/srv/irclogs.ubuntu.com/2004/10/02/#ubuntu-devel.txt

mdzdaniels: did you announce ubuntu19?12:06
danielsmdz: yes, to -users12:54
danielsgone now12:54
thomnpmccallum: wassup?01:12
npmccallumthom: seb128 is taking care of it, thanks01:13
=== parkerc_ [~parkerc@cs671066-129.houston.rr.com] has joined #ubuntu-devel
=== Kamion [~cjwatson@host81-153-126-219.range81-153.btcentralplus.com] has joined #ubuntu-devel
Kamionmdz: much quieter, thanks, didn't know we'd started this02:49
Kamionmdz: the d-i builds that are used are all in debian-cd/tools/boot/warty/; if you regard those scripts as configuration files, it's configurable02:50
mdzcould someone send an announcement to the ubuntu-devel list?02:50
mdzlamont: is linux-restricted-modules happily building?03:52
lamontnope.  missing build-depends: bzip203:52
lamontthat'd be 1.1-403:52
lamontbzip2 -c fireglcontrol.qt3.gcc3.3.4 >fireglcontrol.qt3.gcc3.3.4.bz203:53
lamont/bin/sh: line 1: bzip2: command not found03:53
lamontmake[1] : *** [fireglcontrol.qt3.gcc3.3.4]  Error 12703:53
lamontbut it got that far...03:53
lamontdid daniels get his ftbfs info?03:54
mdzdunno03:55
mdzI'll fix l-r-m now03:55
lamonthanks03:57
lamontthanks, even03:57
=== lamont wanders back over to his test machine to stare at postfix some more
lamontoh, so samba... do we just want to disable sendfile?03:57
mdzlamont: if in doubt, we should do what debian and upstream say is the right thing03:59
mdzlamont: new l-r-m uploaded; please keep an eye on it, people are waiting03:59
mdzKamion: still here?04:03
mdzKamion: I'm looking for bug #1300904:06
=== lamont grumbles at kamion, d-i, and postfix.
lamontKamion: could it possibly be that postfix's postinst gets run before the hostname gets set?04:07
lamontthat would explain it.04:07
mdzpostfix gets installed by debootstrap04:08
mdzwhere the hostname is set, as far as I know04:08
lamontwas just able to reproduce the bug (aliases not built), but purge/reinstall postfix and all is well.04:09
lamontsigh.\04:09
=== lamont wonders how painful it is to take an ubuntu archive and produce a CD image...
=== lamont finds a, um, clever solution to his postfix issues.
mdzlamont: l-r-m happy now?04:19
lamontyep04:20
lamontand uploade.04:20
lamontd04:20
lamontls lrm still i386 only?04:21
mdzlamont: yes, for now04:30
=== parkerc_ [~parkerc@cs671066-129.houston.rr.com] has joined #ubuntu-devel
mdznvidia, at least, comes in amd64 as well04:31
jdubahr04:44
jdubmorning04:44
mdzmorning04:52
mdzjdub: are you going to follow up to Subject: Common Desktop goals04:54
mdz?04:54
mdzjdub: (cross-posted to debian-desktop and ubuntu-devel)04:54
jdubyeah04:56
jdub(and userlinux discuss!_04:56
mdzhah, yeah05:09
jdub(that's where i first saw it)05:10
jdub(bleagh)05:10
=== lamont_r [~lamont@dialup-4.228.48.17.Dial1.Denver1.Level3.net] has joined #ubuntu-devel
=== lamont_r screams
mdzreinstall, back later05:18
=== lamont_r [~lamont@dialup-4.228.87.216.Dial1.Denver1.Level3.net] has joined #ubuntu-devel
fabbionemorning guys05:55
jdubyo fabbione 06:05
fabbionehey jdub 06:15
fabbionemdz: are you still around?06:15
=== lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-devel
fabbionedaniels: you around?06:36
bob2I think he's netless06:36
bob2for now06:36
fabbioneok06:40
danielsfabbione: hey dude06:51
danielsi just had to pop out for like 4 hours06:51
danielsso I started uploading my ubuntu20 stuff just before I left06:51
danielsit's on xfree86_4.3.0.dfsg.1-6ubuntu20.diff.gz now06:51
daniels(no .orig.tar.gz, just the binaries and diffs and stuff)06:51
fabbionedaniels: what did you add to ubuntu20?06:57
fabbionebecause i have the fix for the brasilian keyboard too06:58
fabbionecan you send me the usual interdiff too?06:59
fabbioneso i can resync the repo here?06:59
fabbioneand btw.. since 19 is not in the archive we shouldn't bump to 2007:00
fabbione(just cosmetic)07:00
fabbioneat the end nobody cares of a missing version07:00
danielsfabbione: all that's in 20 is my ati stuff (the backport, plus fwpll, plus the #911 resync)07:21
danielsa patch for s/synaptic/synaptics/ in dexconf07:21
danielsthat's it07:21
danielsif you drop me the brazilian keyboard diff, i'll merge that and throw you an interdiff07:22
fabbionejust throw me the interdiff and i will merge here07:23
danielser, xlibmesa-dri-dbg is uploading right now07:23
danielsyou could have to wait a while :)07:23
fabbioneno problem :-)07:23
danielsxlibmesa-dri-dbg_4.3.0.dfsg.1-6ubuntu20_i386.deb                                                 8% 4236KB   1.4KB/s - stalled -07:24
fabbionebut i think we should let users test ubuntu19 for a little while before pushing another update07:24
daniels(it's not really stalled, it's just that the uplink is saturated)07:24
danielsyeah07:24
danielsi totally agree07:24
danielsthat's why i'm not pimping this yet :) i just want to get this out there to some specific ati people07:24
jdubdaniels: you've done rad ati stuff?07:26
danielsjdub: r4xx support and tv out on ibooks07:27
jdubtv out on ibooks!07:27
jduboh, not with rage128 though, right?07:27
danielsjdub: um, that should already work07:28
jduboh?07:28
danielsyeah, with aty128fb07:28
fabbionedaniels: did you check for that RADEONPPL symbol missing on ppc?07:31
fabbionedaniels: otherwise it would be a really good idea to get the packages built on ppc too07:31
danielsfabbione: yeah, that was me being a tool07:33
danielsnote to self: 'paramaters' is not a word07:33
fabbioneeheh07:34
danielsit was a one-character fix07:34
danielsthat I just hand-hacked the diff for07:34
fabbioneyou suck :P07:38
jdubdaniels: whoa.07:38
fabbionemdz: ping08:09
fabbionemdz: deping ;)08:15
=== pitti [~martin@195.227.105.180] has joined #ubuntu-devel
=== mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel
mdzgack08:38
mdzthis ext3 filesystem has a strange layout, parted can't resize, yada yada08:39
mdzs/ext2/ext3/08:40
mdzthough the filesystem is ext3 and recognised as such08:40
=== jdthood [~jdthood@aglu.demon.nl] has joined #ubuntu-devel
mdzjdthood: hi08:43
jdthoodHi there08:44
=== jdthood guesses that mdz has questions about ALSA packaging
mdzI do?08:45
mdzactually I have questions about parted :-/08:46
mdzit can't resize my ext3 filesystem, which has brought by install to a screeching halt08:46
jdthoodmdz: I think you are mistaking me for someone else.  :)08:48
mdzyou are not Thomas Hood?08:49
jdthoodmdz: Yes, but I don't have any special knowledge about parted.08:49
mdznah, I didn't mean to imply it08:50
mdzbefore you /joined I was moaning about my parted problems08:50
jdthoodmdz: The reason I thought you were going to ask me questions about ALSA packaging is that (1) you uploaded alsa-utils 1.0.5-1ubuntu2 and (2) I was involved in putting out the new upstream alsa-utils 1.0.6-1.08:52
jdthoodCongratulations on ubuntu, by the way.08:54
mdzthanks; have you tried it?08:54
mdzI don't think we'll be pushing in a new upstream alsa-utils this close to release unless it fixes something important08:55
mdzthough we are considering backporting the kernel stuff08:55
jdthoodmdz: Re: trying ubuntu -- No, I haven't tried it but I will follow its progress.08:55
jdthoodmdz: Re: alsa-utils -- You have to make judgments with one eye on your schedule, I know.  It may be helpful if I tell you a bit about what's been going on.  I was recently recruited to the alsa packaging team.  I have made a few changes that I thought were important to make before sarge.  One of them was fixing alsaconf.  ubuntu has removed alsaconf entirely, so that won't affect you.08:57
mdzalsaconf scares me08:58
jdthoodmdz: What may interest you more is that I have improved the alsa initscript.  This ships in the alsa-base package, of course.  I am pretty sure that the new script is less buggy than the previous one.08:58
mdzjdthood: oh, what sort of bugs did you fix?08:59
jdthoodmdz: force-stop is now more useful, for one thing.08:59
mdzas you probably noticed, we've just made a couple of small changes to it08:59
mdzthe unmuting is probably of general interest, but I wanted to get it right (including multiple cards) before sending it upstream09:00
jdthoodmdz: Yes, unmuting is a headache.09:00
jdthoodmdz: An important change to alsaconf is that it writes an /etc/modprobe.d/sound file that contains an "install" line that makes the module loader call a hook script after loading the sound card driver.  This hook can be used to do "alsactl restore".09:01
jdthoodThus, alsactl restore happens after module load.09:02
jdthoodWe still do "alsactl restore" in the initscript start method too because integral sound card drivers are never loaded by the module loader.09:02
=== pitti [~martin@195.227.105.180] has joined #ubuntu-devel
jdthoodThe above doesn't work properly on a udev system, though.09:03
jdthoodOn a udev system the "alsactl restore" has to be done from a dev.d script.09:04
jdthoodWe haven't implemented that yet.  Supporting udev isn't a priority for sarge.09:04
=== azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel
=== seb128 [~seb128@ANancy-111-1-17-20.w80-13.abo.wanadoo.fr] has joined #ubuntu-devel
seb128morning10:28
jdubhey seb128 10:29
seb128hey hey jdub 10:29
seb128jdub: n-c-b 2.8.2, ok for update ?10:43
jdubyes thanks10:45
seb128thanks to you :)10:48
fabbionedaniels: how is the upload going?12:32
seb128jdub: I need to update gnome-cups-manager the current version doesn't include the .mo files ... I think that's ok ? :)12:36
seb128version = package, a missing line in the .install to fix, that's all12:36
danielsfabbione: good good12:38
danielsfabbione: do you need the interdiff now?12:38
danielsi can throw it to you if you need it12:38
jdubseb128: sure12:43
Kamionmdz: #13009? if it's not on bugs.debian.org, we don't have it :(01:02
Kamionlamont: I'm *fairly* sure the hostname gets set before debootstrap nowadays ...01:02
=== daniels [daniel@fooishbar.org] has joined #ubuntu-devel
lamontKamion: hrm.01:07
=== lamont up way to damn early.
lamontactually just up for a minute, checking to see if I'm back online or not...01:08
Kamionlamont: holy crap, man01:09
danielslamont: dude, that's inhuman01:10
lamontwell, the 12 year old wanted up early to work on homework...01:10
lamontsleep is good.01:10
=== lamont goes back to bed. back in about 35 minutes or some such.
fabbioneKamion: thanks for the procmail stuff.. i managed to move everything out of MUA's filters01:10
fabbionedaniels: please upload the patch now if you can so we don't desync the trees01:11
Kamionfabbione: cool. my procmailrc is a bit gnarly, I should tidy it up01:11
=== lamont giggles.
fabbioneKamion: ehehe well.. i didn't use all the bits, but it works fine01:11
lamontI'm associated with the _other_ hilltop.01:11
fabbionehey lamont 01:11
fabbioneyou are supposed to go back to sleep01:11
danielsfabbione: ok, i'll start uploading to people.nny.com/~daniels/01:11
lamontfabbione: check.  later01:12
fabbionedaniels: ok thanks01:13
fabbionethe interdiff should be pretty small anyway01:13
danielsyeah01:13
danielsit's ojust 099j, 099k, the dexconf synaptics tuff, and, a resync of 91101:14
danielsthat's tit01:14
danielsfabbione: interdiff uploading now01:15
danielsok, it's up at http://people.no-name-yet.com/~daniels/xfree86/xfree86_4.3.0.dfsg.1-6ubuntu19-to-20.diff01:17
fabbionetahnks01:24
fabbionedaniels: patch is merged01:34
fabbionedaniels: you need to ask somebody to build ubuntu20 on ppc01:34
fabbioneand verify the fix01:34
fabbionebrb01:34
Kamionmdz: did you get a chance to test with a modified debootstrap script for pcmcia-cs?01:42
jdubso this wacky acpi stuff01:43
jdubdoes the -26 patch mean we'll be able to load DSDT foo to make pooey laptops work?01:44
jdub"allow config to specify custom DSDT"01:44
jdub(hope that's not build time)01:44
thomyou append the DSDT to the initrd AIUI01:48
jdubi thought that was a different patch?01:55
=== jdub can't even build the stupid intel iasl thingy
pittiYeah, another stupid hal bug found and fixed01:59
jdubpitti: have you looked at the latest release?02:00
pittijdub: not yet02:00
pittijdub: we cannot sync that anyway, don't?02:00
pittiI debugged the stuff on mojo's computer02:00
pitti(he let me access it as root)02:00
jdubpitti: if we thought it safe enough and important enough, we probably could sync it.02:01
pittijdub: hmm. 0.2.97 had severe regressions02:01
pittijdub: I hope that 0.2.98 is better02:01
pittijdub: anyway, I don't trust this hal thingy too far, I would really prefer backporting bug fixes instead of syncing02:02
pittijdub: or just fixing the things on our own, as we have done until now02:02
jdubpitti: probably worth checking some of the bugfixes tho :)02:02
pittijdub: yes02:02
pittijdub: actually I tried to download it yesterday, but my URL is outdated.02:03
pittihave to google for it, I suppose02:03
=== jdub is surprised by disk speed once hdparm is set up nicely
jdubpitti: hal.freedesktop.org?02:03
pittijdub: this is impressively, I agree02:03
Kamionjust a shame it seems to break some systems02:03
pittijdub: actually I thought I looked there. Anyway, thanks02:03
pittijdub: yep, I looked there. It has hal-0.2, from Dec 200302:04
jdubhrm, bong02:08
jdubhttp://freedesktop.org/~david/dist/02:09
pittijdub: thanks!02:09
pittijdub: the buffer overflow I just fixed is still present in 0.2.29802:13
pitti0.2.98, that is02:13
pittijdub: I will send the patch to David02:13
jdubcool02:13
fabbionedaniels: diff -Naurd xc.orig/ xc > ../debian/patches/991_ubuntu_mga650_hack.diff 02:18
fabbionedaniels: it might even work :-)02:18
pittijdub: #1450, request for upload permission02:19
Kamionmdz: did you intend to remove binutils from Base along with ksymoops?02:24
=== Kamion is wary of that change
seb128jdub: what do you think about #1550 ?02:40
jdubseb128: i can take that one02:49
seb128jdub: you don't need, I just want to know what do youn think ... nautilus as default ?02:50
jdubseb128: yeah, i think the ftp handling is good enough now02:50
jdubseb128: especially with the keyring support02:50
seb128ok, I'll make a patch and send it upstream02:50
jdubthanks :)02:50
seb128np :)02:50
=== thom [~thom@amnesiac.heapspace.net] has joined #ubuntu-devel
=== pitti [~martin@195.227.105.180] has joined #ubuntu-devel
=== parkerc_ [~parkerc@cs671066-129.houston.rr.com] has joined #ubuntu-devel
=== lamont wonders if his connectivity is as crappy as it seems..
Kamionthom: would you be available to do some jigdo stuff for me today?04:05
thomsure04:05
Kamioninstructions are:04:06
Kamion  * Make ISO images as normal, and generate jigdo files (using either04:06
Kamion    mkisofs and jigdo, or mkisofs with the included JTE patch)04:06
Kamion  * Copy the .jigdo and .template files to a machine exporting the04:06
Kamion    mirror via HTTP04:06
Kamion  * On that machine, run mkjigsnap (see the man page) against those04:06
Kamion    files to generate the needed snapshot ready for download. I'd04:06
Kamion    suggest http://<mirror>/jigit/ as a download point.04:06
KamionI'm working on the first bit now04:06
Kamiondoes having an ssh trigger to do steps 2 and 3 sound feasible?04:06
thomyeah, sounds like the best way04:07
seb128jdub: gnome-vfs 2.8.1  ... ok for upload ? (dunno if I have to ask for each gnome 2.8.n releases :)04:12
=== Sledge_ [~steve@80.46.37.1] has joined #ubuntu-devel
Kamionhey Sledge04:13
Sledge_hey guys04:13
jdubseb128: ...04:14
seb128jdub: ....04:14
seb128:p04:14
jdubseb128: yes, ok04:14
Kamionso, need to make debian-cd on little generate jigdos, and therefore need to give it some URLs04:14
Sledge_Kamion: yup04:14
jdubseb128: in general, it's okay unless you think a change is controversial04:14
seb128ok04:15
jdubseb128: or too big04:15
KamionSledge_: should I set the fallback URL to http://archive.ubuntulinux.org/cdimage/jigit/, then?04:15
Kamion(assuming that's where we're going to put the snapshot archive)04:15
seb128don't worry, I don't make quick upload of new versions without testing :)04:15
Kamionseb128: aww04:15
Sledge_Kamion: I think you need to add a /snapshot/ in there too04:15
jdubseb128: ehhhhxcellent :)04:15
Sledge_just checking04:15
seb128:)04:16
KamionSledge_: hm, presumably we'd want to have a snapshot tree per release; we want to keep the one for warty around when we're developing hoary04:17
Sledge_yup04:17
Kamionso maybe http://archive.ubuntulinux.org/cdimage/jigit/warty/snapshot/04:17
Kamionyay long URLs :)04:17
Sledge_yes, something like that04:17
Sledge_the mkjigsnap script is set up to create dated snapshots in fact04:18
jdubseb128: upstream are killing wifi applet, using netstatus instead ;)04:18
KamionSledge_: do I leave JIGDOFALLBACKPATH unset? I guess mkjigsnap creates that04:18
Sledge_so the easiest way to use that would be to put symlinks in04:18
seb128jdub: yes, I've read that04:18
Sledge_Kamion: JIGDOFALLBACKPATH is used when creating the jigdos in the first place04:18
Sledge_mkjigsnap reads that jigdo rather than modifying it04:19
Sledge_the JIGDOFALLBACKPATH can be added by hand easily, in fact04:19
Kamionright, but it looks like debian-cd creates a hardlink tree yourself if you set that04:19
Kamions/yourself/itself/04:19
Sledge_yes, it does04:19
KamionJIGDOFALLBACKURLS I understand I need to set04:19
Sledge_gah, sorry04:19
Sledge_misparsed what you wrote04:19
Sledge_JIGDOFALLBACKPATH doesn't need to be set unless you're going to generate the images on the mirror04:20
Sledge_otherwise it wastes time creating a snapshot tree on the CD build machine04:20
Sledge_JIGDOFALLBACKURLS wants to point to the right place so that people downloading the jigdo bits by hand get the right stuff04:21
Kamionexport JIGDOFALLBACKURLS="Ubuntu=http://archive.ubuntulinux.org/cdimage/jigit/$CODENAME/snapshot/"04:21
Kamionlet's try that, then04:21
Sledge_yup04:21
KamionI think I'll need to frob debian-cd a bit to deal with the Ubuntu label though04:21
Sledge_ok04:22
Kamionleaving JIGDOINCLUDEURLS unset until I figure out what needs to happen there ...04:22
Sledge_you probably don't want it anyway04:22
Kamionwell, we do have a few mirrors04:23
Sledge_it's in case you want to include a standard block in all the jigdo places04:23
Sledge_ok04:23
Sledge_Kamion: the normal setting tat manty uses is JIGDOFALLBACKURLS="Debian=http://gluck.debian.org/cdimage/snapshot/cd/$ARCH/Debian/04:24
Sledge_on gluck04:24
Sledge_you may need to s/Debian/Ubuntu/ on that04:24
Sledge_or set it to Ubuntu=http://archive.ubuntulinux.org/cdimage/jigit/$CODENAME/snapshot/Ubuntu04:25
KamionI was thinking of dropping /Ubuntu from the path actually ...04:25
Sledge_ah, ok04:25
Kamionwe don't have a Non-US equivalent04:25
Kamionbut yeah, I've done s/Debian/Ubuntu/ on tools/jigdo_create04:25
Sledge_fine04:25
Kamionjust trying build.sh i386 now04:26
Kamionjigdo-file --cache=/srv/cdimage.no-name-yet.com/scratch/tmp/jigdo-cache.db make-template --force --files-from=/srv/cdimage.no-name-yet.com/scratch/tmp/jigdofilelist --image=/srv/cdimage.no-name-yet.com/scratch/debian-cd/i386/warty-i386-1.raw --jigdo=/srv/cdimage.no-name-yet.com/scratch/debian-cd/i386/warty-i386-1.jigdo --template=/srv/cdimage.no-name-yet.com/scratch/debian-cd/i386/warty-i386-1.template --merge=/srv/c04:27
Kamionthat's a command-and-a-half04:27
Sledge_:-)04:27
thomyow04:31
KamionSledge_: any way to convince it not to look at installer-$(arch-that-isn't-the-current-one)?04:34
KamionMatch of `dists/warty/main/daily-installer-amd64/20040801ubuntu13.0.20040918/doc/INSTALLATION-HOWTO' at offset 1400627204:34
KamionMatch of `dists/warty/main/daily-installer-i386/20040801ubuntu13.0.20040918/doc/manual/cs/apa.html' at offset 1402470404:35
Sledge_remove those from the jigdofilelist04:35
Kamionah, right04:35
KamionI think I might well end up switching to JTE, this isn't so fast ...04:35
Sledge_:-)04:35
Sledge_you noticed...04:35
Kamionit's bearable for daily builds certainly, but our preview release from start of CD builds through downloads and testing to release took an hour and five minutes04:36
Sledge_ok04:36
KamionI didn't realize earlier that I was going to be under quite such time pressure :)04:37
Kamionanyway, will start out with plain jigdo and speed it up later04:37
KamionSledge_: what does the Filename= field in the .jigdo do?04:46
Sledge_it specifies the default output filename for the .iso file04:47
thomKamion: btw, there's a file called make-torrent.sh in my ~ on little, and bittornado is installed05:32
=== mdz_ [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel
=== sivang [~user1@80.179.82.15.forward.012.net.il] has joined #ubuntu-devel
thomKamion: so you can integrate BT into the build too :-)05:32
Kamionthom: awesome, will wait until after I've finished my side of the jigdo stuff though :)05:33
thomsure05:33
thommdz_: are you happy for me to upload the new bittornado to warty? (upgrade from 0.2.0 to 0.3.0)05:34
mdz_thom: universe, right? sure05:34
thommdz_: nod05:34
Kamionmdz_: have you been following the parted stuff on #ubuntu?05:36
Sledge_Kamion: how's the jigdo thing going?05:36
KamionSledge_: just been rebuilding a couple of times to try to get all the filenames right, usual debian-cd tediousness05:37
Kamionum, "tedium"05:37
Sledge_:-)05:37
Sledge_sorry, that should be :-(05:37
Kamionmdz_: it looks like we may have to upgrade to parted 1.6.11 plus a Debian patch to fix it, which involves a soname change05:37
sivangbittornado is a full fledged BT gui?05:40
Kamionmdz_: the current version in unstable contains all our fixes, so in principle it's syncable, but I will have to review the entire diff05:40
thomsivang: no05:45
KamionSledge_: what's the SNAPSHOT= thing in the generated .conf file05:45
KamionSledge_: er, sorry, pressed Enter too soon05:45
thomit's much faster than bittorrent05:45
KamionSledge_: what's the SNAPSHOT= path in the generated .conf file relative to?05:45
Sledge_Kamion: relative to the place where the <foo>.conf is found05:46
KamionSledge_: ah, I think I see how to lay it out now05:47
Sledge_ok05:48
mdz_Kamion: no, I have been offline a great deal05:48
mdz_Kamion: what breakage does 1.6.11 fix?05:48
mdz_Kamion: if it fixes the "this ext2 filesystem has a rather strange layout", I might be sold05:48
Kamionmdz_: destruction of partition tables generated by modern versions of Windows05:49
mdz_eek05:49
=== elmo [~james@83.216.141.215] has joined #ubuntu-devel
mdz_elmo: I removed it from base and forgot to add it back to supported05:49
elmook05:49
sivangthom : conflicts with bittorent. just tried it and it "killed" the original client05:49
KamionSledge_: is there any way to have the final .jigdo and .template files generated on little (our actual CD image build machine) rather than on auckland (the machine where mkjigsnap's going to be run)?05:50
thomsivang: yep, the programs all have the same name05:50
mdz_elmo: fixed05:50
KamionSledge_: hm, actually, never mind, I'm being confused05:50
azeemsivang: this is fixed in unstable, AFAIK05:50
Sledge_Kamion: absolutely, that's exactly what I was expecting you to do05:50
Kamionmdz_: did you intend to remove binutils from base too?05:50
=== doko [doko@dsl-084-057-043-060.arcor-ip.net] has joined #ubuntu-devel
mdz_Kamion: it should go away on its own, shouldn't it?05:50
mdz_I think it was only there for ksymoops05:51
Kamionok, as long as that's what you intended05:51
KamionI saw it and went "eek" :)05:51
Kamionwill need a new debootstrap upload, then05:51
Kamionack/nack?05:51
mdz_ack05:51
sivangthom : we'd better change them? it it reasonable for someone to have both clients present?05:53
sivangthom : *is05:53
dokonpmccallum: I see you changed the bash prompt in skel.bashrc, but not in etc.bashrc. what was the reason for the change?05:53
Kamionthom: could you grab little:~cjwatson/jigit/mkjigsnap?05:54
thomsivang: as azeem said, it's fixed in the version i'm uploading05:55
Kamionthom: I think it needs to be run as 'mkjigsnap -n warty-i386 -m <path to archive mirror> -o <path to cdimage mirror>/jigit/warty -j <path to cdimage mirror>/daily/current/warty-$(arch).jigdo -t <path to cdimage mirror>/daily/current/warty-$(arch).template -k Ubuntu', for arch in amd64 i386 powerpc05:56
npmccallumdoko: the only difference I can see is the ':', right?  Probably just a typo...05:56
thomKamion: 'k, will look in a sec05:57
dokook. then I'll sync that. thanks for the color prompt anyway.05:57
Kamionthom: I think you'll also need to change the cdimage archvsync trigger to exclude jigit from the sync, otherwise it'll get trashed05:58
npmccallumdoko: yeah, I hate bw prompts -- when you do ls 2x, you cant tell where one ls ended and one begain06:02
dokois there any reason that we don't source /etc/bash.bashrc from /etc/profile? I'd like to sync that with the skeleton files06:05
npmccallumdoko: I have no idea06:14
Kamionthom: ok, there are now .jigdo and .template files in /cdimage/daily/current/ which should be usable06:33
Sledge_Kamion: how do I go about getting the jigit package into ubuntu?06:39
KamionSledge_: in general, mail ubuntu-devel@lists.ubuntu.com; will need to convince mdz/jdub06:40
KamionSledge_: once we've demonstrated it to work with our ISOs it should be a much easier sell06:40
Sledge_of course06:41
Kamionanyway, I think the ball's in thom's court for the time being, I'll wait until he bounces it back to me :)06:43
Sledge_:-)06:43
Sledge_Kamion/thom: can you mail me when the bits are in place on archive.ubuntulinux.org so I can d/l and test too?06:44
KamionSledge_: will do06:45
thomyep.06:45
thomjust having some tea06:45
elmooi06:45
thomand celebrating australia losing to england at cricket06:45
elmowhat the hell are you guys running on auckland06:45
Kamionuh, not me?06:46
elmono, talking about running I mean06:46
Kamionelmo: hardlink snapshot of the archive at CD generation time so that jigdo is even possible06:46
elmowe need to not be doing stuff on auckland - the torrent stuff was a mistake that happened 'cos it was at the last minute. 06:47
Kamionelmo: correct, all the jigdo stuff will happen on little, and bittorrent is being moved there06:47
elmokamion: what needs the harlinks?  jidgo server or jidgo user?  does the snapshot need to be publicly available?06:47
elmokamion: okay, if it's happening on little, what's thom doing?  his r00t powers scare me :P06:47
Kamionelmo: in order to do a jigdo download, the client *must* have access to a fallback snapshot of the archive at the point when the CD was generated06:48
Kamionelmo: most of the downloads happen from a local mirror, but you have to have the fallback snapshot otherwise the download will fail as soon as anything in the archive changes06:48
elmokamion: so jidgo users all come hit up our BW?  reeeesult06:48
Kamionelmo: no more than they would by grabbing the ISO, usually much less06:49
elmoyeah06:49
Kamion(this is the same as what gets done on gluck, BTW)06:49
=== Sledge_ hides from elmo
elmoI thought most of the cdimage stuff had moved to maswan?06:49
Kamionthe jigdos are built on gluck; if maswan syncs the snapshot archives too, it's his bandwidth I guess06:50
Kamionsee gluck:/org/cdimage.debian.org/www/snapshot/06:50
Kamionelmo: the thing I wanted done on auckland is little:~cjwatson/jigit/mkjigsnap, which is Sledge's script to hardlink-snapshot an archive; I need to have auckland run it after it finishes the cdimage archvsync06:52
Kamion(it's a bit of a race technically, but as long as the CD build time doesn't exceed the archive's StayOfExecution then it works)06:53
Sledge_Kamion: how long are you thinking of keeping each snapshot around?06:53
elmoSOE is a day06:54
Kamionwell, we're building stuff daily, so absolutely not more than a week except for say the warty release jigdo06:55
Sledge_ok06:55
Kamionor sounder-test images I guess :-/06:55
Sledge_the mkjigsnap code deliberately creates dated snapshots, so it should be easy to clean up old ones with rm -rf <DATE>06:56
Kamionelmo: are we still in "AARGH NO RUN AWAY" territory?06:56
Sledge_time to go home and eat...06:58
elmokamion: one sec07:00
Kamionthom: I think I've got bittorrent integration working now07:21
Kamionpretty easy once I had the script07:22
sivangKamion : how much time after a package has been uploaded until it's on the archive for downloading?07:23
Kamionit's better to direct questions to the channel rather than to me personally07:24
Kamionsivang: anyway, the archive's "cron.daily" runs once every half-an-hour, source uploads get installed then, after each cron.daily run the buildds pick it up and then the binaries get installed at the next cron.daily after that07:25
sivangofcourse :)07:25
Kamionsivang: so absolute minimum 30-something minutes07:25
sivangok, thanks07:25
sivangyeah. figured that.07:25
sivangtnx07:25
Kamionelmo: who's the admin of the Italian mirror, and shouldn't that mirror be listed in Archive on the wiki?07:40
elmokamion: md07:40
elmoand probably sure, he just mentioned it on irc so it got added to the website by jane/lu/someone07:40
Kamionseems to be just a mirror of the preview release07:43
Kamionit doesn't have MD5SUMS which is why my attention got drawn to it07:43
Kamionthom: does the bittorrent tracker need any extra magic?07:51
Kamionthom: I've tried starting a download, but it's just sitting at "connecting to peers"07:51
=== lamont_r [~lamont@dialup-4.227.235.124.Dial1.Denver1.Level3.net] has joined #ubuntu-devel
Kamionthom: or do things get strange when the client is behind NAT?07:53
thomit doesn't like nat at all08:05
thomwell, unless you can forward the ports correctly08:06
thomKamion: hrm, i think the mirror script is going to have to kick off the bt seeders08:10
Kamionthom: what are those?08:11
thomyou have to have something to seed the tracker, ie you run a bt client on the local machine to get the file into the torrent08:11
danielselmo: dude, thankyou so very much08:12
Kamionthom: that really wants to be done on auckland rather than little, right?08:12
danielselmo: source-only uploads absolutely rock my world08:12
danielselmo: so, I uploaded i386+source+all xfree86 to rookery (not including the orig)08:13
danielselmo: 18 hours on, it's just finished08:13
thomKamion: yep08:13
thomso the mirror script would need to kick them off when it finishes the rsync08:14
thom(we pushed 1500 i386 isos of the preview via BT, btw :-) )08:15
Kamionnice08:16
Kamionthom: that's entirely auckland-side, isn't it? little doesn't know when the rsync is finished08:16
thomyep08:17
thombut little logs in to the archvsync user and runs a script auckland side to run the rsync, doesn't it. i'm saying we tack it on that08:17
Kamionthom: right, absolutely08:24
KamionI've never seen that script08:24
thomKamion: it's exactly what you'd expect :-)08:28
Kamionthom: also, presumably the torrents need to be regenerated when I copy a daily build to release08:37
Kamionthom: so it isn't just 'cp -a' any more08:37
Mithrandirnpmccallum: what TZ are you in?08:54
npmccallumMithrandir: EST08:58
npmccallumMithrandir: my hours are a little funky though (not too bad)08:58
Mithrandirthat's UTC-5 or so?08:58
Mithrandirok, I'm at fairly-normal CEST.08:58
Mithrandirbut I prefer to do work-work in the evening, so if UTC16-ish is ok with you?08:59
npmccallumMithrandir: depends on the day, but generally09:00
npmccallumMithrandir: check my time schedule on the company wiki09:00
Mithrandirok09:00
npmccallumMithrandir: I'm going to run to the store, should be back in about 10 if you need anything09:03
Mithrandirsure, I have a problem with evo I need to track down..09:03
Mithrandirman, I love corba. :/09:03
npmccallumwhat is the best way to download the entire archive (warty/main)?09:04
Mithrandirrsync?09:05
npmccallumI just only want the newest stuff, I was wondering if there was some debian tool to do it09:06
Mithrandiryou can use debmirror or some other mirroring tool as well09:06
=== Kamion wonders why partman's progress bar for resizing a partition doesn't update from 0%
Mithrandirperhaps parted is busted?09:17
Kamionprobably09:19
Kamionor partman, who knows :)09:19
Kamionkernel          /boot/vmlinuz-2.6.8.1-2-386 root=/dev/hda6 ro vga=771 acpi=force quiet splash09:20
Kamionhooray, the user-params stuff works for me09:20
=== lamont_r [~lamont@dialup-4.225.209.51.Dial1.Denver1.Level3.net] has joined #ubuntu-devel
sivangIs there a specific doc for PXE install for ubuntu? or plain 'o debian would suffice?09:48
=== lamont_r wonders if his mail to ubuntu-devel actually made it through...
sivanglamont_r : it made it. about the buildlogs right?09:52
lamont_ryeah09:53
=== lamont_r is out of the log-db-query-engine business
lamont_rwhich is good, since my router appears to have not survived the power hit this morning... :(09:53
sivangohh, i am going to build myself a ubuntu router , you want me thorugh in another for you ? (trying to encourage lamont_r)09:54
lamont_rthat's probably what I'll do - I have a more hardened x86 box, will probably switch from the hppa router to an ubuntu-powered x86 box09:55
=== lamont_r heads to town for a while, just to get away from everything...
lamont_rback on in a while.09:55
sivanglamont_r : you feeling better since that nasty virus? I'm on pills for the last couple of days. only that fixed my dizziness09:56
lamont_rrunning at about 80-90% I'd guess09:57
sivangif only we could patch those..:)09:58
lamont_ryeah09:58
lamont_ranyway, back on in a while10:00
dokoMithrandir, mdz: is there an overview, which packages use spell checking? or just get the reverse dependencies on aspell/ispell/myspell?10:14
Mithrandirdoko: i$language is the norm for ispell at least, aspell-$langcode and myspell-$langcode seems to be the norm for those10:15
mdzdoko: the ones which are high priority are evolution, openoffice and mozilla-thunderbird10:18
mdzdoko: let's say evolution and openoffice, those are in desktop10:19
mdzfollowed by mozilla-thunderbird, abiword, and other stuff in Supported10:19
mdzI believe gaim also does spell checking10:19
dokook, that's a starting point.10:19
Mithrandirmdz: it does.10:19
Mithrandirspellchecking should Just Work for those, in major languages10:19
Mithrandir?10:19
Mithrandirwe need to install extra dictionaries based on $LANG10:19
mdzbut does it?  please check10:19
mdzwe need to get a handle on which are working and which are not, and which spell checking libraries they use10:20
Mithrandirmdz: yes, I'm talking about the goal, not the current state. :)10:20
mdzand ensure that at least the english dictionary is installed by default10:20
mdzthen we can think about ways to get additional dictionaries10:20
mdzfor warty, I am not sure if it is feasible to select them at install time automatically10:20
Mithrandirwe need to check that the dictionaries work properly as well10:20
mdzwe may need to choose between including them by default, or requiring that they be installed later10:20
Mithrandiryeah10:21
dokodictionaries: these are the w<language> packages? or the aspell-$LANG, ispell-$LANG packages10:21
Mithrandirdoko: w$lang is just wordlists10:22
mdzfabbione: here?10:24
sivangmdz : from how it looks curerntly, most 2004 universe DSAs would be fixed by a sync from sid. :) (halfway finished)10:39
mdzsivang: great, so it all comes down to whether they can be synched without changing main10:46
mdzthom: ping?10:46
Mithrandirmdz: you are going to want to kill me.10:47
Mithrandirmdz: evolution needs a TLS enabled glibc.10:47
mdzMithrandir: on amd64?10:47
sivangmdz : yes, however i havn't still reviewed 2002 and 2003, but 2004 is highest priority.10:47
thommdz: hi10:47
Mithrandirmdz: yes.10:48
mdzthom: hey. what's the good word on firefox?10:48
sivangmdz : that is, the universe relavent parts.10:48
Mithrandirmdz: or rather, in general10:48
mdzMithrandir: our glibc is TLS-enabled on i686 (with libc6-i686), is it not?10:48
mdzMithrandir: is it only a matter of turning some bits on, or something worse?10:48
Mithrandirmdz: it's always TLS-enabled on i68610:48
Mithrandirmdz: well.. it hasn't been _that_ much tested..10:48
Mithrandir(on amd64)10:48
thommdz: i've got a few more rejects to clean up, and then a lot more testing10:49
mdzMithrandir: TLS is something which is only used when the application specifically uses it, is it not?10:49
Mithrandirmdz: it uses the DB backend with DB_THREAD and that requires "fast mutexes" to be available.10:49
Mithrandirmdz: nope.10:49
Mithrandirmdz: you can turn it off with LD_ASSUME_KERNEL=2.4.110:49
mdzthom: what's the story with the localisation?  I was thinking that even if they are outdated upstream, we should relax the deps and at least get partial translations10:49
thom(mostly because they seem to have funted with the l10n support since the last release)10:49
mdzthom: then it will feed into the warty translation effort10:50
mdzs/warty/ubuntu/10:50
Mithrandirmdz: another option is to rewrite the backend for evo.  (No, I'm not doing that)10:50
thommdz: that's what i'm planning to test; i'm not sure how much they've changed - hopefully not so much that the old packages won't work at all10:50
mdzthom: ah, I see10:50
mdzMithrandir: the glibc changes should only affect amd64, right?10:50
mdz(please say yes)10:50
Mithrandirmdz: yes, they should10:50
Mithrandirit's a very trivial change to the source package, and I can test it for a few days before we upload if you want that)10:51
mdzMithrandir: please do test it, of course :-)10:51
Mithrandirobviously, yes.10:51
Mithrandir:)10:51
=== pitti [~martin@195.227.105.180] has joined #ubuntu-devel
=== thom goes for an early night and vanishes
=== Mithrandir has learnt far too much about evo today
=== sivang is going for PXE warty install, and than to bed. Night everybody
thom'night10:54
mdzKamion: how does the pcmcia-cs magic work?  does it use laptop-mode, or something else?11:08
Kamionmdz: it tries to start up pcmcia, then looks in /sys/class/pcmcia_socket11:10
mdzhmm11:10
Kamion(it needs to start up pcmcia anyway)11:10
mdzKamion: I had forgotten that pcmcia starts during the installer too...11:10
mdzKamion: I wonder how it is that it worked in d-i but failed during a normal boot (https://bugzilla.ubuntu.com/show_bug.cgi?id=581)11:11
Kamion       -f     Foreground:  do not fork and run as a daemon until after config11:11
Kamion              uring any cards that are already present.11:11
Kamionit uses that option, maybe that's relevant11:11
Kamion(cardmgr)11:12
mdzthat's exactly the opposite of what I'd expect, how weird11:12
mdzthe problem in #581 was that it sat there in the foreground forever11:12
Kamionalthough actually apparently that's to avoid a race condition11:12
mdzif it's harmless to start pcmcia under d-i, it should be harmless to install it too11:13
mdzbut given the symptoms, it does sound like this would solve the problem anyway11:13
Kamion                CARDMGR_OPTS="-f" /etc/init.d/pcmcia start </dev/null 3<&0 2>&1 \11:13
Kamion                        | logger -t hw-detect11:13
Kamionthat's the actual invocation11:13
mdzI have (remote) access to the machine where it was failing before11:14
mdzanyway, in the interest of solving this upstream's way, I'm OK with pcmcia-cs moving to ShipSeed11:15
Kamionit definitely doesn't have PCMCIA? that would rule out having any cards that are causing problems11:15
Kamion    - Add a new template, hw-detect/start_pcmcia, as a kind of workaround for11:17
Kamion      systems where starting pcmcia hangs the system (Dell inspirons..).11:17
Kamion(ddetect 0.87)11:17
Kamionhw-detect/start_pcmcia=false is documented in the boot screens as a workaround for systems where starting PCMCIA crashes11:17
Kamionthere's no similar workaround for post-install if we have pcmcia-cs in Base, though11:18
KamionI think the submitter of #581 must just have been lucky pre-reboot :-/11:18
Kamionso yeah, I think there are enough arguments11:19
mdzI'm fairly certain it doesn't have pcmcia11:19
mdzhmmm11:19
mdzunfortunately, someone uses that machine as their primary desktop to do work11:20
mdzso I've been trying to schedule a time when I could (potentially) break it for testing11:20
Mithrandirwe don't ship any 2.4 amd64 kernels?11:22
KamionMithrandir: no11:25
Mithrandirgood11:25
Kamion*are* there any 2.4 amd64 kernels? :) I thought 2.4 was totally unsupported on amd6411:25
Mithrandirthere are kernels, but nobody should use them11:25
Kamionin fact, we don't have any 2.4 kernels in supported for any architecture11:26
MithrandirI'm removing support for < 2.6 for glibc for amd64, that's why I was asking11:26
mdzKamion: ubuntu-base needs to be added to base; do you have an upload pending or shall I do it?11:40
Kamionno upload pending, but I'm used to debootstrap uploads and can do it easily11:42
KamionI should put my magic autoupdate script somewhere11:42
Kamionhalf of it's in the debootstrap package ...11:42
Kamionthe other half relies on germinate though11:43
mdzKamion: ok, I'll leave it to you, then11:45
Kamionmdz: hm, you closed that util-linux bug with "we haven't fixed this yet", but LaMont backported that fix to warty already11:45
mdzKamion: he did?  he didn't close the bug11:45
Kamion util-linux (2.12-7ubuntu3) warty; urgency=low11:45
Kamion .11:45
Kamion   * Incorporate patch from Samuel Thibault via Debian.  Closes warty#46811:45
=== Kamion is catching up on warty-changes
mdz468   maj   P2   Oth   lamont.jones@canonical.com   ASSI     delay in spawning new getty after logging out on console11:46
=== mdz growls at lamont
mdzKamion: if you have the bug number handy, please reopen11:46
Kamionmdz: what sort of things do you foresee going into ubuntu-base?11:48
Kamionmdz: done11:49
mdzKamion: the little script to enable universe, an upgrade tool for warty->hoary if we need one, stuff like that11:49
Kamionok11:50
Mithrandirmdz: 1545 will be fun to track down when I don't have such a box. :/11:52
Kamionmdz: ubuntu-base doesn't seem to have built11:52
KamionMithrandir: I'll forward you mail from Robert and Goswin if either of them replies11:53
MithrandirKamion: thanks.11:53

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