[00:15] ScottK: I've made the description changes to both the control file and manpage, if you are inclined to be an advocate... http://revu.ubuntuwire.com/details.py?upid=3405 [00:15] I may have time to look later. I'm busy with $WORK at the moment. [00:15] ok, thank you === ember_ is now known as ember === asac_ is now known as asac === emma_ is now known as emma [01:31] Heya folks [01:35] Howdie bddebian [01:35] Heya RAOF [01:37] Any of you know if there is a nice and easy way to really figure out DEPENDSON in Makefile.am without digging through all the dependencies by hand? [01:40] bddebian: Not from Makefile.am, but configure.ac will ideally have a whole bunch of PKG_CHECK_MODULES [01:41] bddebian: They'll use pkg-config to check for the presence of various stuff, and apt-file can inform you of the package providing the relevant .pc file. [01:41] kirkland: i replied on revu (update-motd package) [01:42] kirkland: other than that suggeston i'm quite happy with what i've already reviewed [01:42] any revu admins around? [01:42] RAOF: I don't mean external package dependencies. I mean actual object files within the same package. [01:45] Oh. That did sound like a fairly basic question ;). [01:45] Do you mean minimal dependencies or make-dependencies? [01:46] It should be possible to extract the latter without too much trouble, but extracting the former will likely require source-greppage. [01:47] automake will link all the object files from foo_SOURCES together, right? [02:19] RAOF: Gah, sorry, I didn't realize you were still talking to me [02:20] Makefile.am just has this endless list of: DEPENDSON = ../GameScenario.o ../army.o ../GameMap.o ../templelist.o .... [02:20] new azureus crack, away we upload... [02:20] once vmware fusion boots back up [02:20] stupid wrong button [02:21] problem #4 with macs: Too few buttons = buttons are too big = chances of hitting the wrong one increase. === superm1|away is now known as superm1 [03:11] persia, its amazing how one can be so dead to the world that connecting to IRC is a burden [03:11] NCommander: i was looking for you [03:11] what did I break now? [03:11] NCommander: i've just been added to the MOTU team, but in REVU i still apear as a normal contributor [03:11] heh [03:12] default answer :P [03:12] Its not an automatic upgrade [03:12] We need to manually set you in REVU to be a reviewer [03:13] Altering nxvl to level reviewer [03:13] Done [03:13] oh [03:14] i thought it was automagiuc [03:14] Thank you! [03:14] That's one of "it would be nice" things [03:14] nxvl: It's only automagic after you ask NCommander. [03:14] Speaking of automagic ... [03:15] NCommander: How's your relibtoolization skillz? [03:15] ScottK, you summone RoAkSoAx? [03:15] Uh [03:15] Depends on how much pain you want to force on me [03:15] (I helped fix libtool on OS X when it first came out, take a guess) [03:16] So your inured to the pain by now then I'll assume. [03:16] what who when where? [03:16] ScottK: :D [03:16] RoAkSoAx, sorry, you walked into the punchline of a bad pun. [03:16] NCommander, haha ok np =) [03:16] RoAkSoAx: i know what ScottK is talking about, better run! [03:16] ScottK, what do you need fixed [03:17] nxvl, lol [03:17] nxvl, your an MOTU, you should be teaching me these things ;-) [03:17] NCommander: ok, then RUN! [03:17] NCommander: Would you please have a look at Bug 254681. [03:17] * NCommander runs towards the danger [03:17] Launchpad bug 254681 in courier "please merge courier 0.60.0-1 from debian sid" [Undecided,Incomplete] https://launchpad.net/bugs/254681 [03:17] heh [03:17] i knew this bug was about to show up [03:17] * nxvl HUGS ScottK [03:18] There's a debdiff at the end that is almost right +/- libtool something. [03:18] Ok [03:18] If either one of you will sponsor the end result [03:18] I'll fix it [03:18] Thanks. [03:18] i'm really interested in the result of this package [03:18] i gived up last week [03:19] BTW, after today, I'm going to have somewhat limited Ubuntu machine access until September [03:19] i needed to step out before i will start hating libtool more that i do now [03:19] (the laptop getting a Debian kFreeBSD installation) [03:19] NCommander: kvm [03:19] I don't think kernel virtual machines support freebsd ;-) [03:19] emu? [03:19] qemu [03:19] NCommander: I'm a Debian Maintainer and I don't run it at all. I just do stuff in chroots. [03:20] You can't run a LInux based distro on a freebsd host ;-) [03:20] i use chroot's + kvm's [03:20] You're right. I can't, but I bet you can. [03:20] You have amazing faith in me [03:20] NCommander: september is only a couple of weeks away [03:20] (I probably could force it to run under Linux emulation) [03:20] Well [03:20] I'll probably just bootstrap Ubuntu kfreebsd, and call it Freentu :-P [03:21] * NCommander needs a hobby or something [03:21] shun the heretic [03:21] NCommander: ScottK is amazing giving you emotional support and self-confidence [03:21] * nxvl HUGS ScottK for that [03:21] Argh [03:21] Anyone got a quick link to the .dsc? [03:21] NCommander: At one point I thought it might be something related to parallel building, so there's one thing that needs to come out of that debdiff before you play with it. [03:21] grab-merge is cool [03:21] :D [03:21] what? [03:22] What package is that form? [03:22] NCommander: http://merges.ubuntu.com/ [03:22] Oh [03:22] NCommander: http://merges.ubuntu.com/grab-merge.sh [03:22] That package grabber [03:22] yep [03:22] we love MoM [03:22] I thought I had that script in my bin [03:23] * NCommander looks at the hurd source code on his machine [03:23] Ok, its downloading [03:23] Slowly [03:23] you are sick, aren't you? [03:23] i tryed to boot hurd once [03:23] * ajmitch quietly hides his checkout of the hurd source [03:23] tried* [03:23] * ajmitch ran it on a computer for awhile [03:24] i runned quick screaming lout [03:24] * ajmitch was able to compile packages, and had a webserver on there [03:25] NCommander: I added a thing in rules to export DEB_BUILD_OPTIONS 'parallel=1' when I was testing a theory. I forget to take that out before I attached the debdiff. [03:25] oh damn i thing i'm mixing english with german again [03:25] * NCommander has CVS commit access to the hurd repo [03:25] I think I win [03:25] NCommander: don't even look at my debdiff, they are horrible [03:25] nxvl, I thought you were an MOTU ... [03:25] ;-) [03:26] NCommander: yes i am, and part of being motu is knowing your limitations and steping out of what you can't do [03:26] Hrm [03:26] ScottK: doesn't it? [03:26] Interesting concept, I should introduce it to Debian [03:26] nxvl: Absolutely. [03:27] \o/ [03:27] NCommander: I'm core-dev and I'm looking for help. [03:27] we are not packaging or code gods [03:27] I thought ScottK was [03:27] just normal (well, not so normal) people contributing on what we can [03:27] Normal is relative [03:28] NCommander: yep, that's the why of the "(well, not so normal)" part [03:28] Yeah. I've met nxvl in person. [03:28] ;-) [03:28] heh [03:28] ScottK, is that a good or bad thing? [03:28] ScottK: lucky you [03:29] ajmitch: where are you based? [03:29] and btw, i would say he so lucky about that [03:29] <- evil nerd [03:29] nxvl: ends of the earth [03:29] It was a good thing, but it'd be better if he'd been able to fix courier. [03:29] ajmitch: asia? [03:29] nxvl: New Zealand [03:29] Further south. [03:30] heh [03:30] yes, the end of the earth [03:30] ScottK, how'd you manage to downgrade config.status/config.guess? [03:30] actually i want to go to New Zeland some day [03:30] and Australia [03:30] NCommander: we made A LOT of tries [03:30] NCommander: I have no idea. [03:31] THe patch doesn't apply cleanly to the base tarball [03:31] I do note that we have the libtool from Experimental in Intrepid, so I choose to blame it. [03:31] * ajmitch has been to australia a few times [03:31] NCommander: It's a diff to the latest Debian revision. [03:31] Ah [03:31] ANd now it does :-P [03:31] So is this a FTBFS error, or something else [03:32] FTBFS on maildrop. [03:32] And I should wipe out the parallel=1? [03:32] Yes. [03:32] And the comment line above it and the related debian/changelog entry. [03:32] and who's name going on the changelog? [03:32] Damn [03:32] Yours if you fix it. [03:32] thank god, i can't remember what the error was! [03:32] That's one long changelog [03:33] Welcome to Courier. [03:33] and to bleeding edge development [03:33] :D [03:33] Note that the Debina maintainer's package is still standards version 3.6 something. [03:33] and using older libtool [03:33] BUilding [03:33] er [03:33] really older one [03:33] And? Standards-Version isn't so important [03:33] install build-deps [03:34] Policy hasn't changed markedly from 3.6.x to 3.8 [03:34] StevenK: Agreed, but in this case it's an indication of how long it [03:34] it's been since he seriously thought about the package. [03:34] I recently fixed a package that was still 2.x series in its stanards versoins [03:34] I recently ran the entire Debian archive's GPG signatures [03:34] holly! [03:34] and found a few invalid ones [03:34] (and then promptly spammed the FTP admin with removal requests) [03:35] why DD's enjoy rebuilding the archive [03:35] No, not a rebuild [03:35] I was actually checking every dsc to see if it was a valid signature [03:35] * nxvl quivers [03:35] against the keyring [03:35] Part of importing the archive using dak [03:35] Probably the fastest turn around time on a removal request ever [03:36] (Geoneff went WTF when he was pointed to the packages with invalid signatures) [03:36] NCommander: I have a bug that was filed as an MIR, and turned into a removal request [03:36] er, Ganneff [03:36] rofl [03:36] I've never seena removal fly durign a freeze ;-) [03:37] The running theory is that there used to be a bug in the very old dinstall (before dak) scripts which corrupted the signature [03:37] or something [03:37] I've never seen GPG error out with unknown public key algrothmin [03:37] escuses [03:37] (and I installed IDEA to make sure it wasn't that) [03:37] I have. I've tried to convince pitti to remove something, NEW it, and then get it removed again. [03:37] o_o; [03:37] how'd that fly? [03:38] It didn't, but it made him laugh, which was the idea. [03:38] excuses* [03:38] * NCommander has that horror feeling you'd do that to some archive admin to make them question their grip on reality [03:38] Shoot. I can quit now. According to me email, I'm a lucky winner. [03:38] me/my [03:38] ScottK, can you patebin the actual FTBFS? (I'm waiting for this to build) [03:38] ScottK: on what? [03:39] oh geeze, this configure script recursively calls itself ... [03:39] *winces* [03:39] I don't have it still. [03:39] * NCommander has never seen a changelog with three different people editing the SAME entry) [03:39] NCommander: Welcome to team maintenance. [03:40] It's the Ubuntu way. [03:40] ep [03:40] yep [03:40] i have seen a lot of them [03:40] i thing the record was 4 people [03:40] I'm suprised their isn't a lintian warning [03:40] VCS ftw! [03:40] W: debian-developers-being-social [03:40] All done with no fancy VCS either. [03:40] nxvl: Which one. [03:40] ScottK: that's the scary part [03:41] There is bazaar [03:41] ScottK: don't remember, but i have the image in my head of 4 people in a changelog entry [03:41] nxvl: Find me a VCS that can handle 20.000 source packages and we'll talk. [03:41] heh [03:41] :D [03:41] git \o/ [03:41] NCommander: Yes, if you want to wait for the heat death of the Universe while you branch it. [03:41] Oh, you mean the entire repo? [03:41] (the hardware is a different thing [03:41] BitKeeper :-P [03:41] Yeah. [03:41] Unless you meant a free based one [03:42] We already use LP, so that's clearly not a criteria. [03:42] I dunno, git doesn't massive repos well (as in size, any given checkout of Linux is just 50-60MB, but I can't see it scalling all the way up to 200-300GB) [03:42] ^do [03:42] ScottK: canonical will open LP this year (well, at least that's what has been told) [03:43] Fundamentally in Universe, until you can stuff the whole thing in one VCS and get agreement on it, the source package is the one and only canonical source for a package and VCS just isn't interesting. [03:43] They've been saying that for years [03:43] nxvl: Yep. It'll happen when it happens. [03:43] NCommander: But they never put a timeline on it before. [03:43] * NCommander is reminded SourceForge.net was on a timeline to be opened sourced [03:44] configure. is. still. running. [03:44] Freeing the source is important, but I don't think it'll do much to change the aspects of the project that I find most problematic. [03:44] *wince* [03:44] ScottK, try and find (or write) a greasemonkey script to make LP usable ;-) [03:44] I have great hopes for \sh and Leonov. [03:45] \sh works for Canonical? [03:45] No. [03:45] what is he working on then? [03:45] holly crap, one more windows specific software for the university [03:45] He's writing a python-qt4 Launchpad client using python-launchpad [03:45] they only make me hate microsft more that i do now [03:45] nxvl, see if it works in WINE [03:46] ScottK, wow .... that's awesome [03:46] nxvl, i've been through that too :P [03:46] It reminds me of GNATS ... [03:46] NCommander: i hate wine too [03:46] :D [03:46] nxvl, red or white? [03:46] heh [03:47] class i over [03:47] read you later! [03:47] * ScottK hands NCommander https://launchpad.net/leonov [03:47] I think bash.org went up to that 404 in the sky ... [03:48] FINALLY [03:48] courier starts compiling [03:48] What a beast [03:49] Configure should die [03:49] It seemed to me you'd find it familiar and welcoming. [03:50] configure isn't so bad if you write the scripts right [03:50] i.e. most GNU projects and Hurd [03:50] Then you get some beautiful abuses [03:50] GCC for instance [03:50] Right, but you seem to revel in painful code. [03:50] ITs the only script I've ever seen that can actually compile binutils, glibc, newlib, hurd, and gnumach all in a single pass [03:50] O_o; [03:51] ^autoconf [03:51] BTW, perforce could handle the entire ubuntu packages repo [03:51] its probably one of the only VCSs that could scale to the size required [03:52] At $900 per seat. [03:52] Open source development on perforce is free [03:52] That's why both OpenWatcom and FreeBSD use it [03:52] Hmmmm. [03:52] Interesting. [03:52] The only time I've run into was at a customer's that was not doing open source development. [03:53] Oh for the love of [03:53] WTF is wrong with this libtool script [03:53] *twich [03:53] git's inability to handle directory history tracking bugs me [03:53] (and the way it handles renames) [03:56] it looks like this libtool is broken compiling CXX objects [03:57] ScottK, well, I kinda figured out your issue [03:57] did you touch this libtool script? [03:58] No. The straight Debian package fails the same way on Intrepid. [03:59] for some reason [03:59] Half the script is corrupted [03:59] I copied a known working libtool [03:59] and it works [03:59] How big of a hack do you want this to be? [04:00] I'll settle for working and no one will yell at me if I upload it. Maintainable would be nice. Bonus points for a nice patch we can send to Debian. [04:00] Does the Debian one build? [04:00] (I'd be extremely suprised) [04:01] Not on Intrepid. They both build on Sid. Neither build on Experimental. [04:01] Ok [04:01] Note that Intrepid and Experimental have the same libtool. [04:01] configure generating a bum libtool [04:01] None of the configuration variables are being set properly [04:03] OK. So how do we fix it. [04:03] Well [04:03] Copy the working libtool from the base of the source directory into maildrop [04:03] Which gets maildrop building [04:03] But its ugly as hell [04:04] OK. What's the non-ugly answer? [04:04] I think its cached the old libtool [04:04] All the autocrap cache files are sitting here [04:05] OK. So maybe I just open a pristine tarball, drop the debian dir in there and see if it builds. [04:05] I think these cache files are from upstream [04:05] Oh. [04:05] Crap. [04:05] Yeah [04:05] Crap [04:05] I'm seeing if I can rebuild autotools [04:05] (in maildrop) [04:06] Want to see if clean source just builds? [04:07] Wiorking on it. [04:07] * NCommander looks at autofun's manual [04:07] There's probably an M4 macro missing [04:10] Build started, so we'll know in a few hours ... [04:10] build will fail. [04:12] ?? [04:12] Why does courier take so long to build? [04:13] It doesn't really take hours. It just feels like it. [04:14] GAH!!! [04:14] I'm checking now on a clean build [04:14] Hobbsee, ? [04:14] my battery has died, in my music player. [04:15] and i didn't bring the USB cable to charge it, or anything. [04:15] Ok [04:15] On a clean courier [04:15] Maildrop builds out of the box [04:15] so we can blame debian on this one [04:16] Excellent. [04:16] Ugh [04:16] Mark yourself off in debian/changelog for having cleaned it up and attach a debdiff to the bug. [04:16] He inlines all the patches [04:16] Ugh. [04:17] So we have to sort the real changes from the libtool cruft. [04:18] tasty. [04:20] how do some of these people become DDs [04:20] How do they pass T&S [04:20] Some of them did before there was T&S. [04:20] Some of them are no doubt the reason there is T&S. [04:22] ugh [04:22] *twichs* [04:22] I think all these config.sub files in the diff file are doing it [04:23] are the .po files susposed to be in the diff o_o; [04:24] Some of those are likely bogus and some aren't. [04:24] The timestamp is from a few versions ago [04:24] I just removed them all [04:24] If it builds [04:24] Fine [04:25] It's building [04:26] ScottK, I'm trying to figure out if I should hurt the Debian upstream or the courier upstream [04:26] The former for doing fun stuff like .sub/.guess when out of the box works fine, and the later for using autofun [04:29] So I removed that from the diff and I'm going through to make sure no other crap snuck in. [04:30] I'm currently at line 10,729 of the .diff. Not kidding. [04:30] Mine is down to 6000-ish [04:31] I removed everything expect the "seems to still be needed", and everything in debian [04:32] I'm trying to preserve the .po file changes so it's a little less simple. [04:32] I left those in [04:32] And mine was still down to 6000 [04:32] He included config.status/config.guess at least a dozen times [04:33] er wait [04:33] Sorry [04:33] Mine is 13166 [04:33] Miscounted [04:33] (had a stray / 2o ro something ;-) [04:37] bddebian: Still interested in discussing that dependency thing? That sounds like upstream doesn't actually know how to use automake :) [04:38] RAOF: I'm trying to help upstream. He inherited the build system from someone else [04:39] Ah, right. So what's actually happening is that the build is currently producing a pseudo-library in one directory and trying to use it in this one? [04:40] Apparently :) [04:40] hi everybody [04:40] ScottK, its past the point where it was FTBFSing I think [04:40] Probably what wants to happen is adding some foo_LIBADD=../thelibdirectory/foo.a, I think. [04:41] OK. [04:42] do you know where i should ask for a software in Universe ? [04:43] RAOF: That's one possibility but he doesn't want a static library that gets installed if possible I think. Something like a libtool convenience library or such? [04:46] ScottK, I spoke too soon [04:47] OK. What now? [04:48] bddebian: Yeah. The goat book (I think) has a section dedicated to "internal convenience libraries". [04:50] bddebian: http://sources.redhat.com/autobook/autobook/autobook_92.html#SEC92 [04:50] NCommander: I just finished mine and started it to build. [04:51] ScottK, well, if you want a super-sized hack to make it build, through "cp $srcidr/libtool $srcdir/maildrop/libtool" after configure [04:51] and it will work :-/ [04:51] RAOF: Thanks man [04:53] Maybe if this one doesn't work I'll try that. .diff.gz looks way cleaner in any case. [04:59] ScottK, I might have run down the configure.in issue with libtool [05:00] (I had to rearrange some of the rules) [05:00] er [05:00] nm [05:03] Nope. Mine didn't work either. [05:04] Its frustaitng [05:04] NCommander: you fixed courier? [05:04] I have a work around which works [05:04] Buts not a real fix [05:05] :( [05:05] NCommander: Are you up for fiddling around with it some more? [05:06] Clearly you're making progress. [05:06] I told you my "fix" [05:06]