attack any FTBFS packages | 05:34 |
| === jbailey still dreams of an rss feed, and a one click 'claim' option. I've wanted that for new arch porting in Debian for ages too. =) | ||
| fabbione | lamont: afaik doko did build all of them already | 05:35 |
| lamont | on at least i386, yes | 05:36 |
| fabbione | well... should we start? | 05:37 |
| lamont | elmo: any great ideas besides @no_auto_build for blocking builds of the CxxApps? | 05:37 |
| lamont | jbailey: that's what IRC is for. :-) | 05:37 |
| lamont | just pick a letter, it's all yours./ | 05:37 |
| jbailey | Pitty the poor sod stuck with 'l' =) | 05:38 |
| lamont | l != lib* | 05:38 |
| lamont | those are 4-octet letters | 05:38 |
| elmo | lamont: not really, sorry | 05:38 |
| lamont | elmo: np. I think that means you don't have to disable syncing of the apps, though. | 05:38 |
| lamont | since it's really just 'don't build any of these until all the libraries are in the archive', right? | 05:38 |
| doko | lamont: yes, or else you have to check for build-deps on C++ libs, which I didn't do | 05:39 |
| lamont | right | 05:40 |
| fabbione | 12 -rw-r--r-- 1 sparcbuildd sparcbuildd 10235 May 17 17:39 buildd.conf | 05:40 |
| lamont | doko: is the app list golden at this point? | 05:40 |
| fabbione | meh! | 05:40 |
| fabbione | it got huge :) | 05:40 |
| lamont | fabbione: it's an 8500+ byte _line_ | 05:40 |
| fabbione | lamont: yes i know :) | 05:40 |
| doko | lamont: no. wait, I'll update the list before I go ... | 05:41 |
| === lamont goes on a paste frenzy | ||
| lamont | doko: thanks. holler when it's golden | 05:41 |
| daniels | fabbione: the X order you listed is right | 05:41 |
| daniels | i'm just fixing xorg now so it's happier with Xrender.h | 05:41 |
| fabbione | daniels: ok. | 05:42 |
| lamont | l-r-m-2.6.10 is in the list?? wow | 05:42 |
| === lamont phears | ||
| fabbione | we can just skip that one forever | 05:42 |
| daniels | oh wait, there's a simpler option | 05:42 |
| daniels | i'm a fucking moron | 05:42 |
| daniels | fabbione: stick with the xorg you have | 05:42 |
| lamont | nah - we're breaking hoary-* builds right now... | 05:42 |
| daniels | i'll do the render/xrender transition | 05:42 |
| lamont | fabbione: you gonna add that to the quotes file??? :-) | 05:43 |
| fabbione | lamont: your pleasure :) | 05:43 |
| lamont | nah, you can have it | 05:43 |
| fabbione | i need more coffee :) | 05:43 |
| fabbione | ok i will :) | 05:44 |
| daniels | fabbione: ok, new render/xrender versions with changes in the usual place | 05:54 |
| daniels | i'm putting a new xorg up now, with explicit versioned build-deps on the new versions | 05:55 |
| fabbione | daniels: let me test first | 05:55 |
| daniels | sure | 05:56 |
| daniels | nothing uploaded anywhere but p.u.c/~daniels/newx/ and chinstrap:~daniels | 05:56 |
| fabbione | daniels: do you want to bump xrender Build-Dep on render-dev version too? | 05:57 |
| fabbione | or there is no need to? | 05:58 |
| daniels | nah, no need | 06:00 |
| fabbione | i confirm with the new *render* xorg is building again... | 06:00 |
| daniels | autoconf is smart enough to deal with it; imake is not | 06:00 |
| fabbione | ok | 06:00 |
| fabbione | restart xorg build from scratch | 06:00 |
| daniels | fabbione: i believe the phrase is 'the truth, the whole truth, and nothing but the truth' | 06:01 |
| fabbione | daniels: i remember a similar one, but apparently ameringlish says that one | 06:02 |
| daniels | weird | 06:03 |
| === fabbione hugs ccache | ||
| lamont | fabbione: actually, daniels is more correct in this case. :-( | 06:06 |
| fabbione | lamont: he must be because of his experience with the court :) | 06:06 |
| doko | fabbione, lamont, elmo: updated cxxlibs.txt and cxxapps.txt in my home on chinstrap | 06:06 |
| daniels | fabbione: haha | 06:07 |
| daniels | yeah, so much experience with american courts :P | 06:07 |
| daniels | i'm going back to bed now, tired as hell | 06:07 |
| fabbione | daniels: are render and xrender signed too? | 06:08 |
| daniels | i'll be back at 2300 UTC (7h from now) | 06:08 |
| daniels | fabbione: yep | 06:08 |
| fabbione | ok | 06:08 |
| daniels | i'm phoneable before then if xorg is ftbfs though | 06:08 |
| fabbione | daniels: ok. | 06:09 |
| fabbione | doko: ok thanks | 06:09 |
| daniels | fabbione: i'm just updating x11proto-core and xtrans to fix their pre-depends now | 06:09 |
| fabbione | do they need pre-depends too????? | 06:09 |
| fabbione | i didn't see the problem here | 06:09 |
| elmo | doko: should I freeze both? | 06:10 |
| daniels | they had a Pre-Depends on x-common (>= 1.0) | 06:10 |
| fabbione | ah ok | 06:10 |
| daniels | which needs to be changed to 0.99 for the work-around-the-buildd thing | 06:10 |
| daniels | nothing major | 06:10 |
| fabbione | doko: the 3rd one is the source package, right? | 06:10 |
| daniels | fabbione: ok, x11proto-core and xtrans updated in the usual spots | 06:11 |
| daniels | night all | 06:11 |
| fabbione | daniels: night | 06:11 |
| doko | elmo: hmm, the library list for syncs only, or else we can't upload bug fixes | 06:12 |
| elmo | uh, ok | 06:14 |
| lamont | doko: don't all the lib packages have -ubuntu versions? | 06:17 |
| lamont | which would let the merge process do its thing... | 06:17 |
| fabbione | yup | 06:17 |
| === lamont disappears into the other workspace to update buildd.conf on 12 machines | ||
| doko | lamont: yes, but not yet | 06:18 |
| fabbione | doko: he is blocking the CxxApps | 06:19 |
| elmo | ok, cxxlibs + cxxapps blocked from syncs | 06:21 |
| elmo | cxxlibs blocked from uploads except by doko's key | 06:21 |
| doko | heh, then I can upload gcc-defaults and buid-essential? | 06:21 |
| fabbione | doko: you should be greenlight for it now | 06:21 |
| fabbione | once they are built, we need to update all the chroots on the buildds | 06:21 |
| doko | elmo: maybe add amu's and jbailey's key? | 06:22 |
| fabbione | elmo: mine and daniels for X please | 06:22 |
| fabbione | we have c++ libs in X | 06:22 |
| elmo | meh | 06:22 |
| fabbione | elmo: it's probably easier to just bless them manually? | 06:23 |
| doko | elmo: ready to start? | 06:25 |
| === fabbione makes the drums sound for doko | ||
| fabbione | doko: probably tell to #u-d :) | 06:25 |
| elmo | added amu, jbailey, fabbione and daniels' keys | 06:27 |
| fabbione | elmo: thanks | 06:27 |
| === lamont tries to decide if he wants to be able to fix bugs, or if he'll be busy dealing with other things | ||
| fabbione | lamont: i did test X build order locally... | 06:32 |
| fabbione | better you keep free is something goes wrong on the buildd | 06:32 |
| fabbione | s/is/if | 06:32 |
| fabbione | AYE | 06:40 |
| fabbione | another Xorg FTBFS | 06:40 |
| fabbione | there... fixed... | 06:48 |
| fabbione | hopefully this is the last one | 06:48 |
| fabbione | elmo: can you please let nvu from sparc in? | 06:58 |
| fabbione | i am sure it has been built sanely | 06:58 |
| fabbione | it was building way before we started the transition | 06:59 |
| elmo | eh? | 06:59 |
| elmo | it's not in the list to exclude? | 06:59 |
| elmo | hum | 07:00 |
| fabbione | yes it is | 07:00 |
| fabbione | that's why i was asking :) | 07:00 |
| elmo | err, rather it's in the apps list | 07:00 |
| elmo | and doko told me only to exclude source stuff | 07:00 |
| fabbione | so it should go in.. | 07:00 |
| fabbione | well i am sure it is built properly | 07:00 |
| fabbione | danke | 07:01 |
| === fabbione kicks Xorg in the code | ||
| fabbione | humpf | 07:07 |
| fabbione | this will need an Imake patch to get fixed properly | 07:07 |
| === lamont told the neighbor that he'd come look at his install issue in about 45 min... so I need to wander off for a few soon.. | ||
| === lamont looks to see if he can update chroots yet | ||
| fabbione | lamont: i don't see build-essential around yet | 07:18 |
| fabbione | yay there it is another package | 07:18 |
| lamont | fabbione: I'm busy being confused... | 07:21 |
| fabbione | lamont: ehehe | 07:22 |
| lamont | we should have gcc-defaults and build-essential there for all architectures after the :33 pulse, I think | 07:22 |
| fabbione | yes i think so | 07:22 |
| fabbione | build-essential missed the :33 | 07:23 |
| fabbione | gcc-defaults source is there.... | 07:23 |
| fabbione | or it seems like | 07:23 |
| fabbione | ok i got xc to compile.. | 07:24 |
| fabbione | let's see xc-debug | 07:24 |
| fabbione | bah crap | 07:25 |
| fabbione | it does a make clean | 07:25 |
| === lamont runs off for about 8 minutes | ||
| fabbione | i will need to get some food soon | 07:26 |
| fabbione | wow.. this was easy :) | 07:35 |
| fabbione | lamont: the GCC34_MMX macro is simply broken | 07:36 |
| fabbione | it doesn't know about gcc-4 | 07:36 |
| fabbione | so killing the -D works | 07:37 |
| fabbione | even if it is not the cleanest solution | 07:37 |
| lamont | build-essential_11.0ubuntu1 is installed in all the data center chroots | 07:40 |
| lamont | we are _GO_ for library infusion | 07:42 |
| fabbione | lamont: ok thanks | 07:42 |
| lamont | at least from my perspective | 07:42 |
| fabbione | but we are not ready for the libs yet | 07:42 |
| fabbione | we need xorg first | 07:42 |
| lamont | ah, well then. get on it. | 07:42 |
| fabbione | i am almost done i think | 07:43 |
| fabbione | there are a couple of things that still don't match | 07:43 |
| fabbione | ok if i got everything right, this is the last xorg build (at least on i386) | 07:46 |
| fabbione | i just need to revisit a second the upload order | 07:47 |
| === lamont back in a bit | ||
| fabbione | elmo: i am almost ready to start uploading X packages | 08:22 |
| fabbione | some of them will need NEW love | 08:22 |
| elmo | ok | 08:24 |
| fabbione | this is going to be the upload order: | 08:39 |
| fabbione | in parallel: | 08:39 |
| fabbione | render_0.9-0ubuntu2_source.changes | 08:39 |
| fabbione | x-common_0.99_source.changes | 08:39 |
| fabbione | x11proto-core_6.8.99.7-1_source.changes | 08:39 |
| fabbione | xtrans_0.2+cvs.20050513-1_source.changes | 08:39 |
| fabbione | (3 of them needs NEW) | 08:39 |
| fabbione | xrender_0.9.0-0ubuntu5_source.changes | 08:39 |
| fabbione | (build-dep on render-dev) | 08:40 |
| fabbione | and again in parallel: | 08:40 |
| fabbione | xft_2.1.7-1ubuntu1_source.changes | 08:40 |
| fabbione | xcursor_1.1.3-1ubuntu1_source.changes | 08:40 |
| fabbione | (both build-dep on xrender-dev) | 08:40 |
| fabbione | and as last xorg | 08:40 |
| fabbione | this should do | 08:40 |
| fabbione | yeah x is installing | 08:59 |
| fabbione | dh_install --sourcedir=debian/tmp | 09:02 |
| fabbione | COME ON | 09:02 |
| fabbione | install bitch! | 09:02 |
| fabbione | lamont: are you back? | 09:11 |
| fabbione | elmo: first 4 packages on the way | 09:11 |
| === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
| fabbione | elmo: katie hates me :) | 09:23 |
| fabbione | or hates daniels | 09:23 |
| elmo | REJECT | 09:24 |
| elmo | Rejected: Unknown distribution `unstable'. | 09:24 |
| elmo | that you mean? | 09:24 |
| fabbione | i don't get the rejects.. daniels did the packages | 09:24 |
| fabbione | what package was that? | 09:24 |
| elmo | render_0.9-0ubuntu2_source.changes | 09:24 |
| elmo | x-common | 0.99 | source | 9 minutes old | 09:24 |
| elmo | x11proto-core | 6.8.99.7-1 | source | 9 minutes old | 09:24 |
| elmo | xtrans | 0.2+cvs.20050513-1 | source | 9 minutes old | 09:25 |
| elmo | ^-- all got to NEW | 09:25 |
| fabbione | ok.. thanks | 09:25 |
| fabbione | i am repreparing render | 09:26 |
| elmo | eh, what's the point of x-common? | 09:26 |
| fabbione | elmo: to make the X11 symlink disappear | 09:27 |
| fabbione | as far as i understood at least | 09:27 |
| elmo | a whole package for it? | 09:28 |
| fabbione | elmo: daniels is killing X11R6 and soing modular | 09:28 |
| fabbione | x-common will get more stuff later | 09:28 |
| fabbione | new render is up | 09:28 |
| fabbione | you can let them in at any time | 09:29 |
| fabbione | s/soing/going | 09:30 |
| elmo | I assume these packages have to all be in main? | 09:31 |
| fabbione | elmo: yes | 09:31 |
| elmo | ok, processed | 09:32 |
| fabbione | great | 09:32 |
| fabbione | did they make the :33 daily? | 09:33 |
| fabbione | sorry for pushing/asking but it's like 15 hours that i am here :) | 09:33 |
| fabbione | if i can save 30 minutes, that would be really nice :) | 09:33 |
| elmo | hmm | 09:34 |
| lamont | back | 09:34 |
| fabbione | lamont: rocking.. the first 4 pkgs have been accepted | 09:34 |
| elmo | hsssssssssssssssssssssssst | 09:35 |
| === fabbione watches elmo melting down | ||
| elmo | hppa broke cron.daily | 09:36 |
| elmo | once the current one finishes, I'll rerun it | 09:36 |
| fabbione | ah ok | 09:36 |
| fabbione | thanks | 09:36 |
| === lamont boggles | ||
| fabbione | hmm it looks like the list of packages for no_auto_build is a bit too long | 09:42 |
| elmo | ? | 09:42 |
| fabbione | ah crap no | 09:42 |
| fabbione | never mind | 09:42 |
| elmo | btw, cron.daily done | 09:49 |
| fabbione | rocking | 09:49 |
| elmo | is there any reason you can't just upload all this stuff and let the buildds figure it out based on the b-d's? | 09:49 |
| fabbione | elmo: because i am not too confident in these packages. they got very little testing | 09:50 |
| elmo | ok | 09:50 |
| fabbione | so i prefer to break only one piece at a time if the Build-Deps are wrong | 09:51 |
| elmo | these -dev packages have .c files in /usr/include?!?!?!?!?!!?!?!?! | 09:51 |
| fabbione | eh?????? | 09:51 |
| elmo | drwxr-xr-x root/root 0 2005-05-17 20:45:55 ./usr/include/X11/Xtrans/ | 09:51 |
| elmo | -rw-r--r-- root/root 3070 2005-05-17 20:45:55 ./usr/include/X11/Xtrans/transport.c | 09:51 |
| elmo | -rw-r--r-- root/root 31104 2005-05-17 20:45:55 ./usr/include/X11/Xtrans/Xtrans.c | 09:51 |
| fabbione | daniels !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! | 09:51 |
| fabbione | bah | 09:52 |
| fabbione | that can be fixed as soon as daniels wakes up | 09:52 |
| fabbione | brb | 09:52 |
| fabbione | i need 5 minutes break at least to see my wife | 09:52 |
| elmo | sure | 09:53 |
| fabbione | re | 10:01 |
| lamont | hrm... does ccache do gcj stuff too? | 10:05 |
| elmo | no | 10:05 |
| === lamont grumbles | ||
| jbailey | lamont: Shouldn't be alot of gcj stuff. | 10:05 |
| jbailey | lamont: It should be mostly arch: all with a bit in the postinst. | 10:06 |
| jbailey | lamont: That rare exception being the bits that use JNI | 10:06 |
| lamont | ah, that makes sense... the gcc-4.0 test build is -A | 10:06 |
| fabbione | ok it looks like all the 4 packages did build properly... | 10:11 |
| fabbione | xrender is on the way | 10:12 |
| fabbione | this one should still be safe.... | 10:15 |
| fabbione | actually also xft and xcursor should be safe... | 10:16 |
| fabbione | there.. up | 10:17 |
| fabbione | as soon as these 3 packages are available we should be able to upload Xorg | 10:17 |
| lamont | fabbione: which packages am I babysitting? | 10:19 |
| fabbione | lamont: xrender_0.9.0-0ubuntu5_source.changes (chinstrap) | 10:19 |
| fabbione | xft_2.1.7-1ubuntu1_source.changes (me) | 10:19 |
| fabbione | xcursor_1.1.3-1ubuntu1_source.changes (me) | 10:19 |
| fabbione | they should hit the buildd at the next :33 daily | 10:20 |
| fabbione | the other 4 already builded fine | 10:20 |
| fabbione | all ACCEPTED | 10:21 |
| fabbione | elmo: if you want to speed up things, you could run .daily again | 10:21 |
| elmo | running | 10:22 |
| fabbione | rocking | 10:22 |
| elmo | done | 10:28 |
| lamont | May 17 21:29:28 buildd: breezy: total 3 packages to build. | 10:29 |
| fabbione | lamont: i guess it's them :) | 10:30 |
| fabbione | lamont: are they building or did they destroy the chroots? | 10:37 |
| lamont | pretty much uploading, without digging around too much | 10:42 |
| fabbione | ok | 10:42 |
| lamont | built on all 4 | 10:43 |
| elmo | I'm running cron.daily again | 10:43 |
| elmo | AFAICS everything's built | 10:43 |
| fabbione | ok | 10:43 |
| fabbione | uploading xorg now :) | 10:44 |
| fabbione | if this one break we are fucked | 10:44 |
| lamont | fabbione: no. we'll be annoyed. | 10:44 |
| lamont | there's a subtle difference. | 10:44 |
| fabbione | lamont: eheheh | 10:45 |
| fabbione | well the point is that i am tired and i am not 200% sure i can manage to debug X | 10:45 |
| fabbione | so.. somehow somebody is fucked until daniels wakes up :) | 10:45 |
| fabbione | so you and elmo can be annoyed | 10:45 |
| fabbione | i will be fucked for the next 2/3 hours | 10:45 |
| fabbione | :) | 10:46 |
| fabbione | xorg is up | 10:46 |
| fabbione | if this one goes ok, we can push all the libs from doko | 10:48 |
| fabbione | xorg_6.8.2-11_source.changes ACCEPTED | 10:51 |
| lamont | make[4] : *** No rule to make target `gnu/java/net/protocol/https/Handler.java', needed by | 10:52 |
| lamont | +`gnu/java/net/protocol/https/Handler.class'. Stop. | 10:52 |
| lamont | make[4] : *** Waiting for unfinished jobs.... | 10:52 |
| lamont | make[4] : Leaving directory `/build/buildd/gcc-4.0-4.0.0/build/x86_64-linux/libjava' | 10:52 |
| lamont | oh doko!!!! | 10:52 |
| lamont | gcc-4.0_4.0.0-7ubuntu3 times 4 | 10:52 |
| fabbione | lamont: yeah he told me that ubuntu3 was FTBFS | 10:54 |
| lamont | do we have -4? | 10:54 |
| fabbione | not yet | 10:55 |
| fabbione | he told me one minute before leaving | 10:55 |
| fabbione | lamont: keep a close eye on xorg | 11:00 |
| fabbione | it will pull in all the crap in one go | 11:00 |
| elmo | xorg's building | 11:23 |
| fabbione | elmo: cool.. on all 4 arches? | 11:23 |
| elmo | (sorry about the delay, I was (re)building a 2nd archive.ubuntu.com, and it was slaughtering jackass' performance | 11:23 |
| elmo | I assume so | 11:23 |
| fabbione | no problem about the delay.. i am almost dead anyway :) | 11:24 |
| fabbione | lamont: ? | 11:24 |
| elmo | yep, all 4 | 11:24 |
| fabbione | cool | 11:24 |
| fabbione | we need to wait Xorg, upload a clean x-common and the libs from doko | 11:24 |
| elmo | clean x-common? | 11:25 |
| fabbione | elmo: yes. 0.99 had a Dependency hack to build xorg -11 | 11:26 |
| fabbione | once .11 is built, the hack can be removed | 11:27 |
| === elmo puts his hands in his ears and sings LALALALA | ||
| fabbione | ahhaha | 11:27 |
| fabbione | elmo: either 2 x-common or 2 xorg uploads | 11:27 |
| fabbione | considering the size and the arch: all we opted for x-common | 11:27 |
| fabbione | we did agree that it was dirty | 11:28 |
| lamont | elmo: I told them it was evil at least... | 11:28 |
| fabbione | but definetely faster :) | 11:28 |
| lamont | we need to make sure that we have a URL for the morgue-ed source for the 0.99 x-common package to put on the evilness-you-must-endure-when-porting page | 11:28 |
| fabbione | lamont: there will be no need | 11:28 |
| lamont | oh? | 11:29 |
| fabbione | the reason is that xorg-common is arch: all | 11:29 |
| fabbione | so once it is built, it will be there and cooperating properly with x-common 1.0 | 11:29 |
| lamont | and how will xorg build with the old new x-common installed on a debian chroot? | 11:29 |
| fabbione | at the end of the transition xorg will not exists anymore | 11:30 |
| fabbione | there will tons of little tiny source packages | 11:30 |
| lamont | ah, ok | 11:30 |
| fabbione | unfortunatly we did hit a bad timing for X | 11:30 |
| fabbione | because daniels was splitting xorg to the modular tree, killing X11R6 at the same time that we need to do the C++ transition | 11:31 |
| lamont | right | 11:31 |
| fabbione | so look at this sequence of uploads as something you won't need after | 11:31 |
| lamont | right. But I will need it this week, when I finally get hppa started on building breezy. sigh | 11:32 |
| fabbione | lamont: once i386 build the arch: all, you will have no problems | 11:32 |
| lamont | are we planning to bump all the cxxapps, or just let time and transition deal with that? | 11:33 |
| fabbione | no idea about the apps | 11:33 |
| fabbione | we will need to ask doko | 11:33 |
| fabbione | he should be around soon enough | 11:33 |
| lamont | fabbione: once i386 builds the second x-common arch:all package, then it can take the place of the first? huh??? | 11:33 |
| fabbione | lamont: yes. | 11:34 |
| lamont | even though I have a hoary xorg? | 11:34 |
| fabbione | the problem is the interaction between xorg-common -10 postinst and x-common | 11:34 |
| lamont | which seemed to me to be the whole reason for the hacked upload.... | 11:34 |
| lamont | fabbione: that's hppa's current state... has xorg -10 | 11:34 |
| fabbione | lamont: you still get _all.deb from i386 | 11:35 |
| fabbione | that will be -11 | 11:35 |
| lamont | ok | 11:35 |
| fabbione | if you are building arch: all locally.. well.. yeah you will need to go trough the same harassment | 11:35 |
| lamont | it does mean that I need to have my mirror current before I get to xorg though.... | 11:35 |
| fabbione | yeps | 11:35 |
| lamont | only arch-all building locally is either testing, debian uploads, or glibc hacking. damn locales | 11:36 |
| fabbione | eh | 11:36 |
| fabbione | ehehe | 11:36 |
| === lamont watches the glibc build go through Testing IBM500.... | ||
| === fabbione wonders how long is going to take | ||
| lamont | fabbione: until about :20 ish | 11:56 |
| lamont | based on history, that is | 11:56 |
| fabbione | ok | 11:58 |
| === fabbione gives back a bunch of packages in Dep-Wait for gcc-4 & co | ||
| fabbione | i need to find a way to automatically set the packages in Dep-Wait | 12:00 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!