[00:02] <ScottK> slangasek: kde4libs built on ia64 after your qt4-x11 fix.  Thanks.
[00:07] <calc> is usr/share/icons/locolor still a legitimate icon theme dir?
[00:08]  * calc was wondering if he should remove that along with the gnome dir from OOo since iirc that was an old kde 2.x dir
[00:11] <Riddell> calc: it's just an icon theme like any other
[00:12] <Riddell> it's the KDE 1 icons which do still exist in kdeartwork for anyone strange enough to want them
[00:12] <Riddell> no point keeping them in OOo anyway
[00:20] <calc> ok
[00:21] <calc> that removed ~ 250KB from openoffice.org-common :)
[01:47] <Hobbsee> sbeattie: yeah.  maybe mine just hates the world or something, although i think i saw someone else mention it in -bugs a fwe days ago
[03:58] <slangasek> ScottK: aha, great \o/
[03:58] <slangasek> ScottK: now it's just hppa (strigi) cluttering up component-mismatches
[03:58] <ScottK> slangasek: I got started on retries.
[03:59] <ScottK> Unfortunatlely sparc didn't like something about your qt4-x11 change.
[03:59] <slangasek> I had already given back strigi once this morning on hppa, still FTBFS with the same error; apparently qt is uninstallable there
[03:59] <slangasek> bah, sparc shouldn't have cared about that change at all, it's just being a whiner
[04:06] <picklesworth> is there some accepted standard for passing files between running apps that talk to each other over dbus?
[04:06] <picklesworth> or should I just do whatever makes sense? :)
[04:17] <slangasek> bleh, qt on hppa needs manual bootstrapping; phonon and qt have a circular dep.
[04:18] <ScottK> Ouch.  When did we do that?
[04:28] <slangasek> ScottK: qt4-x11 (4.5.0~+rc1-0ubuntu1), 19 Feb 2009, AFAICS
[04:33] <lamont`> wtf does network mangler ((or whoever) decide to put both search and domain directives in named.conf?
[04:33] <lamont`> slangasek: ew
[04:33] <lamont`> er, resolv.conf that is
[04:35] <slangasek> you object to having both?
[04:40]  * lamont makes a note to remember to edit network mangler's postinst _BEFORE_ removing it remotely next time
[04:46] <mneptok> lamont: i'm in Boston, and will accept assassination contracts for Red Hat's Westford office ;)
[04:47] <lamont> mneptok: I do not publicly issue contracts for wetwork
[04:48] <lamont> it's amusing how many people think that having domain and search directives in resolv.conf actually does anything
[04:48] <lamont> (other than helping pre-1990 programs)
[04:50] <lamont> meh.  not worth figuring out what all dbus is doing under the covers for the networkmanager stop call in prerm
[04:50] <lamont> I just hate fighting network mangler, is all
[04:51] <lamont> the cute part this time was that since it got dragged in on the upgrade, the machine got ipv6 addresses only.. not sure why it hated the ipv4 dhcp response
[04:51] <lamont> (and it was an old-n-crufty machine, so not sure I care enough to worry about it)
[06:18] <slangasek> lamont: huh?  of course search and domain directives do something...?
[06:51] <calc> OOo 3.0.1-9ubuntu1 will be uploaded in about 2hr
[06:51]  * calc headed to bed now
[07:47] <shahan> hi, i was wondering if sed supports backreferences in ranges using patterns?
[08:55] <robert_ancell> does anyone know of a public sftp server I can connect to to test bug 27109?
[08:57] <seb128> hey robert_ancell
[08:57] <robert_ancell> seb128: hey seb
[08:58] <hyperair> robert_ancell: set up a ssh host locally?
[08:58] <seb128> robert_ancell: don't you have ssh access to the GNOME boxes to upload tarballs or to some canonical servers?
[08:58] <robert_ancell> hyperair: yeah, was thinking of that but looking for easier option first :)
[08:58] <hyperair> robert_ancell: or register for an account at alioth.
[08:58] <hyperair> robert_ancell: but really, you can't get any easier than sudo apt-get install openssh-server =p
[08:59] <robert_ancell> hyperair: does that do sftp?  Or is sftp just scp?
[08:59] <hyperair> robert_ancell: sshd does sftp and scp
[08:59] <hyperair> robert_ancell: they're different protocols i think,  but they're both provided by sshd.
[09:00] <robert_ancell> hyperair: cool, my local server was already working for sftp :)
[09:01]  * robert_ancell finds it confusing that nautilus refers to sftp as "SSH"
[09:01] <hyperair> robert_ancell: hahah =p
[09:01] <hyperair> well it is ssh://
[09:01] <hyperair> sftp:// and ssh:// are equivalent.
[09:01] <robert_ancell> 2 products for the price of 1!
[09:01] <hyperair> heh
[09:02]  * hyperair prefers rsync through ssh
[09:28] <bryce> kirkland: here's the page I mentioned the other day - http://qa.ubuntu.com/reports/bug-fixing/jaunty-fixes-report.html (and let us know how the race went)
[11:40] <Ng> slangasek: is it just me, or did your upload of hotkey-setup break its init script?
[11:41] <Ng> lala, nm, there's a bug report already
[11:41] <lamont> slangasek: only one of them does, removing all effects of the other
[11:41] <lamont> I think the last one matters, but haven't bothered to check
[11:51] <Ng> slangasek: bug #356157
[12:37] <cjwatson> I've answered the mail to ubuntumembers about Ubuntu community training with proper references; I imagine I'm not the first to have done so but thought I'd mention it so that others can delete the mail that was spammed to an unreasonably large number of people ;)
[12:43] <lifeless> Should we disable 'contact this group' for, oh I don't know, every group ?
[12:45] <Hobbsee> lifeless: can you?
[12:45] <cjwatson> lifeless: it certainly ought to have a bloody great warning
[12:45] <cjwatson> lifeless: since if nothing else it doesn't go to anything like a mailing list, so members of the group won't see followups
[12:46] <cjwatson> (actually I think the total result would be worse if it did, but nevertheless ...)
[12:46] <lifeless> Hobbsee: its just code
[12:46] <Hobbsee> cjwatson: oh, it does make it relatively clear that you're about to send mail to 450 people.  But it doesn't include the blink tag or anything, or smack them in the face.
[12:47] <Hobbsee> presumably people do read that
[12:47] <Hobbsee> lifeless: i know it's theoretically possible.  :)  I'm just more wondering if a) it would be permissable to be done, and b) it would get done without being quickly reverted
[12:49] <elmo> cjwatson: there's bugs about both the flawed semantics of c-t-u with groups, and that c-t-u works with groups of people > e.g. 50
[12:49] <elmo> FWIW
[12:53] <cjwatson> elmo: thanks
[12:54] <mvo> doko: I noticed that you reverted the change for the supported/unsuppored logic in python3.0 - with my latest pycentral changes it should no longer die if a runtime is neither support/unsupported. changing this also means that rtinstall is never run for python3.0 (because its not in debian_defaults).
[12:54] <mvo> (not sure if that matter for 3.0 though)
[12:55] <doko> mvo: it should be fine for 3.0, we don't have any third party stuff packaged for 3.0 yet (and we should not, because we do know that 3.0 is only short term supported)
[12:55] <mvo> ok
[12:56] <Ng> mvo: the patch in the latest compiz upload has at least one notable regression - bug #355330
[12:57] <Ng> mvo: but perhaps the gnome applet is doing the wrong thing
[12:57] <seb128> Ng: as you can see mvo already milestoned the bug ;-)
[12:58] <mvo> Ng: yeah, I noticed this morning
[12:58] <seb128> Ng: we discussed it this morning on #ubuntu-desktop it breaks other case too, session dialogs being one of those for some users, gnome-do too
[12:58]  * mvo prepares a fix
[12:58] <Ng> seb128: that must be a different bug then, 355330 is unmilestoned. I'll read #ubuntu-desktop backscroll and dupe it
[12:59] <Ng> thanks :)
[12:59] <seb128> Ng: bug #355379
[12:59] <seb128> Ng: well, milestoned -> has a jaunty task
[13:00] <seb128> that's how jaunty bugs are tracked
[13:00] <Ng> oh :)
[13:01] <Ng> gar, all the dupes of this have dupes already
[13:02] <Ng> there :)
[13:08] <mvo> sorry, working on a fix as we speak
[13:27] <cjwatson> nggh, my master installer translation files all broke
[13:32] <ogra> cjwatson, ah, btw, i noticed on the partman page where we have $(RELEASE) there is also a stray $(OS) at the top line
[13:33] <ogra> (i was assuming you look into ( vs { anyway for that page, but didnt want to leave it unmentioned
[13:33] <ogra> )
[13:34] <cjwatson> ogra: I mailed ubuntu-translators about it and have been told that it's fixed
[13:34] <ogra> ok, then it should show up soon i guess
[13:34] <cjwatson> ogra: but I'll need to do an import, which is part of why I ran into the problem mentioned above ...
[13:34] <ogra> ah, k ... i'll report if i see it fixed, i'll do a daily install every day anyway
[14:15] <dschulz> hi all
[14:16] <dschulz> anyone noticed the error in hotkey-setup package after updating?  there's an error in the initscript
[14:19] <Ampelbein> dschulz: bug #356157
[14:20] <dschulz> yes
[14:20] <dschulz>  found it
[14:44] <facundobatista> Hi all
[14:56] <facundobatista> tkamppeter, I wanted to tell you that I'm here if you want to try anything (I commented the bug #348316... with the updated cups package I don't see my printer anymore, if I downgrade to your PPA version I see it again)
[14:58] <tkamppeter> facundobatista: When you upgrade to CUPS 1.3.9-17 you must remove the usblp module from the blacklist. The official CUPS package does not contain the new libusb-based backend, as the USB backend was not what caused the bug.
[14:59] <facundobatista> tkamppeter, oh! I try this now
[15:04] <facundobatista> tkamppeter, ok, now I removed it from the blacklist, and loaded it (sudo modprobe usblp), and I see it with lsmod
[15:08] <facundobatista> tkamppeter, now I see the printer (it even appears automatically when unplu/plug the cable), I can send a test page to the printer, it puts a job in the queue, the job disappears, I have a notification that the job was printed, and the printer goes to idle state
[15:08] <facundobatista> tkamppeter, but the printer just doesn't do anything, no printing at all
[15:14] <qense> It seems there is a typo in the latest version of hotkey-setup on jaunty
[15:15] <Hobbsee> https://launchpad.net/bugs/356157
[15:15] <qense> ah
[15:15] <qense> already reported, thanks for the link
[15:16] <qense> sometimes I get the feeling I'm just too lazy, sorry for letting you search up things I could have found myself
[15:16] <Hobbsee> it was mentioned in here a couple of hours ago, so was a matter of scrolling ;)
[15:17] <qense> ok :)
[15:17] <qense> a huge list of duplicates
[15:26] <Tm_T> Hobbsee: as I mentioned elsewhere, makes kinda worried when you notice who's the maintainer of the package, don't they test them? (;
[15:27] <Hobbsee> Tm_T: at this point of the release?  yeah...
[15:27] <Tm_T> Hobbsee: indeed
[15:29]  * Hobbsee ponders uploading the fix, but notes it's now just been assigned to slangasek
[15:30] <Tm_T> Hobbsee: to be done when?
[15:31] <Hobbsee> Tm_T: asap, i guess?
[15:31] <Tm_T> I hope so (:
[15:32] <Tm_T> still, that's kind of things that shouldn't go to users at all IMO
[15:34] <Hobbsee> anytime you want to start responding again launchpad, feel free...
[15:34] <Hobbsee> 43 seconds and no timeout.  Good!
[15:34] <Tm_T> whoa
[15:34] <Tm_T> "answer to me, baby"
[15:35] <Tm_T> (baby being genderless in this case)
[15:35] <ogra> poor baby
[15:35] <thiebaude> someone asked me, how many devs are working on 9.04?
[15:35] <Tm_T> ogra: not as poor as me, anyway
[15:35] <Tm_T> thiebaude: depends on how you count
[15:36] <ogra> Tm_T, shit happens in development releases, so such stuff goes out to users ... thats why they are called *development* releses ;)
[15:36] <ogra> *releases
[15:36] <thiebaude> Tm_T: canonical wise
[15:36] <Tm_T> ogra: I know, but some level of professionalism in core team, please (;
[15:37] <Tm_T> ogra: I'm too used to messages like "I commit this but haven't tested at all" and everything breaks in upstreams
[15:37] <ogra> even prtofessionals make mistakes ... we're all humand
[15:37] <robbiew> Hobbsee: Thanks for sponsoring the fix for bug 356157 :)
[15:37] <ogra> *humans
[15:37] <Tm_T> ogra: I know (:
[15:37] <Tm_T> ogra: still weird, no testing apparently
[15:37] <Tm_T> robbiew: (:
[15:38] <robbiew> Tm_T: noted
[15:39]  * robbiew will reinforce the importance of testing *before* uploading
[15:39] <Tm_T> robbiew: FYI I'm usually the test monkey in many places oftentimes
[15:39] <ogra> robbiew, pfft, that takes out all the fun communication with users ;) how else shall we train our diplomacy ?
[15:40] <Hobbsee> robbiew: no problem.
[15:40] <Tm_T> ogra: war is best way to do diplomacy anyway!
[15:40] <Hobbsee> robbiew: pbuilder hooks or piuparts ftw!
[15:41] <cjwatson> Tm_T: the usual (unfortunate but understandable) pattern is that it was tested but then a slightly different version was uploaded by mistake. I don't think we need to throw around comments about professionalism for a single pre-release mistake
[15:42] <robbiew> Hobbsee: liw as been looking into piuparts
[15:42] <robbiew> cjwatson: +1
[15:42] <Tm_T> cjwatson: I know, it was meant to be a humour, sorry if not apparent
[15:42] <Hobbsee> robbiew: indeed.
[15:42] <tkamppeter> facundobatista: Your printer (HP LJ P1002) is too new, it is not yet supported.
[15:43] <robbiew> Tm_T: this late in the release...humour is the first thing to go :P
[15:43] <Tm_T> cjwatson: I guess this is the "finnish honesty" you can see in Linus' comments too (:)
[15:43] <Tm_T> robbiew: to go in, you mean
[15:43] <Hobbsee> robbiew: just add more beer, and the humour will likely come back ;)
[15:43] <facundobatista> tkamppeter, not supported by who? gnome? kernel? ubuntu?
[15:44] <Tm_T> Hobbsee: or puke
[15:44]  * calc feels less stress today after getting 8 targeted bugs finished last night :)
[15:44] <robbiew> Hobbsee: heh...oh, it comes back...about 2am London time on the day of the release
[15:44] <Tm_T> robbiew: when 2 months more testing is announced?
[15:44] <Hobbsee> robbiew: yeah, i remember that well...
[15:45] <tkamppeter> facundobatista: Linux. The printer most probably needs to get a firmware file uploaded into it whenever it is turned on. This firmware file is not yet available wioth the driver packages.
[15:47] <cjwatson> thiebaude: from Canonical, on the general order of 50 (that isn't an exact number); but many Ubuntu developers are not employed by Canonical, and of course many of Canonical's staff are working on things other than Ubuntu or only contribute to Ubuntu occasionally
[15:48] <Tm_T> indeed
[15:48] <Tm_T> cjwatson: is team numbers kept somewhere in one place?
[15:48] <Tm_T> team member numbers that is
[15:49] <cjwatson> Tm_T: to get the answer to thiebaude's question you need to take the intersection of a couple of teams
[15:49] <cjwatson> (one of which is private)
[15:50] <Tm_T> cjwatson: aye, also just being part of the team doesn't mean youre currently working
[15:50] <slangasek> Ng, Hobbsee, nxvl_: ah, whoops - sorry :/
[15:50] <Tm_T> cjwatson: would be nice to build a graph of team sizes etc from release to release
[15:53] <facundobatista> tkamppeter, I'm using the Foomatic/foo2xqx driver... in the driver's page, it says that it supports HP LJ 1005, but at the same time it says that I shouldn't use the package from Ubuntu, but install the tarball they provide there
[15:53] <facundobatista> tkamppeter, do you know why?
[15:58] <tkamppeter> facundobatista: This is because the author insists to download and install the firmware files from the internet during the compile and build process and not by a user utility as HP does ("hp-plugin").
[15:59] <facundobatista> tkamppeter, but there's a firmware for my printer in that case... but it's not yet in the packages?
[15:59] <tkamppeter> So you do not need to try to rebuild foo2zjs. You can try to set up your printer with "hp-setup", but probably it does not get recognized as HPLIP's current model list does not contain it.
[16:00] <tkamppeter> facundobatista: Building foo2zjs starts getweb with the model name of the printer as argument.
[16:00] <tkamppeter> facundobatista: The foo2zjs version I packaged one day before Feature Freeze did not yet list the printer in "getweb".
[16:01] <tkamppeter> facundobatista: Did you download the newest version? Did it add support for the P1002?
[16:02] <facundobatista> tkamppeter, well, I have a confusion here... my printer is HP 1005, but through usb I see it as P1002, do you know why this duality?
[16:03] <tkamppeter> facundobatista: This is extremely strange for me.
[16:03] <tkamppeter> facundobatista: Can you report an upstream bug on HPLIP, on https://launchpad.net/hplip/?
[16:04] <tkamppeter> facundobatista: They do not fix hardware bugs, but they could take into account this version of the P1005 in their driver.
[16:04] <tkamppeter> facundobatista: In which country did you buy that printer?
[16:05] <facundobatista> tkamppeter, I have in my disk the foo2zjs tarball from one month ago, and in its INSTALL file it says to execute "./getweb P1005", so it was supported then
[16:05] <facundobatista> tkamppeter, Argentina
[16:06] <facundobatista> tkamppeter, I'll bug the hplip project about how it recognizes my printer
[16:06] <tkamppeter> facundobatista: You should perhaps try to load the P1005 firmware file into your printer and see whether it works then. Then your printer principally works but the wrong self identification breaks the automatic setup.
[16:07] <facundobatista> tkamppeter, how can I load the firmware? using hp-plugin?
[16:07] <cjwatson> slangasek: FYI I finally managed a successful installation with grub2 today, after a few bug-fixes
[16:09] <tkamppeter> download it with getweb and convert it with arm2hpdl. Then do "sudo -s" and after that "cat <converted firmware file> > /dev/usb/lp0",  then do "exit".
[16:09] <facundobatista> tkamppeter, oh, ok
[16:09] <robbiew> cjwatson: sweet!
[16:14] <facundobatista> tkamppeter, did that, the printer *did* something (like ininting), and went back to idle/ready state
[16:14] <facundobatista> tkamppeter, I printed a test page, but it stayed in "processing", and then received a notification asking if the printer was disconnected... ??
[16:17] <tkamppeter> Seems that the printer crashed internally, due to a firmware which is not 100% compatible with it.
[16:17] <facundobatista> tkamppeter, which project should I bug about this firmware? hplip?
[16:22] <facundobatista> tkamppeter, otoh, which is your opinion about trying the hplip binary to set up everything?
[16:35] <tkamppeter> facundobatista: Simply try it, but HPLIP will not recognize your printer as supported due to the bad self-identification. Please bug the HPLIP project about the firmware.
[16:35] <facundobatista> tkamppeter, ok!
[16:35] <facundobatista> tkamppeter, thank you!
[16:40] <MacSlow> bryce, hey there... does http://launchpadlibrarian.net/24174661/Stacktrace.txt look like something familiar to you (in regards to known bugs/issues in xlib or xcb)?
[17:08] <slangasek> cjwatson: grub2> ah, great!
[17:13] <cjwatson> slangasek: Julian Edwards informed me a while back that they have an experimental test rebuild system in Launchpad, and I forgot to actually give it a test run. I'd like to do so now; it uses the normal buildds so it'll create a *lot* of pending builds, but they'll all be scored down to 4
[17:13] <cjwatson> slangasek: would that be OK?
[17:13] <slangasek> cjwatson: that seems ok, yes
[17:18] <cjwatson> I might do it just for the primary architectures
[17:21] <lamont> wtf does gimp think that doing forcing a fullsize window for the pic is anything close to the right answer?
[17:21] <slytherin> seb128: Do you have any plan on cherry picking Jan Schmidt's enhancements to DVD playback (http://noraisin.net/~jan/diary/?p=90). This will require changes to gst-plugins-base, totem and gst-plugins-bad. If not, do you mind if I work on it?
[17:21] <seb128> no
[17:21] <lamont> or is that just the same stupidity that makes firefox think that the fact that it finished rendering a page means that I want it to deiconify even though I just iconfied it?
[17:21] <seb128> slytherin: feel free to work on it
[17:22] <slytherin> seb128: Even when I am cherry picking the changes it introduces new API in -base, so I will need FFE, right?
[17:22] <seb128> yes, you can work on it, I can't tell if it will be accepted or not I'm not the one deciding there
[17:23] <seb128> I also need to reply to your upload on the list
[17:23] <slytherin> seb128: Ok.
[17:23] <Tm_T> robbiew: hmm, should I poke you when I get issues with this new package too?
[17:23] <seb128> it seems to be a "let's change the build system, we don't know why but that seems to work better so it must be something to get"
[17:23] <seb128> slytherin: which I disagree with
[17:24] <robbiew> Tm_T: heh...you can, but please file a bug also ;)
[17:26] <Tm_T> robbiew: well I don't know if this worth a report, just asking should this happen? http://paste.ubuntu.com/145536/
[17:27] <robbiew> Tm_T: yuk
[17:27] <robbiew> slangasek: ^^?
[17:28] <slangasek> bluh
[17:28] <slytherin> seb128: I know. I don't have any argument in favour in the change except that it works better. As I said I am not expert in autotools.
[17:28] <slangasek> robbiew: I'll poke at it and do evil things
[17:28] <seb128> slytherin: "work better"
[17:28] <seb128> slytherin: on what basis do you assume that?
[17:28] <robbiew> slangasek: heh...ok, thanks
[17:28] <cjwatson> slangasek: https://edge.launchpad.net/ubuntu/+archives, FYI
[17:28] <cjwatson> i386 only (by accident but actually that's probably OK)
[17:30] <slytherin> seb128: I did quite a good amount of testing with many DVDs I have and I have feed back from at least 2 users on the bug that the change worked. I wish I could gather more feedback.
[17:32] <seb128> slytherin: we have no idea of what the change do, it might change things in a better way for your 3 users and break for a thousand user out there
[17:32] <picklesworth> Tm_T: http://ubuntuforums.org/showthread.php?t=1117615
[17:33] <picklesworth> You'll need to fix /etc/init.d/hotkey-setup yourself, unfortunately. This post shows how: http://ubuntuforums.org/showpost.php?p=7023434&postcount=6
[17:33] <picklesworth> I'm sure the package will never upload an untested package again ;)
[17:33] <slangasek> I'm fairly certain of the opposite
[17:34] <slangasek> I'm too old to learn that lesson :P
[17:34] <picklesworth> heh
[17:34] <Tm_T> picklesworth: I know
[17:34] <slangasek> anyway, I have a fix that should work for recovering from the broken state
[17:34] <slangasek> I suppose I should test it first though :-P
[17:35] <Tm_T> slangasek: aye, that's my meaning at this
[17:35] <slangasek> yep, fix checks out
[17:36] <slangasek> robbiew: don't look at my patch
[17:36] <robbiew> slangasek: lol...ok
[17:36] <Tm_T> slangasek: meaning, is, getting it fixed properly, not making you feel sad (:
[17:36] <slangasek> cjwatson can look at it if he wants, though
[17:37] <slangasek> Tm_T: right :)  well, fix is uploaded now
[17:37] <cjwatson> only if I know where it is
[17:37] <slytherin> seb128: I have made multiple calls for testing multiple times on IRC (#ubuntu+1) for packages in my PPA and didn't get any regression report.
[17:37] <slangasek> cjwatson: hotkey-setup 0.1-23ubuntu12
[17:37] <cjwatson> ah
[17:37] <Tm_T> slangasek: good, I'll keep poking if something comes up
[17:38] <slangasek> cjwatson: using prerm failed-upgrade \o/
[17:38] <cjwatson> slangasek: the test rebuild is a bit of a wash, they all fail to upload
[17:38] <seb128> slytherin: I've seen no evidence recently that DVD playback is broken for the majority of users
[17:38] <slangasek> cjwatson: heh
[17:38] <cjwatson> slangasek: I'm asking muharem about it
[17:38]  * slangasek nods
[17:40] <calc> argh, a catalan language update for OOo managed to break the build :\
[17:41] <slangasek> make the translation coordinator fix it
[17:41] <calc> slangasek: it was a GSI attached to a bug report and aiui the guy who used to do catalan (jordi mallach) no longer does it
[17:42]  * calc will just be dropping it from the next upload probably tomorrow
[17:42] <slangasek> calc: but Ubuntu now has a translations coordinator, who also speaks Catalan :)
[17:42] <calc> slangasek: ok
[17:43] <slytherin> seb128: So what do you suggest, apart from analysing the changes I did.
[17:43] <seb128> slytherin: having somebody to try to understand what the change is doing before pushing it to jaunty one week before freeze this way
[17:44] <slytherin> seb128: Ok. I will try to get hold of someone.
[17:49] <slangasek> ScottK: two steps forward, one step back for qt on hppa; strigi builds, but phonon FTBFS because automoc4 (or something it calls) segfaults :P
[17:50] <cjwatson> slangasek: hotkey-setup> yum!
[17:50] <slangasek> :-)
[17:51] <Tm_T> slangasek: ouch
[17:54] <BUGabundo> we are getting reports on +1 that the daily installer lost support for ext4 using the manual option!
[17:54] <BUGabundo> is this know issue?
[17:54] <calc> hotkey-setup seems almost broken in a way package management can't fix, even purge force-all wouldn't remove it until i fixed the init script
[17:54] <calc> then installing the new version worked ok
[17:55] <BUGabundo> lol calc
[17:55] <slangasek> calc: ahh, but it can be fixed, you just have to channel IanJ when patching the package
[17:55] <slangasek> (fix uploaded, as mentioned above - will be available in ~1h)
[17:55] <calc> slangasek: ah ok
[17:56] <calc> slangasek: oh the ordering of when helper scripts fall back could then move the broken init script out of the way (i'm guessing?)
[17:57] <slangasek> calc: don't actually need to move the broken init script out of the way, I just added code in the prerm of the new package version to skip invoking the init script altogether
[17:57] <calc> eg:
[17:57] <calc> ah :)
[17:57] <d3xter> colin: ping
[17:57] <calc> yea i saw where it tries the new prerm script (had forgotten about that ordering in policy)
[17:58] <calc> slangasek: thursday is just the last day new uploads will be accepted without a really good exception right? (meaning OOo doesn't have to be completely finished building by then?)
[17:58] <BUGabundo> d3xter: I already asked! we just have to wait for a reply on it
[17:58] <slangasek> calc: right - though it would certainly be nice if OOo was done before then...
[17:59] <calc> slangasek: i'm trying to determine when i need to do my next upload since 9ubuntu1 had issues with catalan and thunderbird
[17:59] <d3xter> thx, i've read your msg right after i sent my ^^
[17:59] <calc> slangasek: i think it would be best to wait at least one day to see if any other bug reports come in before uploading again though
[18:00] <slangasek> calc: does OOo really need an upload for an OOo-l10n build failure?
[18:00] <slangasek> makes me think the OOo/OOo-l10n split is worse-than-useless
[18:00] <cjwatson> BUGabundo: known and fixed
[18:00] <calc> slangasek: there were a couple of issues in OOo that didn't get discovered until too late, the thunderbird issue and a desktop file issue
[18:01] <BUGabundo> cjwatson: thanks
[18:01] <cjwatson> BUGabundo: bug 354851
[18:01] <BUGabundo> cjwatson: is there a LP bug I can refer users to?
[18:01] <BUGabundo> bah.. thanks
[18:01] <slangasek> calc: ah, ok
[18:01] <calc> slangasek: yea the split is really not too useful imho, it may become more useful when we eventually go to a full split build
[18:02] <calc> slangasek: the full split build has something like 15 source packages, and won't require building all of OOo just for languages
[18:02] <slangasek> calc: well, AIUI we currently have lots of code duplication in the tarballs of the two source packages; will that be fixed?
[18:02]  * calc is considering forking from debian for OOo 3.1 packaging and then seeing if he can get the changes integrated back into debian version
[18:03] <calc> slangasek: yea long term that will be fixed. currently it is 100% duplication
[18:03]  * slangasek nods
[18:03] <calc> slangasek: all i do is rename the orig.tar.gz :\
[18:03] <slangasek> forking from Debian> yuck, only if you're planning on discarding the debian/ directory and redoing it from scratch :)
[18:04] <calc> i think it is supposed to be fixable by 3.2 (karmic) if things go well
[18:04] <slangasek> lamont: we don't have an hppa porter box?
[18:05] <lamont> slangasek: wongi
[18:05] <calc> slangasek: yea i am thinking of doing a complete repackaging while using his rules file as a guide, the more i see bits of the rules the more scared i become ;-)
[18:05] <Mirv> calc: have you discussed with rene about the idea?
[18:05] <BUGabundo> cjwatson: one more question if I may. can the base ubuntu be installed and then install the ubuntu-netbook-remix package?
[18:05] <calc> Mirv: not yet, i am going to see if it is feasible first then discuss with him
[18:05] <slangasek> lamont: ah, thanks; adding to the wiki then
[18:05] <cjwatson> BUGabundo: no idea, not my field
[18:05] <Mirv> calc: ok.
[18:05] <BUGabundo> oh who is then ?
[18:06] <calc> Mirv: we were originally going to fork for the split build anyway this release, but that looks to be delayed
[18:06] <calc> Mirv: while doing some of the split build work i started noticing that the rules file could do for a major cleanup
[18:07] <cjwatson> BUGabundo: I'd probably start with #ubuntu-mobile
[18:07] <Mirv> calc: anything to make maintaining OOo easier would probably be very welcome, it's just so huge and rene is I guess still the "OOo team" in Debian...
[18:09] <jefferai> Anyone know what should be creating /var/run/network directory in Jaunty?
[18:10] <jefferai> in previous versions it was /etc/init.d/loopback
[18:10] <jefferai> but that doesn't seem to exist anymore
[18:11] <calc> Mirv: yes pretty much just me and rene between ubuntu and debian
[18:12] <calc> Mirv: some of what i want to do wrt packaging is get closer to the packaging layout of upstream... without the braindamage of core## packages
[18:12] <calc> Mirv: so we won't run into each user level application needing to depend on each other
[18:26] <bryce> MacSlow: not offhand; get a full backtrace and file a bug
[18:26] <MacSlow> bryce, that's all I have right now
[18:29] <bryce> MacSlow: ok, well not enough in that backtrace to analyze; just indicates some sort of crash malloc'ing memory in xcb
[18:30] <MacSlow> bryce, yeah... that's from a bug filed against notify-osd... and that stracktrace indicates that the error does not really happen in notify-osd
[18:31] <MacSlow> bryce, I thought you might recognize a "pattern" and happen to know a bug in xlib or xcb related to that
[18:31] <bryce> ah, nope
[18:34] <LaserJock> bryce: quick question, my X locks up, keyboard is dead, but my mouse (either touchpad or external) will move, do you know a likely culprit?
[18:34] <bryce> LaserJock: too generic of a symptom to say
[18:35] <LaserJock> hmm, ok
[18:35] <bryce> LaserJock: that's basically a GPU lockup.  lots of stuff can do that
[18:35] <LaserJock> ah
[18:35] <LaserJock> would not running compiz help?
[18:35] <slangasek> could also be an input dev bug, though?
[18:36] <LaserJock> it started in the last ~ 1 week and it's done it about 3 times
[18:36] <bryce> LaserJock: quite possibly
[18:36] <LaserJock> I have to hard reboot
[18:36] <sbeattie> LaserJock: are you able to ssh in?
[18:36] <LaserJock> good question, I'm not sure
[18:36] <bryce> slangasek: possibly, but I think a gpu lockup is more likely
[18:36] <LaserJock> next time I'll find a machine to ssh with
[18:37] <slangasek> fairy nuff
[18:37] <LaserJock> it's just odd to have the cursor working just fine but everything else just lock up
[18:37] <bryce> LaserJock: if you know it worked properly a week ago, reverting packages until you find the one that causes it could be helpful
[18:37] <LaserJock> I was wondering since the intel drivers have been updated
[18:38] <BUGabundo> LaserJock: I got that on the Dell machines I was testing last week
[18:38] <bryce> I've put in several -intel patches recently, and in my experience with them, fixing one bug seems to cause another somewhere else :-/
[18:38] <BUGabundo> they add the lovely intel 865
[18:38] <BUGabundo> the live cd would freeze but cursor still worked
[18:38] <LaserJock> bryce: I'll right, I'll try going back on -intel and see if that seems to solve it
[18:39] <LaserJock> bryce: it's often enough to  be annoying but not often enough to be easy to debug :-)
[18:39] <sbeattie> bryce: what's useful to do for debugging a lockup like that
[18:40] <bryce> sbeattie: I don't have a good debugging procedure for lockups yet
[18:40] <bryce> sbeattie: any error message you can get is usually handy
[18:40] <bryce> they're usually not in Xorg.0.log tough, so you have to poke around
[18:41] <bryce> backtraces are only marginally useful
[18:41] <bryce> sbeattie: turning off DRI or other X subsystems to narrow it down can be useful
[18:41] <bryce> sbeattie: the most helpful thing is to bisect or revert backwards until you find the change that caused it
[18:44] <LaserJock> bryce: I'll start with 2:2.6.3-0ubuntu1 and bisect my way
[18:44] <sbeattie> bryce: okay. The lockup I saw didn't show anything useful anywhere, only after trying suspend-resume to unwedge it did i start getting "[mi] EQ overflowing. The server is probably stuck in an infinite loop." in the log
[18:44] <LaserJock> I'm pretty sure it started after ubuntu1
[18:46] <bryce> I put up a http://wiki.ubuntu.com/X/Glossary to help explain some of the technical terms you see in X error messages
[18:47] <sbeattie> bryce: note that I do little to no 3d stuff; my wm doesn't composite.
[18:48] <bryce> sbeattie: well... like I mentioned before, gpu lockups are kind of a generic symptom and can occur due to a wide variety of things
[18:48] <kirkland> james_w: where can I find some documentation on bzr-builddeb?  like how to setup my debian/rules properly?
[18:48] <james_w> hey kirkland
[18:48] <bryce> 3d stuff tends to trip up the gpu more, because 3d code is often newer, the registers vary more widely from hw to hw, and there's a lot more to 3d than 2d
[18:49] <james_w> kirkland: it should require no changes to debian/rules
[18:49] <james_w> kirkland: or anything in debian/ at all
[18:49] <kirkland> james_w: howdy
[18:49] <kirkland> james_w: hmm, okay.  where can i find some instructions on making my get-orig-source target correct?
[18:49] <kirkland> james_w: b/c right now, builddeb is not happy with my get-orig-source
[18:49] <james_w> kirkland: do you have a log?
[18:50] <james_w> builddeb was actually broken with get-orig-source until a few days ago, so if you don't have the latest package it won't work
[18:50] <kirkland> james_w: http://pastebin.ubuntu.com/145602/
[18:50] <kirkland> james_w: let me update
[18:52] <james_w> kirkland: ah, it's failing before where that fix would kick in so it's something else
[18:52] <james_w> that's an interesting get-orig-source you have there
[18:52] <james_w> is the intent to use the cwd without ./debian/ as the .orig.tar.gz?
[18:57] <kirkland> james_w: http://pastebin.ubuntu.com/145602/
[18:59] <james_w> kirkland: are you trying to tar up the cwd without ./debian/ to create the .orig.tar.gz?
[19:00] <kirkland> james_w: yes, and without the .bzr
[19:00] <calc> elmo: so thunderbird did actually work in the past for the users who are reporting the problem now?
[19:00] <kirkland> james_w: i'll pastebin the rule too
[19:00] <elmo> calc: yes
[19:00] <james_w> kirkland: you realise that can easily lead to broken packages?
[19:00] <kirkland> james_w: i did not realize that
[19:00] <james_w> ones that aren't a danger to the archive at least, but would still be a pain
[19:00] <elmo> calc: and it works for them again if I s/sensible-ooomua/thunderbird/
[19:00] <elmo> calc: (I haven't tried xdg-mailer or whatever it was you suggested)
[19:01] <calc> elmo: interesting, ok, i will look into the file that the patch that adds "sensible-ooomua"
[19:01] <kirkland> james_w: i can drop that --exclude
[19:01] <james_w> kirkland: well, it makes it easy to create a broken package if you try and build a -2 at least
[19:01] <james_w> kirkland: and you don't have the original .orig.tar.gz in the correct place
[19:01] <kirkland> james_w: http://pastebin.ubuntu.com/145616
[19:01] <calc> elmo: afaict it has used that setting since at least hardy, so maybe the file the patch changes has otherwise changed as well
[19:01]  * calc hopes that made sense :)
[19:01] <james_w> kirkland: as it will create you a new .orig.tar.gz that won't be the same, and so will be rejected by soyuz/dak
[19:02] <kirkland> james_w: hmm, i expected the diff.gz to contain the debian dir
[19:02] <james_w> kirkland: yes, it will
[19:02] <james_w> kirkland: but with .tar.gz the contents of the files is not all that matters
[19:03] <james_w> kirkland: even if you don't touch anything outside of ./debian/ in -2 the md5sum of the .orig.tar.gz would still change
[19:04] <LaserJock> hence why pristine-tar exists for VCSes
[19:08] <james_w> kirkland: so, the immediate issue is that it is not running get-orig-source in a directory that is named screen-profiles or screen-profiles-${VER}, so it falls over.
[19:08] <james_w> kirkland: I could fix that, but I don't think there is a lot of point
[19:10] <james_w> kirkland: as http://www.debian.org/doc/debian-policy/ch-source.html#s-debianrules states "get-orig-source (optional) [...] This target may be invoked in any directory"
[19:14] <james_w> kirkland: so I think you should re-think your get-orig-source. I can help you come up with a good solution that works for you if you like, but first I must eat...
[19:14] <kirkland> james_w: please, you bet
[19:15] <kirkland> james_w: i've recently gotten a lot of requests for other distros to pick up this package
[19:15] <kirkland> james_w: i'm trying to rework my build/release procedure such that it's still smooth for me, but also works for them
[19:15] <kirkland> james_w: siretart_ is handling the ITP for debian
[19:15] <kirkland> james_w: he's the first consumer I'm trying to satisfy
[19:16] <Chipzz> why do you package those seperately anyway?
[19:16] <Chipzz> wouldn't it make more sense to ship these in /usr/share/doc/screen in the screen package?
[19:19] <GuyFromHell> Anyone here using an update XChat and some form of galago notifications? Are you getting an icon on your popups?
[19:20] <maxb> Chipzz: Personally I think it's far more elegant to put a loosely coupled addon in its own package
[19:20] <festor>  I use ubuntu beta 9.04 (now in livecd) and I dont see any ext4 option in ubuntu installer
[19:37] <mcas> hi
[19:37] <mcas> can anyone tell when there will be the next translation sync?
[19:41] <cjwatson> mcas: I saw language-pack builds going through today ...
[19:41] <mcas> ok thx cjwatson
[19:52] <james_w> GuyFromHell: I am, and yes, the stock_person icon if I am not mistaken
[19:53] <GuyFromHell> james_w, may i inquire specs? distro, version, and galago program?
[19:54] <james_w> ubuntu jaunty with notify-osd
[19:54] <james_w> xchat 2.8.6-2.1ubuntu4, notify-osd 0.9.8-0ubuntu1
[19:54] <GuyFromHell> james_w, interesting, mine doesn't work on same configuration unless the string is changed from "notification-message-im" to "notification-message-IM"
[19:54] <james_w> are you using the xchat-gnome variant?
[19:55] <GuyFromHell> james_w, no, the original
[19:55] <james_w> that changes the icons that you get?
[19:55] <GuyFromHell> james_w, my notify-osd seems to be newer, perhaps they added case sensitivity...
[19:56] <GuyFromHell> james_w, i had no icon from the deb install
[19:56] <james_w> ah, it's not stock_person I don't think
[19:58] <GuyFromHell> james_w, i think this is that was meant to be: http://www.cs.rpi.edu/~mukhea2/stock.png
[19:59] <GuyFromHell> is that what you're getting?
[19:59] <james_w> GuyFromHell: I assume you have libnotify installed?
[19:59] <james_w> GuyFromHell: it's similar, but I can't be sure
[20:00] <GuyFromHell> james_w, that's what i have after putting  a patch on x-chat to capitalize 'im' in 'notification-message-im'
[20:01]  * GuyFromHell debates with himself on what to do
[20:01] <GuyFromHell> i guess I'll file a bug report and see if anyone else is getting this behavior...
[20:01] <james_w> GuyFromHell: http://people.ubuntu.com/~jamesw/Screenshot.png
[20:02] <clearscreen> Not sure if this is the right place to ask.. but I'm a new C++ programmer looking for a new open-source project with a relatively small codebase / small group of developers.. any suggestions? I browsed through countless of sourceforge pages but either wasnt interested, it was a huge project, or it was completely inactive (not a single commit)
[20:03] <GuyFromHell> james_w, it may just be my icon theme, lemme see if it supplies this, i thought the icon was part of human only though :/
[20:03] <Ampelbein> clearscreen: try the gnome-project. http://live.gnome.org/GnomeLove There are simple tasks to start and you can get in touch with an active community.
[20:06] <clearscreen> Ampelbein: Problem is that I do want to be hands-on with the code, but am not even closed to experienced enough to deal with something like gnome.. Would probably be clueless about a lot of things and produce crap :P
[20:07] <james_w> GuyFromHell: got it
[20:07] <james_w> GuyFromHell: the fallback icon for non-Human is capitalised, which is probably a mistake
[20:08] <james_w> GuyFromHell: so it seems a bug against notify-osd to make it lower-case is the right thing
[20:08] <GuyFromHell> james_w, so asking for icon -im falls back to -IM?
[20:08] <Ampelbein> clearscreen: thats fine. the folks at gnome are mostly helpful and will point out what is wrong/could have been better. check out the simple tasks listed on the site and you will be learning-by-doing. GNOME is a big project, but there are many small programs which you could start working on.
[20:09] <james_w> GuyFromHell: no, it doesn't. xchat asks for -im and if you are running human you get it. If you aren't running human then it tries to use the fallback, but doesn't exist
[20:09] <james_w> GuyFromHell: your capitalisation change would mean that you always get the fallback whether you run Human or not, which would be undesired I guess. So making the fallback the same as what xchat is requesting and what human provides would seem to be the solution
[20:10] <GuyFromHell> james_w, ah i see what you're saying now.
[20:10] <GuyFromHell> james_w, i'll go ask #dx what they think about this
[20:11] <james_w> sure
[20:11] <kees> Keybuk: I see no reason to keep /dev/kmem if it's disabled by default in our kernels.
[20:11] <slytherin> Can any of the core dev's take a look at bug 349146 and let me know if this is intentional.
[20:12] <kees> Keybuk: I have no opinion about doing an open() test on the special device to see if it should be created.  on the one hand, it'd be nice to play with a kernel that has it enabled for some reason.  on the other hand, no one should use kmem.
[20:13] <clearscreen> Ampelbein: Mmmm I'm not so sure, I'll keep looking around :P
[20:18] <james_w> kirkland: hey, so I assume you set things up like that originally so that you only had a single branch to worry about?
[20:23] <Laney> Keybuk: Do you have a minute to look at an automake problem? (you did the update). The smuxi package now fails to build (http://is.gd/r3I3) with the error "shift: 2219: can't shift that many". I've confirmed that downgrading to the older version renders it buildable again.
[20:24] <Laney> I don't know if this is a regression or a bug in upstream's usage or what
[20:38] <adelie42> If I am starting a new project and want it hosted on launchpad, I just register a branch, correct?
[20:39] <jpds> adelie42: And the project, best ask Launchpad questions at #launchpad
[20:39] <adelie42> k
[21:00] <GuyFromHell> james_w, Just in case you were wondering, a bug has now been filed against notify-osd to rename the icon.
[21:01] <james_w> I saw, thanks
[21:01] <GuyFromHell> james_w, oh didn't see you were in #dx too
[21:09] <johanbr> clearscreen: ekiga is c++ and they're short on developers
[21:15] <sbeattie> tedg: are you no longer maintaining gnome-screensaver's packaging in bzr+ssh://bazaar.launchpad.net/~ted-gould/gnome-screensaver/ubuntu-packaging ?
[21:16] <tedg> sbeattie: I think someone may have moved it over to ~ubuntu-desktop.  I know they were discussing doing that.
[21:16] <sbeattie> okay, thanks, the vcs tag for it is wrong, then.
[21:17] <tedg> sbeattie: Yup, here it is: lp:~ubuntu-desktop/gnome-screensaver/ubuntu
[21:17] <sbeattie> tedg: cool, thanks.
[21:27] <picklesworth> Hm... are the DX team people aware of this bug report?: https://bugs.launchpad.net/bugs/343219
[21:28] <jpds> That's intended bahaviour.
[21:28] <picklesworth> indeed, and it's also wrong behaviour ;)
[21:28] <jpds> If you remove fusa, you get the System menu back, one or the other.
[21:32] <Laney> There is a regression in the version of automake that's currently in Jaunty which is rendering at least one package unbuildable. I've almost confirmed that http://git.savannah.gnu.org/cgit/automake.git/diff/?h=branch-1-10&id=f9a3dde93cbd6497966d45631ec1cf665b09e3a9 fixes it
[21:32] <Laney> I don't know how to work up a patch for a git-maintained package though
[21:33] <dtchen> git format-patch
[21:38] <Keybuk> gnargh, apport is popping up for valgrind again
[21:38] <Laney> bluh
[21:38] <slytherin> any maintainer of human-icon-theme here?
[21:39] <Laney> Keybuk: You forked automake! What's the best way to patch it now?
[21:39] <jpds> slytherin: kwwii.
[21:39] <Laney> the git repo doesn't appear to be current
[21:39] <cjwatson> Laney: apt-get source automake, hack, upload?
[21:39] <cjwatson> Laney: the git repository is the Debian package, by the looks of things
[21:40] <slytherin> jpds: is it just me who find logout icon more like shutdown?
[21:41] <Keybuk> Laney: ?
[21:41] <clearscreen> slytherin: its a man walking about.. looks like logging out to me :P
[21:42] <Keybuk> Laney: what do you mean?
[21:42] <slytherin> clearscreen: are you using human theme?
[21:42] <clearscreen> slytherin: if that's the default theme for 9.04, then yes.. cant remember changing it
[21:42] <clearscreen> I could be wrong though
[21:42] <Laney> Keybuk: Should I add a patchsys or what?
[21:43] <Laney> cjwatson: Right
[21:43] <Keybuk> Laney: I updated our package to a new upstream version, is that what you mean by "forked" ?
[21:43] <Laney> Keybuk: yes
[21:43] <cjwatson> Laney: please don't ever add a patch system in Ubuntu to a package (from Debian) that doesn't have one
[21:43] <Keybuk> Laney: never add a patchsys to a package that doesn't have one
[21:43] <Laney> right
[21:43] <Keybuk> Laney: that's hardly a "fork" - that's a relatively standard practice
[21:43] <clearscreen> slytherin: nevermind :)
[21:43] <slytherin> clearscreen: You can verify in system -> preferences -> appearance
[21:44] <Laney> excuse my terminology then
[21:44] <Keybuk> laney: I used the upstream orig.tar.gz, so when Debian update too, we simply drop out diff
[21:44] <clearscreen> slytherin: yeah, using glossy
[21:44] <Laney> I didn't mean ti imply anything negative
[21:44] <Laney> o
[21:44] <Keybuk> Laney: then I'm unsure what your point is?
[21:44] <Keybuk> Laney: perhaps you should rephrase your exclamation as a question?
[21:44] <Laney> I wanted to know in which format the patch should come
[21:44] <Laney> is all
[21:44] <Keybuk> Laney: what patch?
[21:44] <jcole> didrocks: according to the evolution developers, it looks like your evoltion-mapi package is very out of date
[21:44] <Laney> the one I'm about to supply
[21:44] <Keybuk> Laney: and what does that do?
[21:44] <Keybuk> why haven't you send that patch upstream instead?
[21:45] <Laney> it is upstream, I'm backporting it
[21:45] <Keybuk> is there a new upstream release containing the patch?
[21:45] <Laney> it's a regression vs 1.10.1
[21:45] <Keybuk> Laney: ref?
[21:45] <Laney> no
[21:45] <Laney> http://git.savannah.gnu.org/cgit/automake.git/diff/?h=branch-1-10&id=f9a3dde93cbd6497966d45631ec1cf665b09e3a9
[21:45] <cjwatson> Keybuk: to be fair, he gave a reference above
[21:46] <cjwatson> 21:32 <Laney> There is a regression in the version of automake that's currently in Jaunty which is rendering at least one package unbuildable. I've almost confirmed that
[21:46] <cjwatson>               http://git.savannah.gnu.org/cgit/automake.git/diff/?h=branch-1-10&id=f9a3dde93cbd6497966d45631ec1cf665b09e3a9 fixes it
[21:46] <cjwatson> 21:32 <Laney> I don't know how to work up a patch for a git-maintained package though
[21:46] <jcole> didrocks: this may be why so many people are having issues
[21:46] <Keybuk> ah, I missed that bit
[21:46] <Keybuk> shaky network here
[21:46] <Laney> no worries
[21:46] <Laney> just want to ensure it gets in
[21:46] <Keybuk> Laney: just apply the patch directly to the sources, add a changelog entry (-0ubuntu2) and upload
[21:46] <jpds> slytherin: re: logout icon: true...
[21:47] <Keybuk> Laney: obviously there should be a Launchpad bug for this, targeted to jaunty, etc.
[21:47] <Laney> ok
[21:47] <Laney> debdiff following shortly
[21:47] <kees> Keybuk: did you see my backscroll on my okayness with dropping /dev/kmem?
[21:47] <Laney> watch this space etc
[21:47] <Keybuk> kees: no, I don't keep backscroll
[21:47] <Keybuk> kees: I saw your e-mail about that
[21:47] <kees> 19:11 < kees> Keybuk: I see no reason to keep /dev/kmem if it's disabled by default in our kernels.
[21:47] <kees> 19:12 < kees> Keybuk: I have no opinion about doing an open() test on the special device to see if it should be created.  on the one hand, it'd be
[21:47] <kees>               nice to play with a kernel that has it enabled for some reason.  on the other hand, no one should use kmem.
[21:47] <slytherin> jpds: I filed bug 349146, but looks like it has not come to attention of the maintainers.
[21:48] <Keybuk> kees: well, it'll be created automatically if the kernel supports it
[21:48] <Keybuk> since the kernel has a /sys/devices/virtual/mem/kmem or something
[21:48] <kees> Keybuk: works for me
[21:48] <Keybuk> we just force it anyway
[21:48]  * Keybuk can't remember why
[21:48] <Keybuk> probably because devices.txt said we should
[21:49] <Keybuk> oh, hmm
[21:49] <Keybuk>   * Add /lib/udev/devices/kmem as there's no sysfs node for it yet.
[21:49] <kees> Keybuk: yeah, it's a legacy device at this point.
[21:49] <Keybuk> wonder if that's been fixed
[21:49] <kees> Keybuk: it has (intrepid has it)
[21:49] <kees> dapper and gutsy don't.  hardy has kmem disabled so I can't check
[21:51] <Keybuk> ah yes, it uses device_create()
[21:51] <Keybuk> that's ok then
[22:11] <calc> what type of threading issue would running strace on a process cause not to happen?
[22:23] <cjwatson> calc: anything that's a race condition
[22:23] <Laney> bug 356612 for the automake patch
[22:23] <cjwatson> (I wouldn't put it past there to be other threading problems that strace might mask, but races are the standard answer to this question and threaded code is full of them if written even slightly incautiously)
[22:26] <calc> cjwatson: ok
[22:27] <seb128> calc: that's probably not the issue there though, it's just a function calling xrandr, there is no possible race since it doesn't depend of any other code
[22:29] <calc> seb128: its failing inside of spawn_with_input afaict and doesn't fail if i strace it
[22:30] <calc> spawn_with_input seems to just spawn a thread and and execute the program
[22:31] <seb128> not something which should be subject to races
[22:31] <calc> when i straced i did strace -f -p pid -o output
[22:31] <calc> not sure if that could mask any other types of issues though
[22:33]  * calc looks at OOo 3.1.0 rc1 debian orig size and dies... 764MB just for orig
[22:45]  * picklesworth looks at calc's message... "but how did he write that?" "...he must have been dictating"
[22:47] <calc> picklesworth: heh :)
[22:48] <calc> picklesworth: so 1.5GB+ for OOo source in karmic if it isn't a bug in debian packaging :\
[22:48] <directhex> calc, 764 meg orig? so... we're breaking the 10 gig pbuilder requirement then
[22:48] <directhex> ?
[22:48] <calc> directhex: i have no idea what is going on, i'm still downloading it, i think rene might have screwed up the orig.tar.gz
[22:49] <calc> i can't imagine how it would have gotten that large between a couple of OOo milestone releases
[22:49] <calc> it was only 400MB at 310m6
[22:49] <picklesworth> OO is an impressively complicated office suite if it's not an error
[22:50] <picklesworth> maybe he ran a build in there :/
[22:50] <geofft> You say that like OOo can't be both an error and an impressively complicated office suite...
[22:51]  * jdong wishes update-manager used PolicyKit...
[22:52] <glatzor> jdong, https://edge.launchpad.net/aptdaemon
[22:53] <jdong> ooh
[23:02] <laszlok_> is it known that jockey-gtk doesn't work? maybe this is part of the python 2.6 thing in the title?
[23:04] <laszlok_> that python bug is marked as fix released though...
[23:05] <directhex> "doesn't work"? is claiming unemployment benefits?
[23:05] <bryce> directhex: :-)
[23:07] <clearscreen> directhex: i loled
[23:11] <hyperair> say... can someone test what happens when you run this in a shell: notify-send title 'message&'
[23:11] <hyperair> for some reason, the message gets substituted with something else
[23:11] <hyperair> is that supposed to happen?
[23:12] <james_w> no
[23:13] <directhex> "&" generally stuffs it
[23:13] <hyperair> yeah
[23:13] <hyperair> but is this a bug in notify-osd or notification-daemon?
[23:13] <directhex> using '&' in title means no message at all
[23:14] <jdong> well you're not supposed to use unescaped HTML chars, right?
[23:15] <hyperair> oh so it's supposed to be &amp;?
[23:15] <directhex> seems so - but still broken in title
[23:16] <directhex> hm, no, broken if &amp; is the ENTIRE title
[23:16] <directhex> or... hm, it wasn't working, now it is
[23:16] <directhex> bloody technology
[23:16] <hyperair> no i think notify-osd might have crashed
[23:16] <hyperair> remember that dbus automatically starts it again
[23:17] <hyperair> [51462.618299] notify-osd[5512]: segfault at 18 ip b7698cc3 sp bf841a24 error 4 in libc-2.8.90.so[b7629000+158000]
[23:17] <hyperair> [51787.162592] notify-osd[5960]: segfault at 121c61f8 ip b76e9cd9 sp bf992b74 error 4 in libc-2.8.90.so[b767a000+158000]
[23:17] <hyperair> see.
[23:17] <jdong> hyperair: holy crap. is that from an ampersand?
[23:17] <hyperair> jdong: yes it is
[23:17] <hyperair> jdong: unescaped
[23:17] <jdong> hyperair: ok then you might be onto something more serious.
[23:18] <jdong> hyperair: I'd file it as a potential security issue at this point.
[23:18] <hyperair> hmm
[23:18] <hyperair> wait, i can't reproduce it
[23:18] <hyperair> it's not re-segfaulting
[23:18] <jdong> hmm
[23:18] <cjwatson> run it under valgrind
[23:20] <hyperair> meh i can't reproduce it anymore
[23:20] <hyperair> wonder what went on
[23:21] <kees> hyperair: the body portion goes through pangomarkup, but I wasn't able to crash it...
[23:21] <hyperair> hmm
[23:22] <hyperair> so with regards to this whole pangomarkup thing... are the rules the same as html for symbols?
[23:22] <kees> hyperair: it's a subset of html
[23:22] <hyperair> subset?
[23:22] <kees> hyperair: I was poking at it last week, but it seemed okay
[23:22] <hyperair> hmm
[23:22] <kees> http://library.gnome.org/devel/pango/unstable/PangoMarkupFormat.html
[23:23] <hyperair> hmm i'm more interested in the symbols
[23:26]  * hyperair now has a should-be-perfect irssi libnotify script =D
[23:27] <jdong> hyperair: are ' you& su<a href=localhost>re</a>?"!;
[23:27] <jdong> *ducks*
[23:27] <cjwatson> I was thinking that irssi might be better hooked into libindicate
[23:27] <cjwatson> although there don't seem to be shell bindings for that
[23:27] <hyperair> jdong: sorry i didn't manage to see the notification. could you try again?
[23:27] <hyperair> cjwatson: if there are perl bindings, it'll work.
[23:28] <jdong> hyperair: test normal.....
[23:28] <hyperair> cjwatson: but there aren't perl bindings either
[23:28] <jdong> hyperair: test nasty are ' you& su<a href=localhost>re</a>?"!;
[23:28] <hyperair> jdong: works perfectly =)
[23:28] <jdong> hyperair: just making sure you wrote good perl :)
[23:28] <hyperair> you wouldn't happen to be interested in the script would you?
[23:28] <jdong> hyperair: some of the libnotify scripts floating out there are shell escapable.
[23:28] <jdong> hyperair: putting the script up somewhere would be nice for the world, yeah :)
[23:29] <hyperair> jdong: ah i used the multi-arg system() notation.
[23:29] <cjwatson> I put mine up a while ago, BTW
[23:29] <hyperair> cjwatson: where?
[23:29] <hyperair> cjwatson: it isn't on scripts.irssi.org is it =p
[23:29] <jdong> hyperair: good boy.
[23:29] <cjwatson> http://www.chiark.greenend.org.uk/~cjwatson/code/notifications/
[23:30] <hyperair> cjwatson: that requires a file
[23:30] <cjwatson> yes, it does. deal.
[23:30] <hyperair> hahah
[23:30] <hyperair> and a daemon =p
[23:30] <cjwatson> ditto
[23:30] <jdong> wow. that's really good.
[23:30] <jdong> I had a maildir-ish spool setup before I redid my shell server
[23:30] <hyperair> +
[23:31] <hyperair> also mine are based on the same script i think
[23:31] <hyperair> heh
[23:31] <jdong> yeah, fnotify was a good starting point for me too
[23:31] <cjwatson> it's not really a proper daemon. I doubt you could do without something like that and still work over ssh, anyway.
[23:32] <cjwatson> my laptop isn't necessarily on even remotely the same network as the server on which I run irssi
[23:32] <jdong> my issue was I needed the notion of a notification being "acknowledged"
[23:32] <hyperair> cjwatson: agreed. without ssh-ing back, there really isn't a way
[23:32] <cjwatson> specifically the server can't necessarily open an ssh connection back to it
[23:32] <jdong> much like how with an IMAP mailbox I know when I've read the message elsewhere
[23:32] <hyperair> that's one of the reasons i'm running irssi locally
[23:32] <cjwatson> I suppose I could do it with port forwarding
[23:32] <jdong> I sometimes have 4 of 5 systems logged into IRC, some abandoned....
[23:33] <hyperair> cjwatson: why not a fifo? eventually those hilights will build up you know
[23:33] <cjwatson> which might be a bit simpler 'cos you wouldn't need the heartbeat
[23:33] <cjwatson> hyperair: in practice it seems fine
[23:33] <hyperair> hmm
[23:33] <jdong> hyperair: it's gotta build up for years before that matters though
[23:33] <cjwatson> I don't want to use a fifo because that might cause irssi to block if the server isn't running
[23:33] <hyperair> jdong: good point =p
[23:33] <hyperair> ah yes
[23:34] <hyperair> fork! then you get a mini fork bomb =p
[23:34] <cjwatson> the wrong solution to so many problems ;-)
[23:34] <hyperair> =p
[23:34] <hyperair> exec then.
[23:35] <hyperair> Irssi::command("exec echo msg > fifo");
[23:35] <cjwatson> doesn't interest me
[23:35] <hyperair> hehe
[23:36] <cjwatson> indeed, that'll just lead to a bunch of sh processes building up over time if the fifo is full
[23:36] <cjwatson> don't see how that's an improvement really
[23:37] <hyperair> fork bomb =p
[23:41] <james_w> cjwatson: I was just thinking about how shell binding for libindicate would work, as it's not async like libnotify
[23:41] <james_w> you have to maintain the dbus connection for the length of the indicator
[23:42] <cjwatson> james_w: oh, interesting, you'd need a little session daemon to do it then
[23:42] <james_w> yeah
[23:42] <james_w> you could even do it over dbus
[23:59] <cjwatson> slangasek: feel free to downgrade the universe test rebuild bugs I'm filing if you feel so inclined
[23:59] <cjwatson> I didn't get much output from the test rebuild since it turned out to spam people by mail, but I'm filing what we got