[00:00] <kirkland> slangasek: okay, my focus is on pam/ecryptfs now....
[00:00] <verwilst> does pbuilder use the sources.list of the system?
[00:01] <verwilst> Version: 1.9.0.1+build1+nobinonly-0ubuntu0.8.04.3 is available there...
[00:01] <kirkland> slangasek:  I didn't see any debian/pam-configs/ within the pam package ....  perhaps you meant another package?  i'm downloading the bzr branch (on the hotel's sad wireless) right now....
[00:02] <slangasek> kirkland: which version were you looking at?
[00:02] <verwilst> any idea what i could be doing wrong?
[00:03] <kirkland> slangasek: hmm... maybe i went about this wrong... apt-get source pam pulls pam-1.0.1
[00:04] <slangasek> kirkland: at what Debian revision? :)
[00:04] <kirkland> that sounds conspicuously nascent
[00:04] <kirkland> pam (1.0.1-1ubuntu1) intrepid; urgency=low
[00:04] <kirkland> slangasek: okay, bzr branch has the docs ;-)
[00:05] <kirkland> slangasek: gimme a few minutes, i'll peruse this....
[00:06] <slangasek> yep, your apt mirror is at least 4 hours out of date ;)
[00:07] <kirkland> slangasek: :-)  thankfully, my hotel wifi pull the bzr mirror, though, initially estimates pointed at 4 hours to branch that
[00:08] <slangasek> kirkland: heh, ouch
[00:17] <kirkland> slangasek: any guidelines on the priority front?
[00:18] <slangasek> kirkland: I threw some together in the wiki (https://wiki.ubuntu.com/PAMConfigFrameworkSpec)
[00:18]  * kirkland subscribes to changes to that page, albeit after-the-fact
[00:19] <slangasek> kirkland: but I think this hooks in as an 'optional' module, so I'm not sure the priorities matter as much?
[00:19] <kirkland> slangasek: agreed, optional in all points in the pam stack
[00:20] <kirkland> slangasek: would Priority 0 be acceptable?
[00:21] <slangasek> kirkland: yes, that's what I'm currently using for smbpasswd
[00:22] <kirkland> slangasek: http://pastebin.ubuntu.com/39236/
[00:24] <slangasek> kirkland: s/^\s+(auth|session|password)//;
[00:24] <slangasek> (that's the recent syntax change I made during the implementation)
[00:26] <slangasek> kirkland: oh, and s/Initial/Final/
[00:26] <slangasek> kirkland: because if there's no "Initial" stanza defined it'll fall back to "Final", but not the other way around
[00:27] <kirkland> slangasek: http://pastebin.ubuntu.com/39237/
[00:27] <kirkland> slangasek: nice, (auth|session|password) were redundant
[00:28] <slangasek> precisely :-)
[00:28] <slangasek> kirkland: looks good to me... throw it in there, call pam-auth-update --package in your postinst, and test it? :)
[00:29] <kirkland> slangasek: suggestions on where to install the file?  /usr/share/foo ?
[00:30] <slangasek> kirkland: it has to be installed in /usr/share/pam-configs/
[00:30] <kirkland> slangasek: k
[00:30] <kirkland> slangasek: i see it in the wiki now....
[00:34] <kirkland> slangasek: "Package config declarations are shipped in the packages under /usr/share, and are not user-editable" ... should I install it 444?
[00:34] <slangasek> kirkland: 0644, the default for all files
[00:35] <slangasek> ("not user editable" meaning only that user changes will not be preserved or honored)
[00:35] <kirkland> slangasek: right, but that's chmod-able in postinst, no?
[00:35] <kirkland> k
[00:35] <slangasek> there's no need to chmod it at all
[00:43] <slangasek> can't win
[00:43] <kirkland> slangasek: no manpage for pam-auth-update yet?
[00:43] <slangasek> close enough for now :P
[00:43] <slangasek> kirkland: correct; I wanted to get this into your hands (& into the archive), documentation is next along with a couple of missing features (like, being able to deconfigure a module in prerm remove)
[00:43] <kirkland> slangasek: i don't see a usage statement in the wiki either
[00:44] <kirkland> slangasek: no worries...
[00:44] <slangasek> pam-auth-update --package is all you should ever call from a postinst
[00:44] <kirkland> slangasek: http://pastebin.ubuntu.com/39240/
[00:44] <kirkland> slangasek: oh, maybe i should drop the ecryptfs-utils then
[00:46] <kirkland> slangasek: and the file installed to /usr/share/pam_configs/ should just be the package name?  ie, ecryptfs-utils?
[00:46] <slangasek> kirkland: a simple, descriptive filename; package name is fine
[00:47] <slangasek> or simply "ecryptfs", even
[00:50] <kirkland> slangasek: which pam version should I depend on in the control file?
[00:50] <slangasek> kirkland: libpam-runtime (>= 1.0.1-2ubuntu1)
[00:51] <kirkland> slangasek: i suppose "Recommends" would be most appropriate
[00:52] <slangasek> why not a simple Depends?
[00:52] <slangasek> if you do a 'Recommends', then you have to make the use in the postinst conditional
[00:52] <kirkland> slangasek: okay, f that
[00:52] <kirkland> :-)
[01:02] <kirkland> slangasek: error message... pam-auth-update: Run pam-auth-config --force to override.
[01:02] <kirkland> slangasek: looks like you change names of the bin at some point?
[01:08] <slangasek> kirkland: no, I just fail at looking at the name of the file I'm editing
[01:08] <slangasek> I'll fix that up, thanks :)
[01:08] <ion_> $0
[01:08] <kirkland> slangasek: np.... okay i'm close
[01:09] <kirkland> slangasek: i got added to common-auth and common-session fine, but not common-password for some reason
[01:10] <slangasek> kirkland: s/addit/Addit/
[01:10] <slangasek> case-sensitive field names, sorry
[01:10] <kirkland> slangasek: aha
[01:10] <kirkland> slangasek: got it
[01:11] <kirkland> slangasek: okay, now for the sensitive part....
[01:11] <kirkland> slangasek: i've been telling users all along to use auth-client-config to setup pam manually
[01:11] <kirkland> slangasek: i see your utility is smart enough to detect outside edits of pam.d/*
[01:11] <kirkland> slangasek: i suppose in the postinst i should tear down the auth-client-config configuration, if present
[01:12] <kirkland> slangasek: to properly support upgrades?
[01:12] <slangasek> kirkland: by then it's too late, I think; libpam-runtime will already be installed, and will have already decided to go into manual mode
[01:12] <slangasek> kirkland: I think this is a "you were running a devel release, fix it up by hand" case
[01:12] <kirkland> slangasek: i can see that....
[01:13] <slangasek> noting that "fix it up by hand" probably means nothing more than "pam-auth-update --force"
[01:14] <kirkland> slangasek: shall i rip out the auth-client-config in the same patch as the one where I'm adding this?
[01:14] <kirkland> slangasek: that seems logic to me...
[01:14] <slangasek> I have no opinion on that; maybe ask jdst<tab><tab> ok, he's not here :)
[01:17] <lifeless> asac: well, I'd implement a magic mark filter for that
[01:17] <lifeless> the basic theory graydon notes is accurate
[01:17] <lifeless> though actually we can do better with N-parent diffs
[01:22] <kirkland> slangasek: okay, i have a tested working ecryptfs-utils package
[01:22] <slangasek> w00t
[01:23] <kirkland> slangasek: however, i would like to provide a better upgrade vector for my bleeding edge testers
[01:23] <slangasek> ok
[01:23] <kirkland> slangasek: perhaps a dpkg-reconfigure libpam-runtime?
[01:24] <slangasek> insufficient; libpam-runtime isn't going to have enough smarts to know whether it's safe to overwrite this stuff, so at best you get a debconf prompt asking to overwrite (and in any case, it's not safe to call dpkg-reconfigure from a maintainer script
[01:24] <slangasek> )
[01:25] <kirkland> hrm
[01:25] <slangasek> but what you can do is do the same thing as libpam-runtime's postinst: have a version check, and an md5sum check, and if you recognize the files as being your handiwork (via auth-client-config), call pam-auth-update --package --force
[01:25] <slangasek> y'know, I'm having the hardest time remembering that the program is named pam-auth-update instead of pam-auth-config.  Maybe I named it wrong and should change it?
[01:26] <kirkland> :-)
[01:26] <kirkland> slangasek: you've typed it that way a few times
[01:26] <kirkland> slangasek: i cut-and-pasted it into a shell at some point and it didn't worky
[01:26] <slangasek> :)
[01:27] <kirkland> :-)
[01:28] <kirkland> slangasek: i think i can actually issue an auth-client-config command to undo its changes
[01:28] <kirkland> slangasek: would that help?
[01:28] <slangasek> nope
[01:28] <slangasek> :)
[01:28] <kirkland> ah, it's too late
[01:28] <slangasek> something still has to call pam-auth-update --force
[01:28] <kirkland> as you've said
[01:28] <kirkland> okay
[01:30] <kirkland> slangasek: okay, i see the libpam-runtime.postinst, i'll use something like that
[01:30] <slangasek> I guess you probably only have one md5sum for each file that you need to worry about, so I would suggest including them as variables in the maintainer script instead of having to ship files for it
[01:54] <cjwatson> argh, testing cdebconf changes => officially a pain
[01:54] <lifeless> :>
[01:54] <ion_> cjwatson: Why is that?
[01:55] <slangasek> because you can't install cdebconf on a full system
[01:55] <ion_> slangasek: chroot
[01:55] <slangasek> due to conflicts with debconf (though we should be getting closer to the end of that transition)
[01:55] <ion_> All that can be done by a test suite.
[01:55] <cjwatson> yes yes yes stop preaching at me :)
[01:56] <cjwatson> I'm trying to test something exceedingly specific and need to do a proper system test in an Ubuntu CD environment
[01:56] <cjwatson> also I don't need solutions since I've figured it out
[01:56] <cjwatson> I should have known that saying "test" would bring out the helpful crew
[01:56] <lifeless> I have not claimed helpful  :)
[01:57] <slangasek> https://launchpad.net/~the-helpful-crew/+members
[01:57] <StevenK> And Launchpad is offline
[01:57] <slangasek> ha-ha, made you look
[01:57] <StevenK> :-P
[01:58] <cjwatson> ion_: short answer, if you don't know what I'm trying to do, don't suggest a chroot ;-)
[01:59] <cjwatson> d-i is not amenable to that approach
[01:59] <LaserJock> hmm, for some reason the Edubuntu CD seems to not be working anymore
[01:59] <slangasek> heh, that too
[01:59] <slangasek> you /used/ to be able to chroot d-i...
[01:59] <cjwatson> I wouldn't like to attempt to install as far as tasksel that way
[02:00] <slangasek> ...right :)
[02:00] <kirkland> slangasek: something like http://pastebin.ubuntu.com/39252/ ?
[02:00] <cjwatson> (the bug I'm working on is related to the combination of preseeding and debconf -> cdebconf passthrough)
[02:01] <cjwatson> $ dpkg --compare-versions "" lt "53-1ubuntu6"; echo $?
[02:01] <cjwatson> 0
[02:01] <cjwatson> kirkland: you can leave out the [ -z "$2" ] test
[02:01] <kirkland> cjwatson: thx, copied from slangasek ;-)
[02:01]  * cjwatson <- nitpicking maintainer scripts since 2000
 I'll just pretend that Sam wrote that part of the postinst originally
[02:03] <slangasek> anyway, sometimes it's nice to make it explicit for the benefit of readers :)
[02:04] <slangasek> kirkland: actually, in your case I think you explicitly *don't* want to do this on first install of the package, so better to omit the [ -z ] test and use lt-nl as the operator...
[02:04] <kirkland> slangasek: i thought i might have an off-by-one there....
[02:06] <slangasek> kirkland: lt-nl != le, if that's what you mean
[02:06] <slangasek> -nl is "null versions compare greater than everything instead of less than everything"
[02:06] <kirkland> slangasek: not what i meant
[02:07] <kirkland> slangasek: just that 53-1ubuntu6 didn't seem like quite the right operand
[02:08] <slangasek> kirkland: looks right to me
[02:08] <cjwatson> I think I prefer "No version is Later than any version" as the expansion of -nl, myself; "greater than" isn't so mnemonic
[02:09] <kirkland> slangasek: okay, i'm fading fast...  i will test this more thoroughly asap tomorrow
[02:09] <kirkland> slangasek: upgrade some of my vm's in various conflagurations
[02:09] <kirkland> slangasek: before posting the patch for sponsorship
[02:09] <kirkland> slangasek: but i'm liking what i'm seeing out of your utility!
[02:09] <cjwatson> oh, this is so wrong, I'm using kickstart as a way to get in ahead of cdebconf starting
[02:10] <kirkland> slangasek: i just don't want to smack the people who adopted this early, if you know what i mean
[02:10] <slangasek> kirkland: sure, I understand :)
[02:11] <slangasek> kirkland: fwiw, I normally use lt $fixed version, rather than le $last_known_good, because I figure repeating the operation against uncoordinated ~ versions is the safer course
[02:11] <slangasek> er, $last_known_bad
[02:11] <kirkland> slangasek: i'll try to find a Euro sponsor for the patch tomorrow; if not, i'll be waiting for your review
[02:11] <cjwatson> (if you pass 'ks=http://foo.invalid/bar' to the installer, it will bring up the network and then stop with an error message before starting main-menu)
[02:11] <slangasek> kirkland: good-o, cheers :)
[02:12] <kirkland> night all
[02:12] <slangasek> 'night!
[02:21] <slangasek> kirkland: one thing that's missing here is to check that /etc/pam.d/common-account is also pristine before deciding to overwrite; your package hasn't edited it, so it should just be a matter of using the checksum from libpam-runtime, but we should check that all four files are unmodified before taking an action that will clobber them all
[02:32] <slangasek> kirkland: how I would write this: http://pastebin.ubuntu.com/39257/
[02:32] <ScottK> slangasek: If you have a moment, would you please do the wireshark 1.0.2 backport for Hardy.  I'm trying to administer a "Please use the system" lesson to my local LUG and it'd be handy if that magically appeared soonish.
[02:32] <ScottK> It's Bug #259876
[02:32] <slangasek> ScottK: heh :)
[02:33] <slangasek> I guess that means LP is back up, though
[02:33] <ScottK> Was it down?  I turned off the network and went out in the backyard to work without distraction for the last several hours.
[02:33] <ScottK> (with my laptop)
[02:33] <slangasek> yeah, it's still down if you ask the /topic
[02:33] <StevenK> ScottK: Which was when LP was down
[02:34] <ScottK> Right, of course.
[02:34] <StevenK> It's 1:34UTC, they're early
[02:34] <StevenK> "We still have 30 minutes! Bring it back down!"
[02:35] <ScottK> Well I've looked at one page and didn't claw my eyes out, so that's good news.
[02:36] <slangasek> ScottK: magic moofled
[02:36] <ScottK> Thanks.
[02:38] <spm> slangasek: sorry :-)
[02:39] <slangasek> spm: no worries :)
[06:23] <dholbach> good morning
[06:25] <Hobbsee> hey dholbach!
[06:25] <dholbach> hi Hobbsee
[07:03] <StevenK> slangasek: I guess "Cannot make/remove an entry for specified session" iz the pam bug?
[07:12] <slangasek> StevenK: yes
[07:13] <StevenK> slangasek: Yes, I actually read the bug between asking and you replying. :-)
[07:13] <slangasek> well, hmph
[07:13] <StevenK> Heh, sorry :-)
[07:15] <slangasek> roughly, that was also a "yes; do you need help getting anything un-broken?"
[07:15] <StevenK> Nothing OMG kittens urgent
[07:16] <StevenK> Just need to update a mobile image chroot/squashfs
[07:16] <StevenK> slangasek: Has the fixed package been published?
[07:16] <slangasek> yes, hours ago
[07:17] <StevenK> Ah, my local mirror needs to be booted
[07:26] <slangasek> huh, when did mailman start sorting admindb entries by sender first name instead of by date?
[07:33] <TheMuso> 7~/me has had a mental blank... .la files do go in dev packages right?
[07:33] <slangasek> .la files go to /dev/null ;)
[07:33] <slangasek> yes, they go in dev packages or nowhere at all, depending on your perspective
[07:35] <TheMuso> slangasek: heh right, thanks.
[07:43] <slangasek> kees: so with the new framework, all we have to do to change the default password hash from md5 to something else is to edit debian/pam-configs/unix. <mad giggle>
[07:50] <asac> lifeless: hmm. does that mean that a solution is already available?
[07:56] <lifeless> asac: no more than that thread talks about
[07:56] <lifeless> though you can do a merge and diff yourself
[07:56] <lifeless> and I have some plans that may help eventually
[07:58]  * TheMuso grumbles at libtool changes breaking another package...
[07:59] <TheMuso> And in a non-obvious way as well. The package builds, but no .so files are built.
[08:19] <slangasek> hrm; did something regress on bazaar.lp in the latest launchpad code rollout?:
[08:19] <slangasek> Using saved location: bzr+ssh://vorlon@bazaar.launchpad.net/~ubuntu-core-dev/pam/ubuntu/
[08:19] <slangasek> Server is too old for streaming pull, reconnecting.  (Upgrade the server to Bazaar 1.2 to avoid this)
[08:19] <StevenK> I thought the server was 1.5 ?
[08:20] <StevenK> slangasek: Beg in #launchpad?
[08:20] <\sh> slangasek: see #launchpad
[08:28] <RAOF> StevenK: The server is 1.6rc3, actually :)
[08:28] <StevenK> RAOF: Not according to that error :-P
[08:29] <RAOF> Right; apparently it's got to do with something that bzr 1.5 uses that 1.6 doesn't support.
[08:30] <slangasek> mmk
[08:36]  * persia suspects the error is actually " Server is too new for streaming pull, reconnecting."
[10:10] <amitk> soren: kvm doesn't like being shutdown on intrepid....
[10:13] <soren> amitk: -v
[10:15] <amitk> soren: bug 259336
[10:16] <amitk> bug 259336
[10:16] <soren> could you try with kvm 72, please?
[10:19] <amitk> sure
[10:22] <Keybuk> ooh, I didn't just get quoted in LWN
[10:22] <Keybuk> I made "Quotes of the week" in the LWN Kernel page
[10:22]  * Keybuk is FAMOUS!
[10:28] <soren> Keybuk: Ooh, top spot, no less :)
[10:30] <Treenaks> Keybuk: woo!
[10:35] <emgent> moin people
[10:35] <stefanlsd> hihi
[10:35] <emgent> hey stefan! :)
[11:06] <persia> cjwatson: Could you subscribe me also to the CD image reports for alternate CDs for the ports?  I get ports_daily-live, but not ports_daily.
[11:29] <cjwatson> persia: the reason you aren't getting mails is that the builds are succeeding, not that you aren't subscribed
[11:31] <persia> cjwatson: Ah, I only saw hppa and lpia on cdimages.ubuntu.com, and so thought that I should be getting errors for ia64, sparc, and powerpc.
[11:31] <persia> cjwatson: On another note, do you have any plans for further updates to the ArchiveReorganisation spec, or do you think it's ready for presentation to the TB for review/approval?
[11:33] <cjwatson> persia: oh, I think it only sends out a mail if the entire build fails
[11:33] <cjwatson> that's probably a bug
[11:33] <cjwatson> http://people.ubuntu.com/~ubuntu-archive/cd-build-logs/ubuntu/intrepid/ports_daily-20080821.log
[11:33] <cjwatson> persia: I want to run it by Mark this week while I'm in the office; I already gave him a heads-up on it
[11:34] <persia> cjwatson: OK.  I asked about it at the last TB meeting, and was advised that the idea was accepted, but that the specific proposal had not yet been submitted for review.  Please let me know if I may be of any assistance in helping the process along.
[11:35] <cjwatson> ok
[11:35] <cjwatson> thanks
[11:51] <\sh> hmm...python2.4 will still be available in intrepid, right?
[11:55] <\sh> Riddell: please apply http://archive.linux-server.org/debdiffs/python-qt4_4.4.3-1ubuntu1.debdiff and reupload to intrepid, thx
[11:55] <\sh> argl...FC
[12:15] <soren> IIUIC, it's ok for a GPL project to copy code from an LGPL project. Is that right?
[12:16] <cjwatson> yes
[12:17] <soren> But not the other way around?
[12:17] <cjwatson> remember that the code retains its original licence
[12:17] <soren> Oh, it does?
[12:17] <cjwatson> an LGPL project can import GPL code if it likes - it's just that the work as a whole may now only be released under the terms of the GPL
[12:17] <soren> I thought the LGPL said that you were free to change the license to GPL.
[12:18] <cjwatson> actually, yes, that is true in the LGPL case
[12:18] <soren> Ok..
[12:18] <cjwatson> so indeed it would just become pure GPL then
[12:18] <soren> So if a GPL project takes code from an LGPL project and improves upon it, the LGPL project can't use those improvements. Correct? (without changing their own license, of course)
[12:18] <cjwatson> that's correct
[12:19] <cjwatson> unless the owners of the GPL code relicense it back
[12:19] <soren> Figures..
[12:19] <soren> cjwatson: Ok, thanks for clearing that up.
[12:22] <cjwatson> soren: ok for me to merge your server seed stuff now?
[12:22] <soren> cjwatson: Oh, by all means.
[12:23] <cjwatson> I merged my tasksel branch and fixed a cdebconf bug that was causing problems there
[12:23] <cjwatson> server seed merged
[12:23] <cjwatson> I'll get the debian-cd stuff done shortly
[12:23] <vargadanis> hey there... I want to build a GUI with a different look and feel from the OS's default using either GTK (preferred) or QT...
[12:23] <vargadanis> what are my options?
[12:24] <cjwatson> http://developer.gnome.org/arch/gtk/themes.html
[12:25] <cjwatson> links to art.gnome.org which has a good deal of tutorial information
[12:26] <vargadanis> cjwatson: so you say that with this I can set the theme of my application without modifying the theme of the whole desktop?
[12:27] <cjwatson> you can do that sort of thing with local resource files, yes
[12:27] <cjwatson> or local styles
[12:27] <soren> You can apply a different theme to your application. For an example, see rawstudio.
[12:28]  * cjwatson <- not an expert
[12:29] <vargadanis> cjwatson: you are more of an expert in the issue than me ^_^
[12:29] <vargadanis> soren: I will do that, thanx for the tip
[12:29] <soren> vargadanis: np
[12:32] <vargadanis> soren: do you think that I can even let's say create custom widgets and say set the background image for them just by these local rc files?
[12:32] <soren> I imagine so. I've never used it myself.
[12:33] <vargadanis> soren: alright... I will look into that
[12:40] <Riddell> mvo: synaptic recommends apt-xapian-index, do you plan to get that into main?
[12:40] <mvo> Riddell: yes I have not written a MIR for it yet though
[12:42] <Riddell> mvo: ok, I need one for adept so I'll write one
[12:44] <mvo> aha, excellent
[12:51] <persia> ogra: About denemo.  The last I heard from LaserJock was that it had three chunks of stuff that needed MIR.
[12:51] <persia> Those being csound, lilypond, and libaubio
[12:52] <ogra> right
[12:52] <persia> I think csound is best dropped to suggests, as it's an optional audio rendering system, and timidity can also be used to render.
[12:52] <ogra> i asked pitti about lilypond and he didnt find any harm in MIRing it ... i didnt know about the others back then
[12:52] <persia> denemo doesn't make any sense without lilypond, so that, and the other 4 sources that go with it ought drop in.
[12:53] <ogra> as i mentioned before, i dont really need the midi functionallity in edubuntu but want the notesetting capabilities it offers in edubuntu
[12:53] <ogra> sure it makes sense if you want to write down notes and prnt them out in music class
[12:54] <persia> THe issue is really with libaubio.  It depends on JACK.  JACK depends on freebob.
[12:54] <ogra> we never had any midi stuff in edubuntu
[12:54] <ogra> but denemo was always there
[12:54] <persia> if libaubio doesn't depend on JACK, then denemo and lilypond become less interesting for Studio, which is bad.
[12:55] <persia> If JACK doesn't depend on freebob, we drop support for all firewire audio interfaces, which is bad.
[12:55] <persia> freebob is dead upstream.  There is a replacement ffado project, but it has yet to reach it's first release.
[12:55] <ogra> what happened to the plan ubuntustudio had to snaitize jack for main ?
[12:55] <persia> It's waiting on a ffado release.
[12:55] <ogra> will that happen in time for intrepid ?
[12:56] <persia> No.
[12:56] <ogra> :(
[12:57] <persia> ogra: That said, while I'm not comfortable with a dead upstream in main, someone else might want to maintain it.
[12:57] <persia> Alternately, if you look at the feedback to LaserJock's mail to the edubuntu-users mailing list, it seems that at least a couple users would rather install Studio directly for teaching music.
[12:58] <ogra> well, but it was part of edubuntu the last years ... i'm not really comfortable to drop the feature
[12:58] <ogra> and it never did any midi stuff without installing the universe components either
[12:58] <persia> ogra: Makes sense.  Let's wait to hear more from LaserJock, who has been following this more closely.
[12:58] <ogra> right
[13:04] <JontheEchidna> mvo: Did you get my email about software-properties-kde ?
[13:11] <mvo> JontheEchidna: yes, thanks a lot! I haven't looked at the merge yet, sorry. I will do it today
[13:11] <JontheEchidna> mvo: Ok, I don't want to  be pushy or anything, but with the feature freeze coming up I got a bit nervous...
[13:11] <JontheEchidna> thanks :)
[13:11] <mvo> thank you :)
[13:33] <hwilde> fyi system has been stable for 14 hours since removing console-kit-daemon  (244218)
[14:06] <kirkland> cjwatson: what's the magic you're putting on the end of your kvm command to get the keyboard mapping back to a sane state?
[14:08] <soren> -k us, probably.
[14:09] <soren> kirkland: Good news on that front, though. It seems we have a fix that works, that will land withing the next couple of days.
[14:09] <kirkland> soren: rock on
[14:09] <soren> kirkland: aliguori strikes again :)
[14:09] <kirkland> \o/
[14:10] <nxvl> good morning
[14:10] <kirkland> nxvl: howdy
[14:12] <nxvl> waiting for some news
[14:12] <nxvl> :D
[14:15] <cjwatson> kirkland: -k en-us
[14:15] <kirkland> cjwatson: cool, thanks
[14:15] <cjwatson> it doesn't quite work, but it at least fixes arrow keys
[14:16] <kirkland> cjwatson: yeah, partitioning in the installer is excruciatingly difficult without arrow keys :-)
[14:18] <LaserJock> I noticed yesterday that in VMware Player my up arrow takes a screenshot :-)
[14:33] <LaserJock> ogra, persia: would it make sense to MIR freebob with the expectation that it'll be replaced ASAP?
[14:34] <persia> LaserJock: As long as someone is willing to maintain freebob, I'm happy with that.  I don't have that kind of HW, so don't feel comfortable doing it myself.
[14:34] <LaserJock> hmm, I don't either
[14:35] <LaserJock> I tried to look in the denemo ./configure for an option to turn it off
[14:36] <LaserJock> there wasn't any obvious flag
[14:36] <persia> See, I'd be opposed to turning it off.  While mScore is gaining mindshare, denemo is still the most popular score editor for ubuntustudio
[14:37] <ogra> oh
[14:37] <ogra> why did obody poit me to mscore yet
[14:37] <ogra> *nobody
[14:37] <persia> I didn't know you wanted it.  It's nicer than denemo, but it kinda needs a real soundfont, which makes it a little awkward if you want it on a CD.
[14:37] <ogra> it doesnt even have any deps
[14:37]  * sistpoty|work hands ogra another n
[14:37] <ogra> (apart from QT)
[14:37] <LaserJock> ogra: because I checked the number of MIRs I'd have to do to get it
[14:38] <ogra> persia, well, my words should have told you
[14:38] <ogra> all i want in edubuntu is a score editor for note setting
[14:38] <persia> ogra: mScore needs JACK too, and jack-in-main is the primary blocker here.
[14:38] <ogra> upstream doesnt talk about any midi deps
[14:39] <ogra> http://mscore.sourceforge.net/en/download.php says alsa, cmake, freetype2 and qt4
[14:39] <persia> It doesn't have any MIDI deps, it has a built-in soundfont player.  It doesn't need a MIDI interpreter (like denemo), but it does need sounds.
[14:40] <persia> fluid-soundfont is the only free GM/GS soundfont anyone has been able to find, but it's *huge*
[14:40]  * ogra likes huge 
[14:40] <ogra> size does matter, you know ;)
[14:40] <ogra> we have ~400M free space on the edubuntu CD
[14:41] <ogra> space os my smallest concern
[14:41] <ogra> *is
[14:41] <LaserJock> more like ~188MB free at the moment
[14:42] <persia> Oh, if you've that much space, mScore is *way* better than denemo, and mScore+MuSE is most of a complete environment ffor iddling about.
[14:42] <persia> Still needs jack-in-main though, which is about freebob, and not about anything else.
[14:43] <ogra> LaserJock, minus WINFOSS
[14:43] <ogra> which is about 200M
[14:43] <LaserJock> well, we can trade lilypond for a few more MIRs for mscore
[14:43] <ogra> i see libjack in the deps though
[14:44] <persia> Right, it's all about freebob.  If someone is willing to support freebob-in-main, we're golden.  If not, we're sunk either way.
[14:44] <LaserJock> so.... we need find a victim/volunteer
[14:45] <persia> And I'll maintain, for the same reasons I did pre-Sevilla, that dropping the freebob dep isn't the right way to solve it, as there are too many users who depend on this.
[14:45] <ogra> pfft ... users ... who cares about *them*
[14:46] <ogra> as long as the devs have fun :P
[14:46] <persia> ogra: Well, at least I'll stand up for people with fancy firewire audio interfaces that make me drool.  If I'm lucky, one of them will demo one for me one day :)
[14:48]  * ogra used to live in an artists colony some years ago ... i would have had 100s of midi devices to test back then (there were also a lot of musicians... it had the biggest rehearsal center in northern germany attached ... (people like mousse-t come from there)) sadly i lost all contacts to these people
[14:48] <LaserJock> well, we're looking at removing note-editing altogether here, it'd be nice to come up with something
[14:49]  * persia checks the freebob mailing list
[14:50] <soren> "freebob"? A free implementatino of Microsoft Bob?
[14:50] <Treenaks> soren: ooh! that should fit well in the new PAM framework!
[14:51] <persia> FreeBoB from BeBoB, which is Bridgeco Enhanced BreakOut Box
[14:51] <soren> Treenaks: That's a very disturbing though.
[14:51] <soren> t
[14:55] <persia> ogra: LaserJock: I'm only seeing spam on the FreeBob lists.  There is a beta ffado out (http://www.ffado.org/?q=release/beta) which might be usable.  Some good reports of it working on Ubuntu in the mailing lists.
[14:56] <persia> Doesn't work with ALSA or pulseaudio yet, but that likely doesn't matter because most users wouldn't expect to use it.
[14:56] <persia> Err, to use the ffado backend.
[14:57] <ogra> if users dont expect it, why would we depend on it ?
[14:58] <persia> ogra: The vast majority of users won't have hardware that requires freebob/ffado.
[14:58] <ogra> then it should become a suggests
[14:58] <persia> From the perspective of having a note editor in Edubuntu, I don't think it matters that these devices don't work well with ALSA or pulse yet.
[14:59] <persia> It can't be a suggests: JACK needs to link against the libraries, and the note editors need to be linked against JACK to use JACK.
[14:59] <persia> I use JACK a lot, but I don't use those backends, because I don't have that hardware.
[15:00] <persia> The contention is that most Edubuntu users won't mind not being able to use firewire audio interfaces with ASLA or pulse, but want a note editor.
[15:02] <ogra> right
[15:03] <cr3> what does this file mean on archive.u.c/ubuntu: Archive-Update-in-Progress-leningradskaya.canonical.com
[15:04] <LaserJock> persia: would we be more likely to find somebody to maintain ffado than freebob?
[15:05] <LaserJock> seems more of a "who's got the hardware" issue
[15:05] <persia> LaserJock: Given the ffado beta, that's what I'm thinking.
[15:06] <persia> I could get some of the lower-end HW that it supports this weekend, but I'm just not confident that I understand the subsystems well enough to be more than a tester.
[15:14] <asac> kees: can you give me your /e/n/interfaces ?
[15:15] <asac> kees: you appear to have a rather complex setup. would like to take a look at that ;)
[15:16]  * cjwatson does a snoopy dance
[15:16] <cjwatson> successful 256MB installation
[15:16] <cjwatson> (desktop CD)
[15:16] <persia> Wow!  Very nice!
[15:16] <ogra> YAY !!
[15:17] <persia> This is compcache showing it's colours?
[15:17] <ogra> now try with 75% compcache and 128M :)
[15:17] <cjwatson> I may do in a bit :)
[15:17] <ogra> though users wont have much fun with such a system once installed :)
[15:18] <ogra> unless we'd carry over compcache into the install
[15:19] <cjwatson> 256MB was the target I wanted to hit, so I'm happy enough; 'free' immediately after booting says that about 128MB is used, so there's decent headroom there
[15:21] <ogra> well, it might be intresting for UME once we get to arm support ... i think there are many 128M arm devices out there
[15:22] <ogra> but thats not on the plate yet
[15:22] <cjwatson> arm> I agree
[15:23] <persia> The Zaurus is an example, and someone got Ubuntu Desktop on one
[15:23] <ogra> right, compcache might make that more comfortable ...
[15:24] <ogra> there are a lot of opportunities ... but all rather intrepid+1
[15:24] <persia> Indeed.  One week isn't enough to play enough.
[15:26] <ogra> well, the feture is in since a week ...
[15:26] <ogra> *feature
[15:27] <ogra> it always depends if you call the enhancement of it another feature :)
[15:27] <ogra> its there for all arches in intramfs ... just a matter of setting the COMPCACHE_SIZE variable in initramfs.conf
[15:28] <ogra> so if we should get arm before final release we can indeed make use of it :)
[15:33] <cr3> does the live cd create anything at all on the installed system, perhaps a swap file for example?
[15:33] <ogra> ubiquity creates a swap partition
[15:34] <cr3> ogra: how can it find that space if the installed system takes all the space on the drive?
[15:34] <ogra> well, it will check if you have enough space and moan afaik
[15:35] <ogra> if you don have the 3.5G you need for the install then adding a swapfile will still eat diskspace
[15:35] <ogra> the only option you would have here would be to go completely without swap
[15:36] <cr3> I meant: what if the filesystem, such as ntfs for example, takes up the whole drive? I suspect ubiquity doesn't do some resizing on the fly, does it?
[15:36] <ogra> it does offer it
[15:36] <ogra> not sure about "on the fly" i dont have ntfs here anywhere so i could test
[15:37] <cr3> ogra: oh wait, you're talking about the installation process, I'm just talking about the live environment
[15:37] <ogra> ah
[15:37] <cr3> heh, that was confusing :)
[15:37] <ogra> well, for that we have compcache now
[15:37] <ogra> which creates a virtual swapdisk in ram
[15:38] <cr3> so everything is in memory? that's what I thought but I suspected more space might be needed to run the live environment
[15:39] <ogra> th elivenev needs 128-256M ... but if you have i.e. a graphics card using shared ram that will eat from your main ram ... which is why our requirements up to ow were 384M
[15:39] <ogra> *up to now
[15:39] <ogra> compcache creates a virtial swap device in ram where read and write operations get compressed
[15:39] <ogra> *virtual
[15:40] <ogra> so you can effectively have 256M plus 64M virtual swap
[15:40] <ogra> which compensates the videoram or whatever takes away from the physical ram
[15:41] <cr3> ogra: cool! thanks for those extra details, much appreciated
[15:41] <ogra> as i said above compcache is avaliable generally ... you can use it in other environments as well
[15:41] <ogra> (ltsp will use it for example)
[15:42]  * cr3 puts compcache on his toread list
[15:42] <ogra> just set COMPCACHE_SIZE (see initramfs.conf for documentation) ... and send flowers to ion :)
[15:44] <cr3> we need telegram hugs
[15:49] <kirkland> slangasek: ping me when you come around....
[15:53]  * LaserJock wonders what jsilber would say if Canonical employees started sending hug-o-grams
[16:15] <cjwatson> persia: I think I've fixed up the powerpc and sparc CD builds, BTW, since you mentioned the ports_daily problems
[16:17] <persia> cjwatson: Cool.  We'll have more arches tomorrow then.  ia64 is still broken?
[16:17] <cjwatson> persia: ENOKERNEL
[16:18] <cjwatson> https://edge.launchpad.net/ubuntu/+source/linux-ports/2.6.25-1.2/+build/622775
[16:18] <persia> heh.  That would do it.
[16:18] <cjwatson> doesn't actually look that complicated a failure, mind ...
[16:30] <kirkland> slangasek: Use of uninitialized value in numeric comparison (<=>) at /usr/sbin/pam-auth-update line 104.
[16:31] <Riddell> mvo: what creates /var/lib/update-notifier/dpkg-run-stamp ?
[16:37] <cody-somerville> cjwatson, You fixed PPC builds? :-]
[16:38] <ogra> cody-somerville, compcache went in on alpha4 release day (some hours to late for alpha4 :( )
[16:38] <soren> Riddell: /etc/apt/apt.conf.d/99update-notifier
[16:38] <cody-somerville> ogra, ah kay
[16:39] <ogra> cody-somerville, casper uses 25% compcache by default and switches compcache off on systems with more than 512M
[16:39] <Riddell> soren: thanks
[16:39] <cjwatson> cody-somerville: not the one you care about, sorry
[16:39] <cjwatson> a separate problem
[16:39] <cjwatson> sorry, I really do need to get to that but it's a pain
[16:39] <cody-somerville> cjwatson, :-(
[16:40] <soren> Riddell: :)
[16:40] <cody-somerville> cjwatson, Did you ever get a chance to read over that e-mail I sent you?
[16:41] <cjwatson> e-mail is a dead loss for me this week; I'm away and my home VPN is down ...
[16:41] <cjwatson> so I can't even check until Monday
[16:41] <cjwatson> (argh, etc.)
[16:41] <ion_> ogra: You spoke of moving the percentage etc. logic from the hook to a script. You began doing the change IIRC – did you finish it, and if not, is there a public branch i could pull from to do it, or perhaps something i could dget?
[16:42] <ogra> ion_, i found a more sane way and kept most of you stuff as is
[16:42] <Keybuk> cjwatson: I assume you mean Tuesday? :)
[16:42] <cjwatson> Keybuk: ooh, I forgot about the bank holiday
[16:43] <ogra> ion_, i didnt change much, just dropped the instane SKIP_COMPCACHE carp again, exported BOOT from init and checked for BOOT=casper and TotalMemory >512M
[16:44] <ogra> just look at initramfs-tools ... if your system is up to date you should have the recent code ... the way it works didnt change at all, its all as you know it
[16:45] <ion_> ogra: Alright
[16:49] <ion_> benc: Is there going to be a -386 kernel for intrepid in the near future, btw?
[16:50] <BenC> ion_: There already is one, but I think it's in universe (which is wrong)
[16:53] <ion_> benc: Ah, there seems to be an 2.6.25-1-386 indeed. Hadn’t noticed it, since linux-image-386 depends on 2.6.24something and i also have hardy entries in sources.list. Will there be one in sync with the -generic kernel version?
[16:55] <LaserJock> asac: around?
[16:58] <BenC> ion_: Not likely
[17:01] <MacSlow> Does Xorg or dbus set $XDG_SESSION_COOKIE?
[17:01] <ion_> benc: But there will be one for the final release of 8.10, right?
[17:02] <ion_> /etc/X11/Xsession.d/90consolekit:if [ -z "$XDG_SESSION_COOKIE" ] && [ -x "$CK_LAUNCH_SESSION" ]; then
[17:02] <lool> ion_: That's something else; it's CK integration
[17:02] <lool> MacSlow: I think it's gdm
[17:02] <ion_> Alright
[17:03] <lool> g_setenv ("XDG_SESSION_COOKIE", ck_session_cookie, TRUE);
[17:03] <lool> MacSlow: ^
[17:03] <lool> ion_: That might set it if you are using startx though
[17:04] <lool> So basically XDG_SESSION_COOKIE is set in consolekit sessions, either started by gdm or ck-launch-session as ion_ pointed out (startx/xinit)
[17:07] <asac> LaserJock: ?
[17:08] <ogra> lool, (nitpick) or by sshd if you log in with -X or -Y
[17:09] <ogra> :)
[17:10] <LaserJock> asac: I ran into bug #255839 and found that a simple mkdir /etc/NetworkManager/system-connections fixed it
[17:10] <LaserJock> asac: would that dir normally be created at install time?
[17:11] <lool> ogra: Oh really?  how is this done?
[17:11] <ogra> lool, cjwatson patched sshd to use CK for X sessions
[17:12] <lool> I recall seeing that, but I can't find it in the changelog anymore
[17:13] <asac> LaserJock: ok thanks. fix committed
[17:14] <asac> LaserJock: attached branch to bug accordingly
[17:15] <LaserJock> asac: awesome, for the first time NM actually works for all my network connections :-)
[17:16] <asac> LaserJock: tell that to those voices that dont want me to make network-manager conflict with ifupdown ;)
[17:16] <ogra> lool, openssh (1:4.7p1-4ubuntu1)
[17:16] <asac> LaserJock: what kind of connections did you have? do you still have your interfaces?
[17:17] <asac> (config file)
[17:18] <LaserJock> asac: I have a static IP eth0, a uni WPA-enterprise (which today killed NM) wifi, and a normal home wifi
[17:18] <asac> LaserJock: ok. did you have all configured in "interfaces"?
[17:18] <LaserJock> no, just the static IP
[17:19] <LaserJock> I was trying to let NM handle wifi and "interfaces" handle eth0
[17:19] <asac> LaserJock: how did you configure your dns? dns-nameserver stanza + resolvconf?
[17:19] <asac> (for the static IP thing=
[17:19] <LaserJock> in Hardy I did everything via the Gnome Network tool
[17:20] <LaserJock> for Intrepid I was manually editing /etc/resolv.conf on every reboot
[17:20] <asac> ok ... that means that all was ifupdown aka interfaces
[17:20] <LaserJock> yeah
[17:20] <asac> LaserJock: do you still have the config you had in hardy?
[17:20] <LaserJock> hmm, yes I do actually
[17:20] <LaserJock> I  made a backup of /etc before I blew it away
[17:22] <LaserJock> hmm, /etc/network/interfaces is empty except for lo in my Hardy config
[17:22] <asac> strange
[17:22] <asac> then you used NM for everything?
[17:22] <LaserJock> oh, I bet I know why
[17:22] <LaserJock> I managed the wifi vs static IP by using the profiles in the Network tool
[17:23] <asac> oh. never heard of such a feature
[17:23] <asac> so where does that store the config ... snippets?
[17:23] <LaserJock> if that info was stored somewhere in /etc I could maybe dig it out
[17:23] <LaserJock> that I don't know let me grep
[17:24] <LaserJock> hmm, doesn't seem to be in /etc
[17:24] <asac> hmm. is that tool gone in intrepid now?
[17:25] <asac> cant find it in menu
[17:25] <LaserJock> yes, it was removed
[17:25] <LaserJock> but you can install it
[17:25] <LaserJock> GNOME start menu applet
[17:25] <LaserJock> bah
[17:25] <LaserJock> gnome-main-menu is the package
[17:26] <LaserJock> man, bad morning
[17:26] <LaserJock> gnome-network-admin *is* the package
[17:26] <asac> whats the binary name
[17:26] <asac> let me look
[17:27] <asac> i dont have that on my system;)
[17:27] <asac> but i have that thing here in hardy
[17:27] <LaserJock> yes
[17:27] <asac> ok gnome-system-tools
[17:28] <LaserJock> right, it was taken out and put in its own package for Intrepid
[17:28] <LaserJock> as it's not installed by default
[17:28] <asac> ok thats saved in user profile
[17:28] <asac> .gnome2/network-admin-locations/test1234
[17:29] <LaserJock> awesome found it
[17:29] <asac> well ... probably that generates the interfaces on the fly
[17:29] <LaserJock> that'd be my guess
[17:29] <LaserJock> asac: why do you want the interfaces?
[17:30] <asac> i want more real-life test cases for the NM 0.7 system-config backend that tries to parse interfaces ;)
[17:31] <LaserJock> so having NM be able to deal with interfaces people have hacked on?
[17:32] <asac> yes ... more or less
[17:33] <LaserJock> right now it seems to write a perfect interfaces, for me anyway
[17:33] <asac> the idea is to ease transition for users that used other setups in the past
[17:33] <LaserJock> asac: well, I can make my own interfaces and see what NM does if you'd like :-)
[17:34] <asac> LaserJock: good. ill bring up the latest and let you know
[17:42] <BenC> ogra: are you using tbird from intrepid?
[17:42] <ogra> i'm using evo
[17:55] <superm1> BenC, thunderbird in intrepid shouldn't be too different than hardy's (according to https://edge.launchpad.net/ubuntu/+source/thunderbird )
[17:58] <mathiaz> slangasek: I'm trying to build the nssov overlay from openldap. But it seems there are issues with the includes - have you already tried to build one of the contrib modules for slapd ?
[17:59] <BenC> superm1: yeah, nm, I found my issue
[18:02] <superm1> BenC, if my memory serves me right, one of those extensions in that list i gave you had to be hand modified to work on the latest tbird.  just rev'ing the version number inside the jar was sufficient ,but the rest should be sane
[18:07] <asac> LaserJock: ok i uploaded new NM to ~network-manager PPA should be there in a few ... once that is done you can change /etc/NetworkManager/nm-system-settings.conf to --plugins=ifupdown,keyfile and sudo killall nm-system-settings
[18:10] <asac> LaserJock: to avoid confusion maybe remove "keyfile" for testing
[18:27] <LaserJock> asac: so before I kill nm-system-settings I'll want to have a "custom" interfaces in place or should I do that after?
[18:28] <asac> LaserJock: edit interfaces, then kill that settings thing
[18:28] <asac> LaserJock: maybe start with a simple config ;)
[18:29] <asac> auto eth0 dhcp ;)
[18:33] <slangasek> kirkland: pong; have you created a config that's missing a Priority: field?
[18:34] <slangasek> mathiaz: if you mean you're trying to build an out-of-tree overlay for openldap, no, I've never done that
[18:34] <mathiaz> slangasek: well - it's the nssov overlay located in contrib/
[18:34] <mathiaz> slangasek: but I've figured it out
[18:35] <slangasek> ok
[18:38] <LaserJock> asac: well, auto eth0 dhcp won't get me far, I don't have a dhcp server around
[18:46] <calc> so how is intrepid on intel graphics now, still having issues?
[18:46]  * calc is considering upgrading his machines, but might should wait until intrepid beta
[18:47] <LaserJock> calc: what kind of issues was it having?
[18:58] <lool> calc: I'm having issues on intel graphics, but didn't investigate
[18:58] <lool> Like refresh / sync issues, it's not clear; windows are "out of date"
[18:59] <slangasek> kirkland: ah, seen your bug now; don't call pam-auth-update in the postrm
[19:00] <slangasek> kirkland: I'm working on adding a --remove option for use in prerm remove
[19:02] <emgent> evening
[19:13] <calc> LaserJock: not sure i've just read people were having lots of problems on intrepid a while back and was wondering what the current state is before attempting to upgrade
[19:13] <LaserJock> calc: it's working quite good for me
[19:14] <LaserJock> calc: only X problems I've had were some Xchat unminimizing weirdnesses I think
[19:14] <calc> ok
[19:15] <LaserJock> calc: I'm personally finding Intrepid mostly better than Hardy currently
[19:16] <LaserJock> I've had a couple hard freezes but that's mostly it
[19:19]  * LaserJock stops thinking of all the little bugs in Intrepid before he changes calc's mind
[19:21] <bytor4232> Hey guys
[19:22] <bytor4232> I did an upgrade of Intrepid Alpha 5 last night, and nothing could log in.
[19:22] <bytor4232>  "cannot make/remove an entry for the specified session"
[19:22] <bytor4232> Had to wipe and reinstall from the alpha5 disc.
[19:22] <slangasek> bytor4232: first bit in the /topic
[19:22] <bytor4232> ah
[19:22] <bytor4232> sweet.
[19:23] <bytor4232> Thanks slangasek, I'll hang tight on upgrading.
[19:23] <slangasek> bytor4232: it's already fixed, upgrading now is fine
[19:28] <bytor4232> slangasek: Cool.  I'll run an upgrade tonight or tomorrow.
[19:30] <hwilde> LaserJock, little bugs are tasty.  just get some chocolate on there :)
[19:39]  * BenC finally has his email setup moved to tbird
[19:41] <bytor4232> BenC: I love tbird.  Its spam filtering is top notch once you start training it.
[19:43] <persia> calc: Do be careful it you've an odd input mechanism.  xinput hotplug doesn't quite autodetect everything cleanly yet.
[19:44] <hwilde> bytor4232, its a legit bayesian neural net :)
[19:47] <BenC> bytor4232: I never had any complaints about evo+bogofilter
[19:47] <BenC> as far as spam filtering at least
[19:50]  * bytor4232 gets a lot of spam.
[19:50] <bytor4232> support@ableteam.com, admin@ableteam.com are pretty big magnets.
[19:51] <JontheEchidna> gmail catches most of my spam
[19:51] <JontheEchidna> bogofilter gets most of the rest
[20:05] <okaratas> hello
[20:18] <BenC> is lrm-generic installed by default on live-cd's?
[20:20] <geser> is it normal that /etc/pam.d/common-{password,session} list pam_permit.so twice? wouldn't be using it once enough?
[20:26] <slangasek> geser: it's a stack padding thing
[20:26] <slangasek> geser: a more elegant solution is still percolating through my brain
[20:27] <slangasek> (the code that writes out the rest of the stack makes assumptions about how far to jump, you see)
[20:30] <bazaar> hi. is it possible to translate a single char with gettext?
[20:31] <bazaar> i mean really a char, not a one character char-array / string.
[20:31] <slangasek> if you mean the gettext C functions, then no
[20:31] <slangasek> string in, string out
[20:31] <lool> Why not convert them to a string?
[20:32] <bazaar> (yeah i mean C) so how can i accomplish to translate user input like 'y' or 'n' on a question like "DO you want to (y/n) ?"
[20:32] <bazaar> i mean at the point where comparison happens.
[20:33] <slangasek> well, you can always walk a character array
[20:33] <bazaar> if tolower(input_char) == _('y') . . yadayda
[20:33] <slangasek> to compare
[20:33] <persia> bazaar: If you're not worried about performance, check that the first character of the string matches and that tha string length is 1.
[20:33] <lool> _('Are you sure (%s/%s)?') % (_('y'), _('n'))
[20:33] <slangasek> lool: "", not '' :)
[20:33] <slangasek> bazaar: also, you can't use tolower() here safely for all locales
[20:34] <persia> slangasek: Won't it do an implicit cast in that case though?
[20:34] <lool> slangasek: Remember this is pseudo code, you didn't recognize any particular langage I hope  :-P
[20:34] <slangasek> bazaar: because a "character" won't necessarily fit in 8 bits
[20:34] <slangasek> persia: you mean the "" vs ''?  not in C...
[20:34] <lool> Ok, off to dinner and probably done for today; /me waves « "night »
[20:34] <slangasek> using '' around something that's not a single char will get you a compiler error
[20:34] <bazaar> persia: right tolower() won't work. i think i'll use first character + length test. thank you.
[20:35] <lool> persia: '' is character
[20:35] <lool> in C
[20:35] <persia> slangasek: Ah.  I must have looked at too much not-quite-C recently then.
[20:35] <slangasek> persia, bazaar: a string length of 1 also fails in such cases; you basically need to convert to wchar_t first if you want to compare a single char in a charset-independent fashion
[20:36] <slangasek> persia: yes, many not-C languages are nice to look at :)
[20:37] <bazaar> do i need to use all the "wide" functions (wprintf etc.) if i've localized my program? or will the printf() stuff work?
[20:38] <geser> aren't there also zero-length characters in unicode which are only used in combination with others?
[20:38] <slangasek> you wouldn't need to use widechars for everything; mostly, only when you're trying to do single-char comparisons
[20:38] <slangasek> (as opposed to single-byte comparisons)
[20:38] <slangasek> geser: <mumble>
[20:38] <slangasek> geser: if you start worrying about that, you have to link in a Unicode library that can do normalization for you :P
[20:39] <geser> bazaar: better yes, if you do formatted output as 5 chars != strlen of 5
[20:40] <bazaar> so if i want to print a message (w)printf("Hello everyone. This will be printed in your language!"); -- which one should i use?
[20:40] <slangasek> geser: I was limiting myself to "here's how you make this work for CJK users", not "here's how you make it cope with Macs"
[20:40] <bazaar> sry: (w)printf(_("Yada Yada"));
[20:41] <slangasek> bazaar: the output of _() is going to be of type char*, not of type wchar_t*; so you would always use printf() instead of wprintf(), unless you have some other reason you need to convert it first
[20:42] <bazaar> ah. i see. ok. thanks. guess i should do a little more manual reading again.
[20:42] <persia> Isn't UTF-8 usually handled OK (terminal width aside) even from C strings?  "僕の本" is more than 3 bytes, but does one typically need to care for output purposes?
[20:43] <persia> Ah, question answered while being typed :)
[20:43] <geser> bazaar: in that case printf() should be ok, but don't try to use a multi-byte char string with printf("%80s", ...) to fit it in one terminal line
[20:43] <slangasek> persia: for output - no.  But if you want to compare two chars, instead of two strings, you need a "char" type that can accomodate kanji :)
[20:44] <persia> slangasek: Right, which is why string length of 1 byte is wrong (but string length of 1 character is right), which is why I should not look at any more loosely typed languages.
[20:46] <kees> asac: how do I mark myself "online" so things like firefox don't think I'm offline?
[20:46] <geser> persia: your example is also good to show that 3 (wide) characters may need more space than 3 terminal "cells"
[20:47] <persia> geser: Indeed, although that's common for CJK.  "エメット" is preferred over "ｴﾒｯﾄ" when writing my name, for example.
[20:48] <persia> (plus, not all characters have a half-width form)
[20:50] <geser> but one to think about it when converting an app from char* to support multi-byte strings (UTF-8) and not simply replace strlen() with wcslen() when trying to compute how much space a string will need
[20:51] <geser> space on the terminal
[21:22] <poningru> hey guys who should I talk to regarding the netboot initrd and the structure in it?
[21:23] <poningru> as in this guy: http://archive.ubuntu.com/ubuntu/dists/hardy/main/installer-i386/current/images/netboot/ubuntu-installer/i386/initrd.gz
[21:23] <Mithrandir> poningru: try #ubuntu-installer, probably.
[21:23] <poningru> danke
[22:01] <slangasek> hmm, is prerm remove + removal of files from the filesystem always an atomic operation wrt dpkg?
[22:05] <DktrKranz> are packages in -proposed (once they've been built and published) available to other packages in proposed too or they become available only when moved to -updates?
[22:08] <sbeattie> DktrKranz: yes, packages in -proposed are built against -proposed as well (though it's a policy I disagree with)
[22:09] <DktrKranz> thanks
[22:10] <slangasek> sbeattie: well, I don't know how you would propose we stage kernel ABI changes without it :)
[23:42] <jcastro> anyone know the component responsible for ejecting the CD at the end of the install?
[23:42] <jcastro> or should I just file a bug against ubiquity?
[23:42] <slangasek> eject-udeb
[23:42] <jcastro> thanks
[23:42] <slangasek> but ubiquity is usually a good place to start anyway, for liveCD bugs
[23:43] <jcastro> well, I wanted to file a feature request for the server install, is ubiquity still the right place?
[23:44] <jcastro> dealing with the eject stuff I mean
[23:44] <slangasek> oh, for server install, no - that's definitely eject-udeb
[23:44] <jcastro> ok