[00:02] <SpamapS> Ok maybe not a crumper ;)
[00:03] <SpamapS> lol wrong window
[00:03] <SpamapS> I just love buffering
[00:03] <SpamapS> slangasek: released to updates
[00:21] <slangasek> SpamapS: cheers :)
[00:36] <infinity> SpamapS: And actually accepted to -updates. ;)
[01:39] <naryfa> If there's anybody advanced here, or if anybody could point me in the right direction I'd appreciate it. I'm trying to change animation in gnome-shell's js/ui. Anybody familiar with the code?
[03:24] <vibhav> Good Morning
[04:03] <vibhav> https://bugs.launchpad.net/ubuntu/+source/request-tracker3.8/+bug/769765 : can it be nominated for oneiric?
[04:04] <vibhav> (for request-tracker4 only)
[04:17] <vibhav> oops
[04:20] <cjwatson> (people have probably noticed that 12.04 is out by now)
[04:20] <vibhav> heh
[04:22]  * vibhav waits
[04:23] <vibhav> cjwatson: Can you nominate https://bugs.launchpad.net/ubuntu/+source/request-tracker3.8/+bug/769765 for precise?
[04:25] <cjwatson> no, I've been awake for goodness knows how many hours, critical judgement is beyond me and I'm about to crash
[04:45] <infinity> vibhav: With recommends being installed by default, that change leads to fairly unpleasant behaviour.
[04:49] <vibhav> Is it installed by default?
[04:49] <infinity> Recommends are.  As in, if you "apt-get install foo" all of foo's recommends get installed.
[04:49] <vibhav> ah
[04:51] <infinity> So, that change means installing mod_perl, mod_fcgi, mod_fcgid, and dbi-perl just because someone wanted RT.  Seems suboptimal.
[04:52] <vibhav> "Add Recommends on all Apache-related modules; although any one can be used to produce a working configuration, installing them all will result in less confusion
[04:52] <vibhav> Says the changelog
[04:52] <infinity> I see that.  But that's silly.  If any one can be used to produce a working config, then do so? :P
[04:54] <infinity> If the problem is that speedy-cgi-perl isn't actually good enough on its own, then the dependencies need to be fixed to deal with that, not make people install every other possible working config.
[04:58] <infinity> It might just be that the deps here are expressing something a bit wrong.  I'm betting that you need libapache-dbi-perl to use RT with mod_perl2.  So, having it depend on "mod_perl | dbi-perl | other_method" is weird.
[04:58] <infinity> Dropping the mod_perl dep entirely might fix it, since libapache-dbi-perl already depends on mod_perl.
[04:58]  * infinity shrugs.
[04:59] <vibhav> SO shall I drop mod_perl dep?
[04:59] <infinity> I'm not sure how the RT packaging then decides which method you have installed and how it configures them, but it feels very wrong to me to just go nuclear and install and configure them all. :P
[05:00]  * vibhav nods
[05:00] <infinity> If the mod_perl/dbi-perl method is the preferred one (which it is currently, being at the head of the list), I'd move libapache-dbi-perl to the front of the list and drop mod_perl from there.  That probably fixes the general case.  Probably still doesn't make it perfect.
[05:01] <infinity> Actually, if every one of those ORed lists started with libapache-dbi-perl (and no mod_perl) I think that would express what the packager wanted, yeah.
[05:02] <infinity> You'll then end up with either mod_perl/dbi-perl, or one of the other CGI options.
[05:02] <infinity> But never mod_perl without dbi-perl, which is what's causing the bug.
[05:03] <vibhav> SO what shall I do?
[05:04] <infinity> I'd recommend the above to the Debian maintainer.  And do the same in Ubuntu.
[05:04] <infinity> Do, drop the Recommends: line entirely, and on all the "foo | bar" deps, remove mod_perl, and move dbi-perl to the first position.
[05:04] <infinity> s/Do/So/
[05:05] <infinity> dbi-perl depends on mod_perl, so that gives exactly what you want, IMO.
[05:09] <infinity> vibhav: Added a lazy copy/paste comment to the bug.
[06:03] <SpamapS> infinity: thanks for the alley-oop there ;)
[06:07] <vibhav> Can https://bugs.launchpad.net/ubuntu/+source/mutt/+bug/965772 be nominated for precise?
[12:06] <melodie> hi
[12:06] <melodie> I want to try to learn how to package, and got the doc, printed it, installed all the build programs.
[12:07] <melodie> I would like to know if there is a particular difficulty when starting if the program to be packaged does not provide a configure ?
[12:17] <jtaylor> melodie: that can be a problem, packaging is easier if it uses a more or less standard buildsystem
[12:18] <melodie> hi jtaylor
[12:20] <melodie> the dev told me he will not go throught the process of making a configure because he can't take the time for that. I have been using his program in several distros along the years and would like to bring it to Ubuntu. In fact, this little program is the key to a type of openbox environment.
[12:20] <jtaylor> how does it build? just a makefile?
[12:20] <melodie> jtaylor, do you think the people here are likely to help me along the process,
[12:20] <melodie> it builds with a Makefile, yes : just "make" and "sudo make install" does the work.
[12:21] <jtaylor> does the makefile honor DESTDIR?
[12:21] <jtaylor> then your usually fine
[12:21] <melodie> I have installed a Ubuntu Mini, and now all works
[12:21] <melodie> does the file honor DESTDIR ? If I write DESTDIR in the packaging files it will go to the dir I say. For now, the "make install" has send the binary to /usr/bin
[12:22] <melodie> is my answer relevant ?
[12:22] <jtaylor> make install should put the files in $(DESTDIR)/$(prefix)/...
[12:23] <melodie> well, I asked the dev, he said I can use destdir
[12:26] <melodie> I have a first question about the doc. Here is the page I am reading now : http://wiki.debian.org/IntroDebianPackaging#Step_1:_Rename_the_upstream_tarball
[12:26] <melodie> how come all the commands are root ? Shouldn't they be done as user ?
[12:27] <jtaylor> why should they be root?
[12:29] <melodie> jtaylor, this is the question I am asking : the howto shows commands done as root : do you think that it is an error ?
[12:29] <jtaylor> I don't see that there
[12:30] <melodie> ie : # mv hithere-1.0.tar.gz hithere_1.0.orig.tar.gz
[12:30] <jtaylor> no commands besides pbuilder need root, some of them (e.g. installing) need fakeroot
[12:30] <melodie> # tar xf hithere_1.0.orig.tar.gz
[12:30] <melodie> all the commands in that page are starting with a hashtag
[12:30] <melodie> this is why I wonder
[12:31] <jtaylor> that just denotes its a shell command
[12:31] <jtaylor> it has nothing to do with which user does it
[12:31] <jtaylor> somethims $ or > is used instead
[12:31] <melodie> ok
[12:32] <melodie> what about the archive format ? I have a tar.bz2 : is that ok or should I rearchive it in another format ?
[12:33] <jtaylor> thats fine
[12:33] <melodie> thks
[12:45] <melodie> I have done the changelog file with the dch command as the howto says, and now I read at debian/compat : but there is no more command to create it and the following file, which should be "control": should I use another howto to know what are the precise steps from now ?
[12:48] <melodie> perhaps the debian doc is not up to date ? or missing details ?
[12:49] <melodie> I could use this one also : http://translate.google.fr/translate?sl=fr&tl=en&js=n&prev=_t&hl=fr&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fdoc.ubuntu-fr.org%2Ftutoriel%2Fcreer_un_paquet&act=url
[12:49] <melodie> ?
[12:50] <jtaylor> debian/compat is just a number in a file
[12:50] <jtaylor> for control just copy the one from the guide and change the values
[12:52] <melodie> ok, therefore I create them manually. ok. :)
[12:53] <pitti> cjwatson: indeed, it should probably depend on p-gi | p3-gi
[13:07] <melodie> I have installed yelp and had to start it from console as I could not find it in the menus. there, it crashed, wich triggered a window which sent (supposedly) a bug report. Is that enough or is it necessary to go report the bug manually too ?
[13:07] <melodie> this is in Ubuntu mini 12 04
[13:30] <melodie> is the minimum version number for Build-Depends always needed, or just the package name is ok ?
[13:32] <jtaylor> if the minimum version is available in all stable releases you can skip it, or if you don't want to backport it
[13:32] <melodie> jtaylor, thanks
[13:34] <dupondje> heh, article on tweakers.net about EA games on Ubuntu :)
[13:39] <melodie> now about the depends : the doc says the codes in the default line are magic. I don't figure out, if I just leave this line alone, or if I leave the default and add the depends ? The program needs libmenu-cache1 and libgtk2.0-0
[13:40] <melodie> ie:
[13:40] <jtaylor> melodie: lets move this to #ubuntu-packaging please
[13:40] <melodie> jtaylor, I didn't know that chan : thanks !
[15:20] <SpamapS> ugh, moving partitions around on my mac seems to have confused grub
[15:29] <vibhav> Can somebody nominate https://bugs.launchpad.net/ubuntu/+source/mutt/+bug/965772 for precise?
[15:31] <SpamapS> vibhav: anybody can nominate it, only core devs can accept it. :)
[15:34] <jtaylor> I think that changed
[15:34] <jtaylor> only bug control can nominate
[15:35] <angelo-c> hi guys, i made some patches to a nearly abandoned software, xoscope, i would be really glad to make my patches going upstream to debian but i don't know how to do, can anyone help me?
[15:36] <cjwatson> so if vibhav is doing lots of this, he should join bugcontrol rather than asking here a lot ...
[15:36] <dupondje> angelo-c: report a bug on the package in debian. :)
[15:37] <angeloc> dupondje: i thought there was an automated procedure in launchpad
[15:39] <dupondje> really? don't think so :)
[15:40] <infinity> There's submittodebian.
[15:40] <infinity> But it's really just a fancy wrapper around debdiff and reportbug.
[16:14] <angeloc> dupondje: i think that really a low percentage of contribution going upstream without an automated system
[16:16] <angeloc> dupondje: and also opening a bug in debian is really a pain!
[16:17] <SpamapS> cjwatson: any advice for /usr/sbin/grub-setup: error: will not proceed with blocklists.
[16:17] <SpamapS> ?
[16:26] <angeloc> dupondje: there is an automated method, using submittodebian you can send a patch upstream, really cool
[16:27] <penguin42> SpamapS: Was that the only thing it said?
[16:28] <SpamapS> penguin42: no, thats just the most annoying part of what it said. ;)
[16:28] <SpamapS> /usr/sbin/grub-setup: warn: This GPT partition label has no BIOS Boot Partition; embedding won't be possible!.
[16:29] <SpamapS> /usr/sbin/grub-setup: warn: Embedding is not possible.  GRUB can only be installed in this setup by using blocklists.  However, blocklists are UNRELIABLE and their use is discouraged..
[16:29]  * SpamapS tries --force because the alternative is a reinstall at this point. :-P
[16:29]  * SpamapS reboots w/ fingers crossed
[16:31] <SpamapS> woot
[16:31] <SpamapS> worked
[16:32] <SpamapS> would be cool if gparted detected a grub install on a disk and told you to run that command. Oh wait, it did detect it, and warned me to stop at least. :)
[16:32] <SpamapS> cjwatson: n/m, fixed w/ --force (based on your response to a user on the grub mailing list :)
[16:43] <cjwatson> SpamapS: I strongly recommend creating a BIOS Boot Partition.  http://www.gnu.org/software/grub/manual/grub.html#BIOS-installation
[16:55] <ali_> hi, how can i query for synaptic progress from a command line, im trying to add unity progressbar to synaptic, and that's my missing part of the puzzle, thanks :)
[17:01] <SpamapS> cjwatson: indeed, I think I will do that after reading that page a bit ago.
[17:01] <SpamapS> cjwatson: does the +mac desktop installer do that btw?
[17:02] <SpamapS> I installed this machine in 11.10, and several things were pretty wonky
[17:03] <cjwatson> SpamapS: it's supposed to
[17:04] <cjwatson> I did fix various bugs in that general area in 12.04; I forget the exact details at this point
[17:04] <penguin42> cjwatson: Did you see Matt Garrett's write up of the Fedora 17 ISO image - it's pretty insane
[17:05] <cjwatson> Yes, I did
[17:05] <cjwatson> Need to figure out how to port that to xorriso
[17:05] <penguin42> I like his blog, it's normally a world of insane bios level pain
[17:22] <tjaalton> >
[17:22] <tjaalton> >
[17:22] <tjaalton> > Regards,
[17:22] <tjaalton> > David
[17:22] <tjaalton> >
[19:14]  * slangasek wonders if anyone would care to verify the SRU fix for bug #989585 using the test case there
[19:16]  * soaringsky gawks at the huge number of duplicates on that bug...
[19:17] <slangasek> that's why I'd like someone to verify that it fixes it so we can publish it ;)
[19:18] <soaringsky> slangasek: I would do it, but my slow internet means that the release upgrade would take hours
[19:18] <slangasek> no worries
[19:21] <IntuitiveNipple> I missed it; which bug? Maybe I can test?
[19:22] <soaringsky> bug #989585
[19:25] <slangasek> oops, turns out that introduces a regression; marking it verification-failed myself :/
[19:26] <IntuitiveNipple> OK ... I was going to test it in one of my VMs
[19:38] <slangasek> cjohnston: I've continued to fail to get a working virtualenv going here; now model-mommy (pulled in via zope.interface) is broken, AFAICS from the output the upstream tarball is simply missing the tests/ subdir.  is it known to not be possible to run summit against the distribution packages?
[19:41] <slangasek> cjohnston: nevermind, model-mommy doesn't seem to be packaged anyway. :P
[21:05] <lifeless> cjwatson: around? I have an oddity with the alternate installer + partman + dmraid01 [mirror then stripe over 4 disks]
[21:07] <lifeless> specifically, the installed doesn't activate the partitions, even though it shows the arrays and shows the partitions
[21:08] <lifeless> this leads to partman trying to run tune2fs -l <arraydevice> rather than on the specific partition
[21:09] <lifeless> running dmraid -ay subsequently activates the partitions for the kernel, but it still trys to run it on the array device..
[21:20] <JanC> lifeless: there is also #ubuntu-installer (although cjwatson is both here & there, of course)
[21:21] <lifeless> JanC: thanks; I'll lurk there and follow up whereever cjwatson does ;)
[21:41] <cjwatson> lifeless: eh, not sure I have enough brain to debug that today :-)
[21:41] <cjwatson> psusi's been doing most of the dmraid-related stuff of late
[21:42] <cjwatson> he might well know better than I
[21:43] <lifeless> cjwatson: thanks
[21:43] <lifeless> cjwatson: I suspect raid01 dmraid isn't handled by the dmraid glue in partman; I can file a bug if you like (as I can apparently get this far nondestructively...
[21:44] <cjwatson> I didn't think partman generally knew much about different dmraid levels; might be more like parted
[21:45] <cjwatson> is dmraid01 one of those where there are multiple steps if you try to trace through the dm tables?
[21:45] <cjwatson> IYSWIM
[21:45] <cjwatson> it might be that parted only takes one step when it needs to recurse, or something
[21:45] <cjwatson> I vaguely recall dealing with this in grub2 once
[21:46] <cjwatson> debian/patches/dmraid.patch in the parted source package is probably the relevant bit of code here
[21:46] <cjwatson> and psusi is the last person who rewrote that :)
[21:46]  * kees still needs to refresh the bit-rotting failed raid/lvm hooks
[21:46] <lifeless> cjwatson: yes, it is.
[21:47] <lifeless> cjwatson: you get three dm devices; array0-1 (a mirror set), array0-2 (a mirror set), array0(a stripe over the -1 and -2 devices)
[21:47] <cjwatson> right, then my jetlagged guess is that dmraid.patch needs to be smartened up for that, possibly by comparison with similar code in grub2
[21:48] <lifeless> cjwatson: does it interrogate the kernel, or reimplement bits?
[21:52] <cjwatson> lifeless: it asks the kernel about the table structure; it's tracing through those manually, but I'm not sure how much of that the kernel would let it do any other way
[21:52] <cjwatson> dmraid.patch is only 133 lines, not that complicated
[21:54] <lifeless> thanks for the pointers
[21:55] <lifeless> cjwatson: as a workaround, is it possible to tell parted to not be smart, and just operate on a given device ?
[21:56] <cjwatson> I don't know but I doubt it
[21:57] <lifeless> cjwatson: interesting; given that dmraid can't be really be configured within the OS I am now curious why parted introspects it at all :)
[21:58] <cjwatson> because it needs to exclude some devices from being shown
[21:58] <cjwatson> otherwise people try to partition things at the wrong level
[21:58] <cjwatson> perhaps I've misdiagnosed the problem though
[21:58] <lifeless> ah yes
[21:59] <lifeless> thats a good reason
[22:00] <lifeless> its a shame the kernel can't just flag devices as being parts of higher order things, but perhaps it does - I haven't looked at the non-kernel interfaces around dmraid
[22:00] <lifeless> non-kernel-internal, I mean.
[22:22] <melodie> gn
[22:29] <matthew-parlette> I'm looking to make alarm-clock more useful with the HUD, but I can't find any links for development support on it, is there anything written down for it yet?