[00:11] <calc> are we going to have ext4 support in intrepid?
[00:11]  * calc notices Fedora9 has preliminary support already
[00:11] <mjg59> calc: Until there are guarantees about the on-disk format being stable, I doubt it
[00:12] <calc> mjg59: oh so the 2.6.25 merge wasn't the final on-disk format?
[00:13] <jdong> mjg59: I roughly read that the last on-disk format  changes are being merged into .25?
[00:13] <jdong> heh he beat me to it
[00:13] <calc> well 2.6.25 just was released so i suppose they don't really know for certain yet ;)
[00:14] <mjg59> jdong: The last planned ones
[00:14] <mjg59> Which doesn't mean there are none
[00:15] <mjg59> It still registers as ext4dev, rather than ext4
[00:15] <jdong> mjg59: ah, I see :)
[00:15] <calc> oh ok
[00:15] <jdong> mjg59: do you follow ext4 development personally?
[00:15] <mjg59> Not heavily
[00:15] <jdong> cool
[06:41] <dholbach> good morning
[06:41] <emgent> morning
[06:41] <dholbach> hi emgent
[06:42] <emgent> heya dholbach, all ready for intrepid UDS ? :)
[06:42] <dholbach> yeah :-)
[06:43] <emgent> go intrepid go heheh
[06:43] <dholbach> ROCK :-))
[06:44] <emgent> dholbach: saw http://www.seqfault.de/blog/index.php?/archives/87-Einfach-mal-Danke-sagen.html ?
[06:45] <dholbach> yeah - that's awesome
[06:45]  * jdong still awaits his t-shirt with that :D
[06:45] <emgent> hehehe
[06:45] <emgent> heya jdong
[06:45] <emgent> cjwatson o/
[06:45] <jdong> yo
[06:45] <dholbach> jdong: nice
[06:47] <cjwatson> morning
[06:48] <jdong> ok when you people start waking up it's a sign I need to start sleeping.
[06:49]  * emgent hungry
[06:59] <pitti> Good morning
[06:59] <emgent> heya pitti
[06:59] <tonyyarusso> Say, anyone know when repos will be opening?  (approx.)
[07:00] <pitti> Riddell: good, I'll process the queue now every morning anyway :)
[07:00] <tonyyarusso> jdong: Same :)
[07:00] <ogra> emgent, !
[07:00] <ogra> emgent, impressing
[07:00]  * tonyyarusso is amused that he's on the list of people in that pic, and therefore would consider buying such a t-shirt if it existed
[07:01] <zul> tonyyarusso: im guessing when its ready :)
[07:01] <tonyyarusso> hehe, likely
[07:03] <emgent> ogra: ?
[07:03] <pwnguin> jdong: the more curious question is, where do these people who sleep normal hours come from?
[07:04] <ogra> emgent, the link :)
[07:04] <emgent> oh lol :)
[07:04] <ogra> :)
[07:06]  * bimberi sees his name and dances a little jig
[07:07] <bimberi> right in the middle of "Thank" :)
[07:23] <cjwatson> tonyyarusso: matter of days
[07:23] <cjwatson> tonyyarusso: the toolchain is well in progress
[07:41] <tonyyarusso> cjwatson: ah, cool
[07:54] <bryce> slangasek: question for you - for 8.04.1 would updates to the binary drivers -nvidia and -fglrx be possible, or is that too much risk?
[07:54] <bryce> (I don't have a strong opinion, but if it's not feasible, I'll un-milestone bugs that would require such updates)
[07:58] <cjwatson> we're in discussions with AMD about fglrx, for the record; I asked in the last call if we might be able to push back on intrusive changes and get a minimally-branched version, though haven't had an answer on that yet
[07:58] <cjwatson> (our discussions with AMD are unfortunately NDAed)
[07:58] <bryce> cjwatson: ok we can talk more on it tomorrow
[07:59] <cjwatson> oh, meh, yeah, we have a call this afternoon
[07:59] <bryce> cjwatson: I guess EnvyNG is also available as an option now
[07:59] <cjwatson> bryce: anything special for it?
[07:59] <cjwatson> (the call, that is)
[07:59] <bryce> cjwatson: not from my end
[08:00] <bryce> cjwatson: I think their engineers have "can't reproduce"'d the five bugs I sent to them
[08:01] <slangasek> bryce: IMHO yes, binary driver updates are something we need to be able to accomodate in the SRU policy elsewise so there's no reason fglrx/nvidia shouldn't be candidates so long as we get reasonable SRU verification
[08:02] <slangasek> s/binary driver/black box executable/
[08:02] <bryce> slangasek: ok good to know.
[08:03] <pitti> bryce: BTW, Alberto prepared some l-r-m-envy packages, I'll take a look at them today
[08:04] <pitti> bryce: bug 221304 FYI
[08:05] <bryce> pitti: excellent
[08:05] <pitti> bryce: we should also think about backporting some new jockey features to hardy once they land in intrepid
[08:06] <pitti> bryce: like proposing new drivers for pinpointed hardware (not generally upgrading fglrx for everyone)
[08:06] <pitti> bryce: but that's way post-UDS
[08:06]  * bryce nods
[08:07] <bryce> pitti: being able to install one version of -fglrx for some hardware and a different version for other hardware would solve one or two wishlist bugs too
[08:20] <pitti> argh, hardy isn't released for a week, and already the first ABI bump in -proposed? *sob*
[08:21] <Hobbsee> haha
[08:21] <Hobbsee> i'ts the crack.
[08:21]  * tonyyarusso guesses there will be more changes than usual between now and 8.04.1
[08:22] <pitti> tonyyarusso: yes, absolutely, since we throw a lot of resources at it
[08:22] <tonyyarusso> Right, and the fact that there are certain things (ie, Firefox) that are a big deal and will certainly need them.
[08:23] <slangasek> pitti: I think our choices were to have it the week after release, or the week of RC ;)
[08:24] <pitti> slangasek: BTW, would you have a chance to look at the hardy-proposed jockey upload? it currently breaks X for people without an xorg.conf
[08:25] <slangasek> pitti: "it currently" -> "the version in hardy"?
[08:25] <pitti> slangasek: yes
[08:26] <seb128> tonyyarusso: do you have specific firefox issues in hardy?
[08:26] <slangasek> pitti: ok - will look, yes
[08:26] <pitti> slangasek: thanks; I mailed you about it with some details yesterday
[08:26] <pitti> slangasek: I just wasn't sure whom to bother for my own SRUs
[08:27] <tonyyarusso> seb128: One minor one that I notice most of the time, but I was under the impression that the point release will include the FF3 final - is that incorrect?
[08:27] <pitti> slangasek: (erm, s/whom/who/, right?)
[08:27] <slangasek> whom is correct :)
[08:27] <pitti> argh, Launchpad, where are thou??
[08:28] <seb128> tonyyarusso: it'll be considered yes, can't say for sure until saying what they actually change and if that's suitable for a stable update, but that should be alright
[08:28] <seb128> s/saying/seeing
[08:28] <seb128> tonyyarusso: but you say that's a big deal, so you have real issues right now?
[08:29] <cjwatson> seb128: the URL classifier is a fairly major one, I think
[08:30] <tonyyarusso> seb128: makes sense.  As far as current issues, the most prominent throughout the cycle has been the some images showing black at certain resizings.  Haven't tested in FF since release though, only Ephy.
[08:30] <seb128> tonyyarusso: those were due to xorg and should be fixed in hardy
[08:31] <seb128> cjwatson: not denying there is some issues, I was just curious to know if tonyyarusso has some specifics one or if the comment was just about hardy having a candidate version rather than the official stable one
[08:31]  * StevenK thinks of Gimp in Gutsy
[08:32] <seb128> StevenK: you have been marked by this one apparenly ;-)
[08:32] <tonyyarusso> seb128: hmm, had one earlier today in ephy, so if it's xorg I would think that had gone away.  I'm not going to pretend my testing has been sufficient to say anything for sure though, and my environment is far from clean, given that it's been continuously upgraded since the repos opened.  I note that https://bugs.edge.launchpad.net/xorg-server/+bug/182038 does indeed claim fix released.
[08:32] <StevenK> seb128: There's no apparently about it ;-)
[08:32] <seb128> heh
[08:33] <tonyyarusso> That "is it ready for my girlfriend" posted raised a handful of legitimate points as well.
[08:34] <pitti> does launchpad.net work for other people? it's just stuck for me
[08:34] <seb128> pitti: works for me but I'm using edge
[08:34] <stgraber> pitti: edge WFM
[08:34] <seb128> let me try the normal one
[08:34] <pitti> right, in fact it's edge that's stuck for me; https://launchpad.net WFM
[08:34] <pitti> seb128: nevermind, edge it is
[08:34] <seb128> pitti: yes, works fine
[08:34] <pitti> hmm
[08:34]  * pitti blames his ISP them
[08:35] <tonyyarusso> pitti: WFM too.
[08:35] <StevenK> Argh, more langpacks
[08:35] <pitti> ah, seems it came back now
[08:35] <Hobbsee> heh
[08:35] <Hobbsee> StevenK: what do you want in front of them?
[08:35] <Hobbsee> ubottu: bug 5
[08:35] <Hobbsee> bug 6
[08:35] <Hobbsee> oh good.
[08:35] <Hobbsee> it mostly works, then
[08:35] <StevenK> Hobbsee: python-central in the ubuntu-mobile PPA
[08:36] <ogra> pitti, you had an outage too ?
[08:36] <pitti> ogra: yes
[08:36] <ogra> seems somewhere in the telekom network a router fell over
[08:36] <ogra> i cant reach canonical irc atm
[08:36] <tonyyarusso> Say, what would be the "proper" way of drumming up support for a Blueprint if it's been pretty much ignored so far?
[08:37]  * Hobbsee fiddles, flicks a few switches
[08:38] <StevenK> Hobbsee: thallium has picked it up, thanks
[08:38] <StevenK> (thallium is an element? Blink)
[08:38] <ogra> tonyyarusso, send money to people :)
[08:38] <ogra> tonyyarusso, beer is also a ood bribe
[08:39] <ogra> *good
[08:39] <tonyyarusso> ogra: You know, I could do that, although it would only be a token bit of funding, not a fair rate for their time.
[08:40] <Hobbsee> StevenK: your python-central is served.
[08:40] <Hobbsee> StevenK: enjoy
[08:40] <tonyyarusso> I'm told telepathy (which wasn't really mainstream when I first had the thought) should make it a lot easier.
[09:02] <ogra> dear deutsche telekom, please fix your network !
[09:02] <stgraber> :)
[09:02] <ogra> *sigh*
[09:03] <ogra> firt the routing was f*cked ... now the DNS is gone
[09:03] <ogra> *first
[09:06] <stgraber> ogra: use opendns, those are fast and usually reliable
[09:07] <ogra> raaaaah
[09:07] <Company> the trick is to not use telekom
[09:08] <Company> and to not use arcor
[09:08] <ogra> did anyone ever try to paste an IP in the xchat serverlist ?
[09:08] <ogra> its impossible by design
[09:08] <ogra> unbelivable
[09:08] <ogra> (as soon as you click an entry it automatically gets the selection and wipes the paste buffer)
[09:10] <stgraber> ogra: use irssi, this one let you connect to IPs :)
[09:10] <\sh> ogra, don't you have any cable tv provider next to your house?
[09:10] <ogra> Company, i have a buisness SDSL line, there is no other provider offering that at a feasable price
[09:11] <\sh> ogra, and xchat works here with an ip address ,-)
[09:11] <ogra> \sh, i dont get the upload rates i need at iesy
[09:11] <\sh> ogra, sdsl? qsc?
[09:11] <ogra> and i didnt mean to use an ip
[09:11] <ogra> i men to copy paste an ip in the server list instead of the server name
[09:12] <ogra> thats impossible
[09:12] <ogra> you have to type it in any case
[09:12] <\sh> well, xchat is evil anyways ,-)
[09:12] <ogra> since selecting the entry will wipe your paste buffer
[09:12] <\sh> ogra, more funny, you enter anything into the serverlist, and if you don't press <tab> key, and click on "ok" it throws away the change
[09:13] <ogra> \sh, i have a 2 year contract with telekom with guaranteed uptimes and direct support (no call center)
[09:13] <ogra> i.e. a business line
[09:13] <\sh> ogra, but telekom...
[09:13] <ogra> \sh, yeah, thats another oddity
[09:13] <ogra> (serverlist)
[09:14]  * \sh ports kmyirc to gnome and python...so everything'll be fixed in the future ,-)
[09:15] <ogra> heh
[09:15] <ogra> well, i usually dont have probs with xchat
[09:15] <ogra> since i rarely change the config :)
[09:19]  * \sh fights with bacula...
[09:19] <ogra> Spads, are there known probs with irc.c.c ? it kicks me out all the time after some seconds
[09:20] <Spads> ogra: no known problems I can see
[09:20] <ogra> hmm, ok, must be my side then
[09:20] <Spads> anyway this is a question for #canonical-sysadmins, not here
[09:21] <ogra> Spads, oh, sorry
[09:21] <Spads> ogra: that said, I'll just point out that I'm having some net connection problems here at home, so it may be that there's some general routing problems in europe right now
[09:22] <ogra> yeah
[09:22] <ogra> i had some before already
[09:22] <ogra> then DNS didnt resolve anything outside of germany ...
[09:22] <ogra> now it looks like its back but i get kicked out all the time ... probably not over yet
[09:23]  * ogra wil just wait
[09:23] <ogra> *will
[09:23] <Spads> yeah, I'm getting *terrible* packet loss right out of the gate here
[09:29] <cjwatson> mvo: noticed in python-apt:
[09:29] <cjwatson>     locale = os.getenv("LANG", default="en.UK")
[09:29] <cjwatson> mvo: shouldn't that be en_GB? UK isn't a valid country code
[09:30] <cjwatson> (not that it matters much, since most people will have LANG set)
[09:31] <mvo> cjwatson: that looks wrong, let me check that
[09:33] <cjwatson> mvo: discovered while trying to figure out where bug 221644 lived
[09:35] <mvo> cjwatson: I was wondering why you digged into those dusty corners in pyhton-apt ;) the mirror list ist just a file, if the mirror is dead, I can remove it from the list and prepare a sru
[09:37] <bolsh> mjg59: I've been following DebuggingGNOMEPowerManager - a nice read
[09:37] <bolsh> Sending the message directly to DBus, things work fine
[09:38] <bolsh> They also work fine via the "Quit" menu item, and the "Suspend" button now
[09:38] <bolsh> I did a recursive-reset of GPM settings to make sure I wasn't using some old settings
[09:38] <bolsh> Closing the lid still has the same problem
[09:39] <bolsh> I don't understand
[09:39] <cjwatson> mvo: how do you keep it up to date? ideally I'd have a full list in choose-mirror too
[09:39] <mvo> cjwatson: I have a small script that gets them from launchpad, give me a sec, I dig it out
[09:40] <cjwatson> if you could mail it to me that'd be lovely
[09:40] <mvo> sure
[09:42] <bolsh> Anyone know a magical dbus-send incantation to turn on a backlight?
[09:56] <fs_> hello
[09:56] <fs_> i need some help with the ubuntu kernel build proccess
[09:56] <fs_> am i at the right place here?
[09:57] <dholbach> fs_: try #ubuntu-kernel
[09:57] <ogra> fs_, #ubuntu-kernel might be better
[09:57] <fs_> ok
[10:07] <pitti> doko: any idea about http://launchpadlibrarian.net/14036302/buildlog_ubuntu-intrepid-i386.cdbs_0.4.52ubuntu1_FAILEDTOBUILD.txt.gz ?
[10:09] <doko> pitti: not yet, have to look
[10:35]  * hunger wonders whether he should switch to intrepid...
[10:37] <cjwatson> if you enjoy having a broken system
[10:37] <cjwatson> (it won't be broken right now, but will be as soon as we start the autosync)
[10:38] <Hobbsee> i can't believe i'm actually pondering not running intrepid until alpha 5 or so.
[10:38]  * ogra usually upgrades a week after uds
[10:38] <RAOF> Hobbsee: WHy is that?
[10:38] <hunger> will hardy is sooo boring. No new packages, nothing:-(
[10:39] <ogra> hunger, you could fix bugs for 8.04.1 ;)
[10:39] <RAOF> Do you have something _useful_ to do with your computer, or other crazy reason?
[10:39] <Hobbsee> RAOF: undecided how much dev work i'll end up doing.  although presumably i'll start with gnome stuff.
[10:39] <ogra> that will give you even new packages in -updates
[10:39] <Hobbsee> RAOF: not overly.  the stupid uni ports stuff means there's no point taking it there
[10:39] <hunger> ogra: Nah, that is your job. I do report bugs, but I don't have the time to fix them for you.
[10:40] <ogra> hunger, but yu have time to be bored by missing updates ? come on :)
[10:40] <hunger> Well, actually I only report bugs when they are more pain to me than using LP:-)
[10:41] <RAOF> Hobbsee: Hm.
[10:41] <Hobbsee> hunger: there is a mail interface.
[10:41] <RAOF> It won't be the same without you ;)
[10:41] <Hobbsee> no, no Long Pointy Stick of DOOM!!!!!!!!!!!!!!!™
[10:42] <Hobbsee> ogra: that was my old habit.
[10:43] <hunger> ogra: Well, usually you do not break stuff that really manages to stop me from using my toy-box.
[10:44] <hunger> ogra: And managing my own packages (like git-core which is too old in hardy) takes time as well.
[10:45] <cjwatson> at the risk of stating the obvious, anyone in #ubuntu-devel is at least implicitly accepting responsibility for doing some development. :-)
[10:46]  * ogra adds a if [ "$(id -un)" = "hunger" ]; do break_in_mysterous_ways .... to all his packages 
[10:46] <ogra> err
[10:46] <ogra> s/do/then
[10:46] <hunger> ogra: Feel free to do that.
[10:46] <ogra> heh
[10:46] <hunger> ogra: IRC_nick != login_name ;-)
[10:49] <slytherin> Can anyone please comment if bug 199116 will qualify for SRU?
[11:04] <Riddell> what's the ubuntu xrandr tool called?
[11:04] <Mithrandir> the xrandr tool is called xrandr.
[11:04] <cjwatson> hmm, I've had two reports of bug 221635 now
[11:04] <mvo> Riddell: you probably want gnome-display-properties
[11:05] <ogra> lp seems down
[11:05] <cjwatson> just slow
[11:05] <ogra> or extra slow at least
[11:05] <pitti> just worked up to 30 seconds ago, now it's stuck here, too
[11:06] <cjwatson> I'd love to know why a signal handler installed with sigaction (SIGCHLD, ...) is getting signum != SIGCHLD
[11:06] <cjwatson> anyone have any idea what might be going on there?
[11:06] <ogra> slytherin, where is the fix in that bug ?
[11:07] <slytherin> ogra: I am preparing it with the patch in upstream bugzilla.
[11:07] <ogra> without patch nobody can tell you if it justifies for an SRU :)
[11:08] <slytherin> ogra: Ok. I will create the fix, test the package myself and then upload the debdiff.
[11:08] <seb128_> slytherin: GNOME 2.22.2 will be uploaded to hardy so you might want to not bother and wait a few weeks
[11:09] <ogra> oh, there is an upstream task on it, right
[11:10] <slytherin> seb128_: I will have to ask vinagre maintainer if he is going to do a release before that. Meanwhile I will provide a fix in my ppa anyway for those in need. Is that fine?
[11:10] <ogra> slytherin, the upstream bug already has a commit entry
[11:10] <seb128_> right
[11:10] <seb128_> it'll not be fixed in stable though
[11:10] <ogra> no need to put work into it
[11:10] <ogra> seb128_, no SRU ?
[11:10] <seb128_> since it breaks the string freeze apparently
[11:11] <seb128_> ogra: not sure if we have a sru policy about not breaking translations in stable
[11:11] <ogra> +  { "MachineSendCtrlAltDel", GTK_STOCK_ABOUT, N_("Send CtrlAltDel"), NULL,
[11:11] <seb128_> ogra: +    N_("Send Ctrl+Alt+Del to active connection"), G_CALLBACK (vinagre_cmd_machine_send_ctrlaltdel) },
[11:11] <ogra> if you drop "Send" its working for most langs
[11:12] <seb128_> ogra: not really, and it would be ugly
[11:12] <ogra> indeed
[11:12] <seb128_> my point is upstream will not fix it in GNOME 2.22.2 so yes we need to sru it if we want the change
[11:12] <seb128_> though I'm not decided if breaking translations in stable is a good idea
[11:13] <seb128_> or adding new strings in this case
[11:13] <seb128_> I guess we will have regular language pack updates so if it's done early before 8.04.1 why not
[11:38] <slytherin> seb128_: ogra: As of now, the fix is made in 2.23.x branch in upstream svn. I will write mail to upstream author asking if it will be available in the version in GNOME 2.22.2.
[11:38] <ogra> slytherin, seb explained before that its not suitable for an SRU
[11:38] <ogra> it breaks translations
[11:39] <slytherin> ogra: But if it is part of GNOME 2.22.2, then it is fine right?
[11:39] <ogra> it isnt
[11:39] <ogra> read what seb128_ wrote above
[11:40] <slytherin> ogra: I read it, but I misunderstood
[11:41] <ogra> slytherin, the transaltion thing has to be clearified before a decision can be made at al
[11:42] <slytherin> ogra: Ok. For now I will just upload the package to  my ppa.
[12:22] <xivulon> Hi all, is there any russian here? I would like a second opinion on bug #220112
[12:29] <seb128_> slytherin: do you think it's an import feature or something many users requested?
[12:30] <cjwatson> xivulon: surely that's a console-setup bug rather than anything wubi-specific
[12:30] <ogra> seb128_, you need ctrl-alt-del for windows logins
[12:31] <slytherin> seb128_: in my opinion it is regression because xvncviewer has option to send ctrl + alt + del. And from my experience using vnc for last 4 years, you need it 70% of the time. :-)
[12:31] <cjwatson> xivulon: ah, never mind, already filed on console-setup
[12:31] <seb128_> slytherin: doing a sru seems to be a good idea then
[12:32] <seb128_> slytherin: we want those sort of fixes for 8.04.1 and it can be translated in rosetta if uploaded soon enough
[12:33] <slytherin> seb128_: Ok. I have created a package and testing it. How do I update .pot file?
[12:34] <seb128_> the package uses cdbs and the gnome rule so it's done automatically
[12:35] <seb128_> you just need to make sure the source is listed in the POTFILES.in
[12:36] <slytherin> seb128_: I will be back in 20 minutes. Will get back if I have probolem.
[12:36] <seb128_> ok
[12:36] <seb128_> you can just look to the pot after the build and make sure it contains the new strings
[12:39] <xivulon> 00000419
[12:39] <xivulon> cjwatson, is it ok if I set it invalid for wubi then? When I see windows keyboard #00000419 I will only preseed local=ru without variant
[12:40] <xivulon> locale=ru
[12:40] <xivulon> which is what I am doing already
[12:40] <cjwatson> xivulon: if Windows has two different keyboard layouts, one of which corresponds to ru and one of which corresponds to ru:winkeys, then there would be a reason to do something in wubi
[12:40] <cjwatson> xivulon: otherwise, just console-setup/layoutcode=ru is appropriate, and wubi shouldn't override it
[12:41] <xivulon> yeah that is what I asked, but had no reply. From what I can gather 00000419=ru+winkeys, not sure if there is another code for ru-winkeys. But I would doubt it
[12:42] <xivulon> since normally windows code for explicit variants are a bit different
[13:12] <slytherin> seb128: 2 questions. 1. I have built the package with 'dpkg-buildpackage -b' and tested it. The .pot change doesn't reflect in debdiff, is that fine? 2. Should I have release as 'hardy-proposed' in changelog?
[13:17] <tkamppeter> pitti, hi
[13:32] <tkamppeter> Riddell, hi
[13:32] <dneary> Hi Till
[13:33] <tkamppeter> dneary, hi
[13:33] <dneary> Should acpid and apmd both be activated in 8.04 for power management to work?
[13:34] <dneary> I'm struggling with https://bugs.launchpad.net/ubuntu/+source/linux/+bug/223848
[13:34] <ogra> apmd is a noop usually
[13:34] <dneary> I'm also struggling to figure out why a printer shared by a Mac, which worked fine until I upgraded, is no longer getting automatically picked up
[13:34] <ogra> only there for systems that actually do apm, else it wont even start up
[13:35] <dneary> I'm working around the suspend issue, my remaining issues are the printer and sound (which isn't working at all)
[13:35] <dneary> ogra: So there's no harm in leaving it checked, then?
[13:38] <Riddell> hi tkamppeter
[13:38] <ScottK> Riddell: Who is your relief today for the archive?  I missed one package yesterday (I'll upload it in a moment).
[13:38] <Riddell> ScottK: seb128, but I can do it
[13:38] <ogra> dneary, checked ? where ?
[13:38] <ScottK> Riddell: Thanks.  I'll ping you after it's uploaded then.
[13:39] <dneary> ogra: In System->Administration->Services
[13:39] <ogra> ah, yes
[13:39] <ogra> its automated, you can leave it on
[13:40] <dneary> OK, printer OK
[13:40] <dneary> Next: sound
[13:42] <shashi>  I am using Ubuntu 8.04 64-bit version, if i install any 32-bit applications like browsers, datbase clients ...etc. The 32-bit based applications not able to reach /etc/resolv.conf file to communicate to the network. Any one tell me how to resolve this issue ?
[13:42] <Hobbsee> shashi: this is not a support channel, and you have no need to put it into multiple channels at once
[13:42] <Hobbsee> shashi: please read the /topic.
[13:48] <ScottK> Riddell: claws-mail uploaded to gutsy-backports, please accept it.  I'd also appreciate it if you would backport devscripts from Intrepid to Hardy (I've tested it - can file a bug if you want it).
[13:50] <dneary> hmmm
[13:50] <dneary> alsamixer looks fine
[13:50] <dneary> But even catting a wav to /dev/audio doesn't give me any noise
[13:51] <Riddell> ScottK: done
[13:51] <ScottK> Riddell: Thanks.
[14:00] <fta> doko, it seems gcc 4.2 is not upgradable in intrepid without trashing a part of gnome. is it just me or is it a known issue ?
[14:00] <cjwatson> it seems a little early to be worried about upgradeability in intrepid
[14:00] <cjwatson> it's going to be much more spectacularly broken soon enough, for a while
[14:00] <Hobbsee> cjwatson: except for any of those who are insane enough to run it.
[14:01] <wgrant> cjwatson: Hardy was never particularly broken.
[14:02]  * \sh loves xorg breakage ... regarding one of our last cycles while our Xorg maintainers were transforming xorg into something better ,-)
[14:04] <dneary> Did anything replace logrotate in 7.10?
[14:05] <dneary> I have some huge log files (messages, syslog, debug.log, daemon.log) that should have been rotated out by now.
[14:05] <doko> I assume that's the wrong external dependency of python-gobject on libffi4
[14:07] <fta> doko, looks like it
[14:09] <fta> doko, or libffi4 itself
[14:11] <tkamppeter> Riddell, are you already in contact with Alex and Lars?
[14:11] <Tonio_> hi
[14:11] <Riddell> tkamppeter: lars yes, we had a good chat yesterday evening, he seems to know what to do
[14:11] <Riddell> tkamppeter: alex I'm yet to hear from
[14:12] <Tonio_> is there a way to invoke language-selector the command line way only (without a X session) ?
[14:12] <Riddell> tkamppeter: lars is on #openusability
[14:13] <tkamppeter> Riddell, Lars has already done good work for the new foomatic-rip, which is the most important point to let Intrepid use PDF as standard print job data format. I was VERY content with him.
[14:13] <dneary> OK - so it's still logrotate, it's run through cron.daily, which would run if my computer wasn't suspended overnight, and there doesn't seem to be any recipe to rotate syslog, daemon.log, kern.log, debug, auth.log, or messages
[14:13] <tkamppeter> Riddell, you are chatting there currently with him? Will enter the channnel.
[14:14] <tkamppeter> Riddell, what about Alex, did you contact him and he did not answer or what is the problem?
[14:21] <fta> doko, 4.2.3-4ubuntu1 doesn't seem to be complete.  http://paste.ubuntu.com/8924/  are those libs no longer built ?
[14:29] <Amaranth> pedro_: can you renew my membership to ubuntu-bugcontrol?
[14:29] <pedro_> Amaranth: no way!
[14:29] <Amaranth> haha
[14:29] <doko> fta: no
[14:29] <pedro_> Amaranth: give me a min ;-)
[14:29] <Amaranth> thanks :)
[14:32] <dneary> Anyone mind sending me their default /etc/logrotate.conf, please?
[14:34] <pedro_> Amaranth: you're welcome, it's done now :-)
[14:34] <Amaranth> pedro_: woohoo, now i can keep marking serious usability bugs as wishlist
[14:35] <pedro_> hahaha
[14:47] <dneary> OK - I think the sound problem is fixed
[14:49] <tkamppeter> pitti, hi
[14:54] <pitti> hey tkamppeter, how are you?
[14:56] <tkamppeter> fine.
[14:56] <tkamppeter> I want to know about bug 217787.
[15:03] <tkamppeter> pitti, If the user has not pam_smbpass installed, CUPS does not work correctly for him, if he has this PAM module installed, CUPS crashes in the module.
[15:04] <fta> doko, no as in "it's a wanted change" or no as in "it's a bug". It seems to me that your "regenerate control file" was a bit too aggressive. all the libs are correctly built but they no longer have a deb associated.
[15:15] <doko> fta: no, it's not a bug, the soname did change
[15:18] <fta> doko, eh?? i'm talking about this: http://paste.ubuntu.com/8944/  seems to be much more than a soname change
[15:18] <fta> or did i miss something?
[15:19] <lool> E: Sub-process /usr/bin/dpkg received a segmentation fault.
[15:19] <lool> Just after unpacking sudo
[15:19] <lool> pitti: ^
[15:19] <lool> And no crash report, damn
[15:21] <pitti> lool: ugh, dpkg crash?
[15:21] <lool> Yeah
[15:21] <pitti> lool: do you have apport enabled?
[15:21] <lool> yes
[15:21] <lool> I'm trying it again
[15:21] <lool> This time, it's stuck
[15:21] <lool> Or busy for a long while
[15:22] <lool> Hmm dmesg wont run, but I can run top; weird
[15:23] <lool> The system is foobar
[15:23] <pitti> that seems seriously screwed
[15:23] <pitti> lool: (FUBAR, you mean? :-) )
[15:24]  * lool reboots
[15:26] <lool> pitti: Dépaquetage de la mise à jour de sudo ...
[15:26] <lool> Erreur de segmentation
[15:27] <lool> (segfault)
[15:27] <lool> reproducable
[15:27] <lool> And still no crash
[15:27]  * pitti removes ~/ubuntu/sudo/debian/patches/dpkg_overflow_exploit.patch
[15:28] <lool> pitti: but kill -segving a bash does produce a crash file
[15:28] <pitti> lool: anything in /var/log/apport.log? does the dpkg crash at leats appear there?
[15:28] <lool> No
[15:29] <lool> pitti: Could it be that dpkg registers a segv handler?
[15:29] <lool> I don't know how the kernel machinery works
[15:29] <pitti> lool: theoretically yes, but the stderr output looks fairly standard
[15:29] <pitti> lool: do you get a segfault in dmesg?
[15:29]  * pitti dist-upgrades to sudo hardy-proposed
[15:29] <lool> I get an OOPS in dmesg indeed
[15:29] <lool> So it's kernel stuff
[15:29] <lool> [  125.486444] invalid opcode: 0000 [#1] SMP
[15:30] <lool> http://paste.ubuntu.com/8947/
[15:31] <ogra> hmm, i didnt get sudo offered even with the latest update
[15:31] <cjwatson> dpkg doesn't handle SEGV
[15:31] <mjg59> lool: I blame unionfs
[15:31] <lool> mjg59: Very likely
[15:32] <lool> invalid opcode is weird; is this a kind of code overwrite consequence?
[15:33] <pitti> dist-upgrade worked fine here (hardy + -proposed, amd64)
[15:33] <pitti> ah, unionfs
[15:33] <lool> Someone running the live CD could try pulling sudo from hardy-proposed
[15:34] <mjg59> lool: That's it hitting a BUG_ON statement
[15:34] <lool> For fun and profit of course
[15:34] <lool> mjg59: Ok
[15:34] <mjg59> So it's because the kernel's got into a state which ought to be impossible
[15:34] <lool> So I dump that in a bug report or is someone here interested in my providing more info?
[15:35] <mjg59> That's probably the best you can do right now
[15:37] <lool> unionfs -> linux-ubuntu-modules-2.6.24 I guess
[15:38] <ogra> lool, checking on a classmate, gimme some mins ...
[15:42] <ogra> lool, same
[15:42] <lool> https://bugs.launchpad.net/ubuntu/+source/linux-ubuntu-modules-2.6.24/+bug/224754
[15:42] <lool> ogra: Thanks for reproducing
[15:42]  * ogra adds a me too to the bug
[15:43] <lool> hrwsr-xr-x root/root         0 2008-04-30 11:39 ./usr/bin/sudo lien vers ./usr/bin/sudoedit
[15:43] <lool> Could it be the weird permissions?
[15:43] <lool> Other links have:
[15:43] <lool> hrw-r--r-- root/root         0 2008-04-30 11:39 ./usr/share/man/man8/sudoedit.8.gz lien vers ./usr/share/man/man8/sudo.8.gz
[15:44] <ogra> urgh
[15:44] <ogra> now my termonal doesnt take input
[15:45] <lool> ogra: Oh it will hose your host, sorry
[15:45] <ogra> lool, doesnt matter, i'm surrounded by classmates, i can just take another one ;)
[15:45] <ogra> where do you see these permissions ?
[15:46] <ogra> mine seem ok
[15:46] <ogra> -rwsr-xr-x root root .....
[15:47] <lool> ogra: dpkg-deb -c on sudo.deb
[15:47] <ogra> ah
[15:48] <ogra> i dont have a h there
[15:48] <ogra> oh, i do
[15:48] <ogra> looked at the wrong file
[15:48] <ogra> does unionfs not like hardlinks ?
[15:49] <lool> we shouldn't be shipping hard links in packages
[15:50] <lool> Yeah, it's a hard link on my desktop
[15:50] <mjg59> lool: Why not?
[15:50] <lool> mjg59: I think it's against policy
[15:50] <lool> Just like device files
[15:50] <lool> But this is from memory, ICBW
[15:50] <mjg59> lool: Seems to be a should rather than an absolute, but I'd be interested to know the reasoning
[15:51] <ogra> hm, hard links dont work across different filesystems, right ?
[15:51] <ogra> which is exactly what unionfs tries here i bet
[15:51] <mjg59> ogra: No, but unionfs should still cope
[15:52] <lool> I see nothing current in policy
[15:52] <mjg59> lool: 12.1
[15:52] <lool> Only prevents you to use hardlinks with conffiles
[15:52] <lool> And source packages
[15:52] <mjg59> lool: Seems to be specific to manpages
[15:52] <lool> Oh and man pages
[15:52] <lool> mjg59: man pages and other stuff
[15:53] <lool> mjg59: I do recall some package (lirc I think) has a RC for shipping device files, but that's another story
[15:53] <mjg59> If it's in the same directory, it shouldn't be a problem
[15:53] <lool> Agreed
[15:53] <ogra> shipping device files ?
[15:53] <ogra> thats just crazy
[15:55] <lool> So hmm do we have unionfs hackers around
[15:55] <lool> Cause it's the second high impact bug with unionfs and we currently plan to ship with unionfs on the devices...
[15:55] <ogra> aufs ftw in intrepid :)
[15:55] <ogra> even though we dont have hot aufs hackers either
[15:56] <ogra> lets see what #ubuntu-kernel has to say :)
[15:57] <\sh> ogra, shipping device files, or actually creating device files during make run was in former times the standard behaviour ,-)
[15:57] <ogra> \sh, udev is used since when ?
[15:57] <ogra> :)
[15:58] <\sh> ogra, well, it wasn't there in SuSE Slackware Nov./94 :)
[15:58] <ogra> lool, do you actually do updates on the MID by default ? (on top of the unionfs)
[15:59] <ogra> i dont think it will break if the hardlink is inside the rofs only
[15:59] <lool> ogra: Yes we do
[15:59] <lool> Or rather we will
[15:59] <pochu> do we have overrides for sections as Debian have?
[15:59] <pochu> or do we trust debian/control?
[16:00] <ogra> you must have plenty of space
[16:00]  * ogra is envious
[16:00] <lool> ogra: Good point
[16:00]  * Daviey had some major problems with unionfs that didn't show in aufs..
[16:00] <lool> ogra: We actually do that to save space
[16:00] <ogra> updates ?
[16:00] <lool> ogra: But I realize how silly it is
[16:00] <ogra> you have the same apps twice then
[16:00] <lool> ogra: We want the standard image to fit on small disks
[16:00] <ogra> one in rofs and once in cow
[16:01] <lool> But if you take into account the place taken by the updates, you lose a lot of space
[16:01] <ogra> yeah, i understand i have similar probs on the cmpc
[16:01] <lool> Depends on the amount of updates to the base system and for how long
[16:01] <lool> But it's going to be interesting after some kernel updates
[16:02] <lool> Hmm that's though to reconcile
[16:02] <lool> I can only propose to drop unionfs, but then we wont ever make the disk space commitments in time
[16:02] <ogra> well, i have /var and /boot on real partitions on the cmpc
[16:03] <ogra> (/var on ext3 speeds up the package tools immensely)
[16:03] <ogra> they are nearly unusable on unionfs
[16:03] <lool> They work decently here
[16:03] <ogra> lucky you
[16:03] <lool> Well it's slow to read the list of installed files/packages
[16:04] <ogra> g-a-i never returns from reading its package lists if i run it on unionfs and have universe enabled ...
[16:04] <Hobbsee> holy cow.
[16:04] <ogra> at some point the device just dies on IO
[16:04]  * Hobbsee starts going thru the moderation lists
[16:05] <ogra> luckily putting var away saves my butt here :)
[16:05] <Hobbsee> looks like more people are forging the address to spam
[16:11] <Hobbsee> right.  moderation queue done.
[16:12] <Hobbsee> cjwatson: you may want to put in spam rules for 'Delivery Status Notification (Failure)' and 'failure notice' and 'Undelivered Mail Returned to Sender' etc, on that list
[16:12] <Hobbsee> (u-d@)
[16:12]  * Hobbsee can now go to bed
[16:13] <cjwatson> Hobbsee: I actually filed an RT request the other day asking for postmaster@ and mailer-daemon@ to be discarded across the board
[16:13] <cjwatson> which should cover mostly the same thing
[16:13] <Hobbsee> cjwatson: ah yes, that would help, too
[16:27] <emgent> heya people
[16:44] <crop[linuxforum]> Hello all
[16:45] <crop[linuxforum]> Are there anything known about 8.10: when it will be, main goals e.t.c?
[16:46] <ogra> obviously in october as the version tells you
[16:47] <ogra> goals will be defined at the developer summit
[16:52] <crop[linuxforum]> ogra, thank you:-) I will watch for development of 8.10
[17:40] <mathiaz> slangasek: I've going through the samba bugs - it seems there is an issue with ucf
[17:41] <mathiaz> slangasek: there are a couple of post-install failed - they seems all related to the ucf prompt
[17:43] <slangasek> mathiaz: bug #s?
[17:46] <CarlFK> if ntp.ubuntu.com seems to off by about 30 seconds (as is another stratum2), should I bother anyone here about it?
[17:51] <cjwatson> CarlFK: #canonical-sysadmin
[17:53] <CarlFK> thanks
[17:53] <ion_> Canonical might as well be running a stratum 1 server. :-)
[17:53] <nxvl> did someone knows who is in charge of shop.canonical.com?
[17:55] <mathiaz> slangasek: bug 221427, bug 221526, bug 221963, bug 223971, bug 224061, bug 224567
[17:56] <slangasek> mathiaz: 221526 says 'swat', which is unrelated to ucf
[17:59] <slangasek> mathiaz: the term.log on #221963 is very interesting :)
[18:00] <Caesar> slangasek: #216990 is turning into a real bobby dazzler, eh?
[18:01] <mathiaz> slangasek: yeah - seems like an issue between ucf and the gnome-frontend
[18:02] <slangasek> mathiaz: really, it looks to me like he had a non-pristine smb.conf, chose 'start a new shell to examine the situation', and then manually killed the frontend for whatever reason
[18:02] <slangasek> Caesar: I'm not sure what a bobby dazzler is, so I'm just going to nod fervently in agreement
[18:03] <Caesar> slangasek: rip-snorter?
[18:03] <slangasek> there we go :)
[18:03] <mathiaz> slangasek: well - bug 224061 has the same issue
[18:09] <slangasek> mathiaz: that one doesn't look like the same bug; actually 224061 is one I've seen before, ucf seems to do something totally nonintuitive and broken when the only differences are whitespace :/
[18:09] <slangasek> mathiaz: so we need to fix that in ucf
[18:10] <mathiaz> slangasek: right - some of these bugs are related to ucf rather than samba
[18:11] <mathiaz> slangasek: samba doesn't use ucf is some wired ways - so I'll reassign the bugs to ucf
[18:11] <slangasek> ok
[18:12] <slangasek> I'm not sure 221963 is any bug that we can fix; I can't see what we should do differently if the user kills the frontend in X
[18:16]  * cjwatson wonders if http://people.ubuntu.com/~cjwatson/tmp/lp-bug-notify.png is something anyone else wants to have
[18:16] <cjwatson> the notification, that is
[18:17] <mathiaz> slangasek: it seems there are are two issues - the one were the user had a root prompt and did an exit
[18:17] <soren> Possibly. procmail+notify-send?
[18:17] <cjwatson> procmail plus bizarre lash-up plus notify-send, but right
[18:17] <mathiaz> slangasek: and then further down, the one where ucf prompted for the potential action
[18:18] <cjwatson> it actually goes through the irssi notification system I set up earlier today ;-)
[18:18] <mathiaz> slangasek: I'll ask the user what he did in the second situation
[18:18] <soren> cjwatson: I'd like to see it, actually.
[18:19] <slangasek> mathiaz: I think the 'exit' itself is probably not the cause of the failure; the user is clearly using the gnome frontend in the first run, and I think he first chose the 'examine with a shell' option, and then subsequently killed the window
[18:19] <soren> I'm running out for bit, though.
[18:19] <cjwatson> soren: http://people.ubuntu.com/~cjwatson/notifications/ has most of the bits
[18:20] <cjwatson> the special key in fnotify is a bit gross, and there's no locking on that file
[18:20] <cjwatson> and the irssi bit is dependent on http://www.leemhuis.info/files/fnotify/fnotify
[18:21]  * cjwatson updates irssi-notify there for the LPBUG stuff
[18:46] <astronut> i can't find a straight answer in any of the published documentation, in the installer's partitioner, does manual -> edit partion -> newsize do a destructive or nondestructive resize, esp ntfs?
[18:47] <astronut> or is it based on the d-i code?
[18:48] <sladen> astronaut: what do you mean destructive/nondestructive?
[18:49] <astronut> sladen: edit partition table and ignore filesystem or actually adjust filesystem
[18:49] <astronut> iui, will changing the filesize of an ntfs partition call ntfsresize to shrink it
[18:49] <sladen> astronut: are the files on the NTFS partition deleted (no).  Are the files moved around (yes, they are moved towards the "front").  Will they still be there when you reboot (yes)
[18:50] <astronut> sladen: nondestructive then
[18:50] <sladen> astronut: Yes, ntfsresize has been used in the Ubuntu installer for ~3 years automatically
[18:50] <astronut> sladen: another possible design: editing the partition size adjusts the partition table but doesn't touch the filesystem. on reboot, the filesystem is corrupt and ya
[18:50] <astronut> sladen: glad to hear
[18:51] <astronut> i wasn't sure if it was in the manual or just the guided
[18:51] <astronut> docs weren't clear and the prompts weren't very reassuring
[18:51] <astronut> got no guided prompt because there weser several partitions on that drive (dell laptop)
[18:51] <hwilde> on that note, it's nto really clear when resizing the partition, it says "New partition size:"  then it has a slider bar.  is that the new size of the existing partition, or the size of the new partition to be created?
[18:52] <sladen> astronut: agreed.  I get scared myself.  File a bug that it should be more informative (especially if you can work out how to phrase it more concisely/clearly)
[18:52] <astronut> hwilde: that's not as obscure
[18:52] <astronut> since it's edit not create
[18:52] <astronut> also it starts at the current value
[18:52] <astronut> sladen: i'm doing it for a friend, i don't use ubuntu normally
[18:52] <hwilde> astronut, i'm sure it's easy for you to understand but not for everyone.
[18:53] <astronut> *nod*
[18:53] <astronut> hwilde: i didn't get a slider though, got a num box
[18:53] <hwilde> most people think they are making a new partition to install ubuntu, and then it says "New partition size"
[18:53] <astronut> what's the name of that control called? spinner?
[18:53] <astronut> this was from the live cd installer, not the bootable one
[18:53] <hwilde> so they think, oh I only need 8G for my new partition, but in reality they are resizing their existing partition
[18:53] <laga> hi!
[18:53] <sladen> astronut: if you've had to come here to check, for every one of you, there will be a thousand who haven't asked.  And it would be good to fix it for them aswell.
[18:54] <astronut> sladen: sure... i just hate LP
[18:54] <astronut> though i might have an account, not sure
[18:54] <laga> is there any way to make my see this bug: https://bugs.launchpad.net/ubuntu/+source/mythtv/+bug/221176 i'm one of the mythtv maintainers in ubuntu and i'd like to triage it.
[18:54] <hwilde> but it wouldn't be that difficult to label both sides of the slider bar:  Re-sizing of the existing partition,    New partition size
[18:54] <astronut> i didn't have a slider bar
[18:54] <astronut> i had a number box, insert size in m
[18:54] <astronut> b
[18:54] <hwilde> desktop livecd installer?
[18:55] <james_w> laga: ask in #ubuntu-bugs for a member of bugcontrol to process it
[18:55] <laga> james_w: thanks.
[18:55] <astronut> hwilde: ya
[18:55] <astronut> 8.04
[18:55] <astronut> i'm using sysresccd to do the resize
[18:55] <astronut> now
[18:56] <astronut> since i couldn't get an answer
[18:56] <sladen> astronut: well, I'm not the words greatest fan of Launchpad either.  However, it's what we've got (though it has improved over time).  If you have have suggestions about how to help the LP team make Launchpad suck Less, file at https://bugs.launchpad.net/malone/+filebug
[18:56] <hwilde> it's definitely a slider bar in the livecd installer with gparted, and it definitely says "New partition size", and it's definitely confusing enough that people have asked me about a dozen times what that means
[18:56] <astronut> sladen: heh...
[18:56] <astronut> sladen: normally ijust use debian and ignore ubuntu except when i get MoM mails
[18:58] <sladen> astronut: fixing d-i will fix the resize translation issue for Debian *and* Ubuntu.  So it still be appreciated if you could file it.
[18:58] <astronut> sure, later
[18:58] <astronut> i'll have to double check the message is th esame
[18:59] <sladen> astronut: very likely.  It's a sed script that replaces every instance of 'Debian' with 'Ubuntu'
[19:00] <astronut> now this livecd isn't booting
[19:00] <astronut> *annoyed*
[19:00] <astronut> ok, i'll just do the manual edit
[19:01] <sladen> astronut: unless you're setting up RAID/LVM you really probably don't need to use the advanced made
[19:02] <astronut> sladen: i said manual, not advanced
[19:02] <sladen> astronut: grab the DesktopCD and double-click the icon on the desktop and you'll get a shiny GUI partition editor (which should automatically guess "what you want to do"(tm) anyway)
[19:02] <astronut> sladen: this thing has a coupel partitions, i get either guided (whole disk) or manual
[19:02] <astronut> couple*
[19:02] <astronut> sladen: been there, done that, didn't happen
[19:03] <astronut> i may not have done many ubuntu installs, but i've done tons of debian installs...
[19:04] <ogra> we must be nearing an UDS ... sladen is back :)
[19:04] <sladen> astronut: so you're using a DesktopCD with "manual partitioning".  Or are you using the "Alternate CD"?
[19:04] <astronut> the former
[19:04] <astronut> it's booting right now
[19:04] <astronut> into the installer
[19:04] <astronut> instead of live environment
[19:05] <astronut> though apparently they're not that different
[19:05] <astronut> i dont' get a slider in manual mode, sadly
[19:05] <astronut> you guys should work on that
[19:05] <astronut> what is it, just a gtk port of whiptail?
[19:08] <astronut> lets hope i don't kill my friend's data
[19:09] <hwilde> !enter | astronut
[19:09] <hwilde> I love ubottu  :)
[19:09] <astronut> sorry hwilde
[19:09] <astronut> sladen: did i meet you at DC8?
[19:11] <astronut> uh...hello... it just returned from resize and they don't look any different
[19:12] <astronut> wait, never mind, the dialog just got hidden... it's still at 0% after like 10 minutes
[19:16] <sladen> astronut: d-i is the alternate installer; Ubiquity (https://wiki.ubuntu.com/Ubiquity) is the Live installer (which hooks back to d-i), for which there are GTK and Qt frontends.
[19:16] <sladen> astronut: maybe.  I was certainly at Debconf
[19:16] <astronut> how lon gshould this take? 40->30 gb resize
[19:16] <sladen> astronut: if the dialogue isn't modal on-top, that's a bug
[19:16] <astronut> i think i clicked on the install dialoge to bring it forward
[19:16] <astronut> by mistake
[19:17] <sladen> astronut: probably quite awhile... depends how full it is;  there's probably 5-15GB to be moved around
[19:18] <astronut> ok, fair enough
[19:18] <astronut> it'd be ncie if the progress bar worked
[19:18] <sladen> again.  If it doesn't.  *PLEASE* file a bug
[19:18] <sladen> you've found 3 issues so far
[19:19] <astronut> there we go
[19:19] <sladen> (1) the wording on the resizer can be improved.  (2) that the progress bar should be modal  (3) progress bar taking ages to go from 0% to 1%
[19:20] <sladen> astronut: what's your LP id.  I'll even go to the trouble of opening the bugs
[19:20] <astronut> ...why does it say unusable space?
[19:20] <astronut> astronut i guess
[19:20] <astronut> if i have one
[19:20] <astronut> it was when they imported debian people's
[19:20] <astronut> is 11589 mb not sufficient to install?
[19:21] <astronut> that's about 10 gigs
[19:21] <sladen> more than enough.  But how much free-space is there in the NTFS?
[19:21] <astronut> i'm saying it ays there's unusable space on disk
[19:21] <astronut> oh, shit i'm guessing there are 4 primary partitions
[19:22] <sladen> (4) issues.
[19:23] <astronut> ok, now that thare's free space but partions, will the guided to a primary -> logical conversion?
[19:24] <sladen> astronut: I have a feeling not.  Find cjwatson
[19:24] <astronut> him i remember meeting at dc
[19:25] <astronut> that'll be the same as di, right?
[19:25] <astronut> i'm waiting on windows chkdsk to make sure nothing broke
[19:27] <astronut> why can't we get rid of the whole only four partion things?
[19:27] <astronut> or is that a matter of the size of the partition table
[19:30] <sladen> astronut: http://en.wikipedia.org/wiki/GUID_Partition_Table
[19:33] <astronut> there's not a way to do primary->logical in DI
[19:33] <astronut> bah
[19:35] <astronut> err, ubiquity... does DI do it?
[19:37] <sladen> oh, it'll get poked to the right place eventually
[19:40] <highvoltage> hey sladen!
[19:41] <sladen> highvoltage: what happened to that fridge list stuff?  I got a message with a launchpad.internal URL in it and ignored it more
[19:42] <highvoltage> sladen: fridge list stuff? I don't think I got anything about it.
[19:42] <sladen> astronut: right.  I've failed all of those issues except the "launchpad suckz" one.  Are there any others?
[19:42] <sladen> highvoltage: oh well.  Never mind.  There are more important things in life
[19:43] <astronut> i saw
[19:43] <highvoltage> sladen: I'll PM you :)
[19:43] <astronut> sladen: ubuiquity should have an option to turn primary partitions into logical ones
[19:44] <mathiaz> james_w: re bug 221963 - why did you set the state back to New ?
[19:44] <james_w> mathiaz: because the files that were asked for were provided, is there more information needed?
[19:46] <james_w> mathiaz: sorry for changing it while you were working on it
[19:46] <mathiaz> james_w: no problem - I tend to not set it to New once I've had a look at it
[19:46] <mathiaz> james_w: For me, NEW means that nobody looked at it
[19:47] <mathiaz> james_w: then I move it to Incomplete until the bug has been Confirmed, Invalid, Won't Fix or Triagged
[19:47] <james_w> mathiaz: ah, ok.
[19:47] <mathiaz> james_w: There isn't a standard workflow
[19:51] <lesshaste> Hi, Can I ask about how to assign a bug to kernel? r toolchain | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application dev has a note saying that but I don't know how to actually do it
[19:51] <lesshaste> grr
[19:51] <lesshaste> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/224561 has a note saying that but I don't know how to actually do it
[19:52]  * lesshaste is moving to #ubuntu-bugs
[19:59] <astronut> sladen: apparently gparted doesn't do taht
[19:59] <astronut> it's funny, you'd think it'd be relatively trivial, it's all done in the MBR, no? or would you have to shift everything down a few k
[20:00] <sladen> astronut: that's what I can't think.  Can't remember if it needs an extra 4kB ading
[20:01] <sladen> or just twiddling MBR bits (it's going to break disk labelling in Windows)
[20:05] <astronut> sounds like it requires a small shift...bah
[20:05] <astronut> still, it doesn't sound overly complicated...should be added to gnu parted
[20:06] <astronut> compultationally complex, but not procedureally
[20:06] <astronut> although i guess if there's no free space, you can't do it
[20:06] <astronut> in my case it should be easy since there's free space ahead of it but no tool to do it
[20:20] <cjwatson> sladen: there are bug reports about everything mentioned above already, and also a few specifications
[20:20] <cjwatson> astronut: it feels rather like saying "yes, this partitioner will not destroy your data and instead will behave sensibly", to be perfectly honest; i.e. stating what should be the obvious state of affairs
[20:20] <cjwatson> hwilde: that slider is fixed in 8.04
[20:21] <mysterycool> Hello.
[20:21] <cjwatson> astronut: you can't do a primary->logical conversion; you'd break whatever operating system lives there
[20:21] <cjwatson> astronut: very bad idea to present that
[20:21] <cjwatson> astronut: and no, ubiquity is not a gtk port of whiptail :-P
[20:21] <cjwatson> astronut: but it is based on the d-i code
[20:22] <mysterycool> I have been checking around ubuntu's web now and i found somewhere that it says "if u want to learn u can find a mentor in the development team". where is the development team and how can I get in touch with a mentor?
[20:22] <hwilde> cjwatson, cool i've only done the server install yet
[20:23] <cjwatson> hwilde: the live CD installer does not use gparted (since 7.04); please don't refer to gparted unless you are actually certain that that's what it is, because it tends to cause confusion
[20:23] <cjwatson> hwilde: (and in any case the "New partition size:" bit wasn't gparted even in 6.06/6.10)
[20:23]  * ogra giggels about "gtk port of whiptail"
[20:24] <Nafallo> hehe
[20:24] <hwilde> take it easy - i'm not the one who made the wording ambiguous :/  people struggle with that and its very important to not mess up the partitions
[20:24]  * toresbe doesn't know how to reliably reproduce or file this weird bug.
[20:25] <toresbe> Using swf player in Firefox makes my monitor randomly jump out of sync.
[20:25] <cjwatson> hwilde: I know, I'm just asking not to use technical terms (gparted) unless you're sure. :-) "partitioner" would be fine.
[20:25] <ffm> How do I install all the builddepends of a package via APT?
[20:25] <hwilde> cjwatson, ok I will only speak in screenshots from now on - but you obviously knew exactly what I meant because you said it's fixed
[20:26] <cjwatson> astronut: the spinner in the manual partitioner is different from what hwilde was talking about; I think the problem is less severe there since there isn't an implicit other partition in that context, but there are bugs about that (e.g. bug 117986)
[20:26] <cjwatson> hwilde: I don't mean to go that far, plain English is fine
[20:26] <cjwatson> I've been confused in the past by people talking about gparted in bug reports, so I just wanted to make a general point, that's all, don't take it personally
[20:27] <hwilde> well you can either educate the entire universe, OR put a label in the window title that says it's not gparted, then people will use the correct terminology
[20:27]  * ogra points ffm to #ubuntu-motu for such questions .... apt-get build-dep <package> 
[20:27] <cjwatson> hwilde: uh, it's not called ls or openoffice.org either. :-)
[20:28] <hwilde> cjwatson, how would anyone know what to correctly call it without you telling them
[20:28] <hwilde> put a label in the window title
[20:28] <cjwatson> they can use a plain English word rather than guessing jargon
[20:28] <cjwatson> "installer"
[20:28] <cjwatson> the window title says "Install"
[20:28] <ffm> ogra: merci.
[20:28]  * hwilde stares at cjwatson
[20:29] <hwilde> if you want people to refer to it by the correct name, put that name on their screen and they will
[20:29] <cjwatson> I don't want people to refer to it by ANY NAME AT ALL
[20:30] <cjwatson> it's the installer. That word "Install" is on their screen.
[20:30] <hwilde> you act like I said bomb on an airplane
[20:30] <cjwatson> no, you're arguing ever more vociferously with me and I'm responding
[20:30] <cjwatson> only people who happen to remember that at one point the installer happened to use gparted for its advanced partitioning mode will be confused and say gparted
[20:31] <cjwatson> that small number of people can easily be reminded when they make a mistake, and it's not a problem
[20:31] <hwilde> gparted gparted gparted yeah I said it...  i'm sorry
[20:31] <cjwatson> putting jargon on the screen for the purpose of educating a small number of old-timers is not worth it
[20:31] <cjwatson> "Install" is all it needs
[20:32] <cjwatson> how am I supposed to correct people who make mistakes if you get really annoyed about it? seems bizarre, it was just a simple factual correction
[20:32] <hwilde> what does it use if not gparted
[20:32] <cjwatson> it's a layer over d-i's partman
[20:32] <sladen> mysterycool: #ubuntu-motu is a good place to start
[20:32] <cjwatson> the GUI part is written from scratch
[20:33] <hwilde> and how would anyone know this just by looking at it
[20:33] <cjwatson> they don't need to
[20:33] <cjwatson> nobody who is coming at it with no experience at all can possibly come up with the name gparted :-)
[20:34] <hwilde> ok I agree
[20:34] <cjwatson> the reason that we went for a from-scratch GUI is that the partitioning "business logic" is a lot harder to get right than the GUI
[20:34] <hwilde> I am sorry that I typed in "gparted" instead of "that nameless partition editor in the livecd installer setup screen when you are resizing an existing partitioner"
[20:34] <astronut> cjwatson: acks
[20:35] <cjwatson> and since we have to maintain two installers with different UIs, it makes sense to duplicate the UI not the underlying logic
[20:35] <hwilde> anyways it was a point about the wording of the instructions, not the underlying partitioner
[20:35] <hwilde> and you knew what I meant so there wasn't really any confusion
[20:35] <cjwatson> in this case, yes
[20:36] <hwilde> all i'm saying is that people frequently ask,   is this the new size of hte existing partition, or the size of the new partition being created
[20:36] <cjwatson> but I *have* spent extra time trying to untangle confusing bug reports where people used the gparted jargon incorrectly, and this wasted my time
[20:36] <astronut> cjwatson: will the breaking count if it's a data partition?
[20:36] <astronut> besides breaking an fstab
[20:36] <cjwatson> so I have a general policy of correcting people when they use it wrongly, so that eventually the meme can go away
[20:36] <astronut> but if i move an ntfs to a whatever, ti'll still just show up as the next availble letter
[20:36] <cjwatson> hwilde: right, this is why we spent some effort fixing it in 8.04
[20:36] <cjwatson> with a better resize widget
[20:37] <cjwatson> astronut: sorry, "breaking"?
[20:37] <hwilde> cjwatson, to avoid future confusion I strongly recommend putting a label on that screen, even if it's "Installer step 5 - Repartitioning"  then people can refer to it by the correct name
[20:37] <cjwatson> hwilde: sorry, I disagree and will not
[20:37] <cjwatson> with all due respect
[20:37] <astronut> 15:21 < cjwatson> astronut: you can't do a primary->logical conversion; you'd  break whatever operating system lives there
[20:37] <cjwatson> astronut: oh, right. Yes, it will break any /etc/fstab (etc.) that refers to it
[20:37] <astronut> that's relatively minimal... it'd be nice to have the option
[20:37] <cjwatson> etc. being potentially scripts that refer to /media/sda4/ or whatever
[20:37] <astronut> *shrug*
[20:38] <astronut> right
[20:38] <cjwatson> I do agree that it's all too possible to dig yourself into a hole with four primary partitions
[20:38] <astronut> in this case it was del
[20:38] <cjwatson> it just seems like a massive bazooka to give people so that they can point it at their foot, so I'm not sure
[20:38] <astronut> dell*
[20:38] <cjwatson> hmm, I thought Dell shipped with three primary partitions; that's certainly how mine came
[20:39] <astronut> cjwatson: hidin orsomething, i don'tknow
[20:39] <hwilde> the hidden restore partition :/
[20:39] <cjwatson> and indeed that's why we weakened the constraints on the auto-resize option in 8.04 to allow operation with three primary partitions
[20:39] <astronut> cjwatson: i'm not sure... thisone has a a tiny fat16, two large ntfs (C: and D:, d has label backup and had some empty folders and some data) and then a hidden fat32 w/ restore data
[20:39] <cjwatson> hmm, I might have to ask some Dell contacts about that
[20:40] <astronut> one folder was "Ghost backups"
[20:40] <astronut> but was empty
[20:40] <cjwatson> it's really awkward to do that without making at least one of them a logical partition
[20:40] <astronut> sounds like it was some backup solution
[20:40] <astronut> no kidding
[20:40] <cjwatson> I believe there'll be a couple of Dell people at UDS, so I'll corner them then
[20:40] <astronut> sounds likeit was designed for ghost to backup the C: into d:/ghost
[20:40] <astronut> UDS? ubuntu desktop summit?
[20:40] <cjwatson> Ubuntu Developer Summit
[20:40] <astronut> was close
[20:41] <astronut> i guess UbConf just doesn't have a nice ring
[20:41] <cjwatson> we went through a number of names :)
[20:41] <cjwatson> I don't think libparted provides a way to do that kind of conversion; you'd have to delete the partition (but leave data there) and create a new one with the same extents
[20:42] <cjwatson> sounds hairy at best
[20:42] <cjwatson> oh, hmm, actually it's not always possible
[20:42] <cjwatson> unless there's room before the partition, the extended partition table would have to go in the first couple of sectors
[20:42] <astronut> right, if disk is ful, you can' add the data
[20:43] <slangasek> you also have to either resize the partition before it on the disk, or move the partition, to accomodate the extra header for the logical partitions
[20:43] <cjwatson> so you'd literally have to shift everything forward 4KB or whatever, which means that if power fails in the middle of the operation, you lose all the data
[20:43] <cjwatson> right, slangasek put it more clearly
[20:43]  * slangasek nods
[20:44] <slangasek> er, not nodding that I put it more clearly, nodding that we're in agreement about the problem ;)
[20:44] <bd_> If a package in universe has been rebuilt without source changes (what in debian would be called a binNMU, not sure what the official term is in ubuntu), and a new version goes into debian, is a manual universe sync needed?
[20:44] <cjwatson> resizing is (I believe) safe in the sense that if power fails you're still OK; or at the very least the windows are much smaller
[20:44] <bd_> http://patches.ubuntu.com/g/gimp-resynthesizer/gimp-resynthesizer_0.15-2build2.patch <-- the diff of the package in question
[20:44] <cjwatson> bd_: no, autosync is only inhibited if the version number contains "ubuntu"
[20:44] <bd_> ah, okay
[20:44] <bd_> I'll just wait for the unfreeze then :)
[20:44] <cjwatson> bd_: the purpose of "build" is precisely to allow autosyncs :-)
[20:44] <astronut> why id someone do a sourceful upload of cyrus to do a rebuild by the way?
[20:44] <bd_> cjwatson: excellent :)
[20:44] <Chipzz> bd_: I don't think that's what debian calls a binNMU
[20:44] <astronut> do you guys not do binary rebuilds
[20:45] <cjwatson> astronut: we don't have binNMUs in Ubuntu, no
[20:45] <cjwatson> largely as an intentional policy
[20:45] <astronut> ah
[20:45] <bd_> Chipzz: well, a debian binNMU is usually accomplished by kicking it back to the buildds
[20:45] <Chipzz> hrrrrm wait
[20:45] <cjwatson> though it might be implemented one day
[20:45] <Chipzz> I'm thinking of a binary only upload
[20:45] <bd_> although in theory a DD could manually do a binary only upload of a package that's not their own
[20:45] <cjwatson> binNMUs are a special kind of binary-only upload
[20:45] <cjwatson> bd_: done it plenty of times. :-)
[20:45] <cjwatson> it's called a porting upload
[20:46] <Chipzz> bd_: a rebuild is not a NMU though
[20:46] <Chipzz> nor a binNMU I thin
[20:46] <Chipzz> k
[20:46] <Chipzz> NMU = Non Maintainer Upload
[20:46] <cjwatson> Chipzz: rebuilds are called binary-only NMUs in Debian
[20:46] <Chipzz> cjwatson: they are? I thought they were just referred to as rebuilds/retries
[20:46] <cjwatson> Chipzz: even if the maintainer requests it, these days it's actually done by buildd triggers, so it's still usually signed by somebody other than the maintainer
[20:46] <Chipzz> ie 'give back' a build
[20:47] <cjwatson> give-back is when it failed the first time
[20:47] <slangasek> historically, all porter builds are "binNMUs"
[20:47] <cjwatson> binNMUs are when you bump the version (which in Debian is possible without bumping the source version)
[20:47] <bd_> Chipzz: 'give back' means something different in debian - it's how you unwedge a stuck build, I believe :)
[20:47] <cjwatson> bd_: it means the same in Ubuntu, Chipzz is just a bit mixed up :)
[20:47] <bd_> ah, okay :)
[20:48] <cjwatson> we should probably have a glossary somewhere
[20:48] <Chipzz> yeah
[20:48] <Chipzz> 21:45 < Chipzz> I'm thinking of a binary only upload
[20:48] <slangasek> more recently, that became confusing and in Debian the term is almost exclusively used now to refer to building a package on a buildd with a changelog-only diff to get the binaries rebuilt against the current archive
[20:49] <cjwatson> right, originally Maintainer != signer => NMU; we've got a bit more sophisticated with terminology since
[20:49] <Chipzz> well, I think the term binNMU is confusing
[20:49] <bd_> Chipzz: it is, but it's one character shorter than rebuild, so :)
[20:49] <Chipzz> because an NMU usually involves changes the maintainer does not know about/has not given permission for
[20:49] <bd_> (not one /keystroke/ shorter though alas)
[20:49] <cjwatson> that's not true at all
[20:49] <cjwatson> it's quite common for a maintainer to authorise an NMU when they're unable to upload for some reason
[20:49] <Chipzz> hrrrm I should rephrase then
[20:50] <Chipzz> "changes not made by the maintainer"
[20:50] <bd_> and binNMUs also happen without the maintainer asking in many cases (library transitions etc)
[20:50] <slangasek> Chipzz: oh, I've more than once had a Debian maintainer be completely oblivious to binNMUs I've done :)
[20:50] <cjwatson> Chipzz: that's not true either, given Uploaders
[20:50] <cjwatson> an NMU is just that the person who made the upload is not one of the set of people who would ordinarily upload the package (originally just Maintainer, now Maintainer + Uploaders)
[20:50] <Chipzz> ugh, I should shut up then I think :)
[20:51] <bd_> cjwatson: however, the buildds are, in a way, part of the set of people(?) who would ordinarily upload the package... ;)
[20:51] <cjwatson> bd_: (of course using binNMUs for library transitions is a relatively recent innovation)
[20:51] <bd_> cjwatson: hm, really? Color me newbie then :)
[20:52] <cjwatson> back in the day, we didn't have enough central buildd control to schedule them across all architectures, so there was no point
[20:52] <Chipzz> cjwatson: it's just that I'm more accustomed to the old meanings I guess
[20:52] <Chipzz> well, meaning...
[20:52] <cjwatson> nowadays that's basically all sorted out so that the release team can do them across the board
[20:52] <cjwatson> Chipzz: I'm accustomed to an older meaning ;-)
[20:52]  * bd_ nods
[20:52] <bd_> now if only debian would take the plunge and move to sourceful uploads too... :)
[20:52] <slangasek> which required wanna-build to first grow binNMU support, yes
[20:53] <Chipzz> in my mind NMU's often were associated with "hostile" intentions
[20:53] <Chipzz> but that may be my flawed perception
[20:53] <cjwatson> that's really something that came later; originally NMUs were much more often friendly
[20:53] <cjwatson> which is one of the things we were trying to get back to in setting up Ubuntu the way we did
[20:54]  * cjwatson -> dinner
[20:54] <Chipzz> cjwatson: enjoy :)
[20:55] <bd_> is it me or did launchpad pick a moderately bad time for scheduled maintenance, what with toolchain uploads and etc going on? :)
[20:57] <slangasek> nah, that's /way/ better than having it scheduled the week of a release
[20:57] <LaserJock> there's never really a good time, IMO :-)
[20:58] <ogra> well, we're not really in a hurry, are we ?
[20:58] <stgraber> LaserJock: christmas break ? :)
[20:59] <jdong> ogra: speak for all y'all non-backports users ;-)
[20:59] <bd_> Well, there is the may 1 target for the toolchain upload.
[20:59] <ogra> jdong, heh
[21:08] <cjwatson> bd_: the toolchain's almost entirely done, so it's OK
[21:08] <cjwatson> dpkg is really the only thing left
[21:08] <bd_> nice :)
[21:10] <bd_> oO koffice2 is building on intrepid already?
[21:11] <bd_> [ https://launchpad.net/ubuntu/+source/koffice2/1:1.9.96.0~that.is.really.1.9.95.3-1ubuntu3/+build/572459 ]
[21:13] <cjwatson> bd_: might just be sparc trying to catch up
[21:13] <cjwatson> if it didn't make it on hardy (or even if it failed), it'll retry
[21:13] <bd_> ah
[21:15] <slangasek> yes, I knew when the buildds started tackling intrepid by the flood of build failure mails :)
[21:17]  * cjwatson wonders whether to keep dpkg-triggers upgrade handling from mid-gutsy
[21:18] <cjwatson> technically removable and would simplify the diff, but I do hate removing upgrade handling
[21:18] <bd_> cjwatson: Would pre-depending on an appropriate version of dpkg be enough?
[21:18] <cjwatson> what, in dpkg? :-)
[21:18] <cjwatson> (I'm merging dpkg)
[21:18] <bd_> oh wait, on dpkg's side :)
[21:19] <bd_> I thought you were referring to some kind of release goal for trigger-using packages :)
[22:09] <elmo> is there a standardized 'is this service running' check for Ubuntu initscripts?
[22:10] <slangasek> no
[22:10] <elmo> k
[22:11] <slangasek> there was discussion among the server team about standardizing on a 'status' command, but it was late in the cycle and deferred in light of overlapping plans for upstart status reporting
[22:11] <elmo> ok - well if it ever exists, it'd be nice if the glibc preinst used it
[22:11] <sladen> elmo: theoretically if you use Keybuk's fancy stuff you should be able to get that and more
[22:12] <elmo> as in, 'no please do not restart the IMAP server I just carefully stopped in prepartion for the dist-upgrade'
[22:12] <kees> jcastro: do you have upstream Blender contacts?
[22:12] <jcastro> kees: sure do
[22:12] <kees> jcastro: can you point them at bug 6671 ?
[22:13] <elmo> and pam
[22:13] <jcastro> kees: will do
[22:13] <kees> jcastro: thx
[22:14] <slangasek> elmo: well, pam at least lets you override the list :)
[22:15] <Keybuk> that's just a bug in Debian initscripts
[22:15] <Keybuk> restart should do just that
[22:15] <Keybuk> it should only stop and start it if it was running to begin with
[22:15] <Keybuk> never start it when it wasn't
[22:15] <Keybuk> :)
[22:15] <slangasek> ?
[22:16] <slangasek> the libpam postinst code doesn't rely on a functional "restart" command anyway; I don't remember why
[22:16] <slangasek> possibly because some init scripts get clever and don't actually re-exec the binary on restart; and possibly because that's not how "restart" is defined in debian policy?
[22:18] <slangasek> (policy does specify that the outcome of a successful 'restart' command is always to leave the service running at the end)
[22:18] <Keybuk> I think my point was that policy sucks :)
[22:19] <kees> I think "graceful" should be the 'restart-if-already-running'.
[22:19] <kees> although, that's probably a bad name too
[22:21] <Nafallo> kees: makes me think of apache :-)
[22:22] <infinity> "/etc/init.d/apache2 random" sound like fun to me.  Then we can just close every bug report with "well, what did you EXPECT it to do?"
[22:23] <kees> Nafallo: yeah, I was trying to think of a "restart"-like command that fails when it's not already running, and IIRC that's true for apache's "graceful"
[22:23] <Nafallo> kees: oh. never actually tried it with apache off to be honest :-)
[23:18] <soren> cjwatson: I just took at peek at bug 221635..
[23:19] <ted1> Sexist bug.
[23:19] <soren> cjwatson: I've not looked too closely, but I noticed that around line 720, you define another struct sigaction that you don't memset(0, blah, blah) first.
[23:19] <soren> ted1: :)
[23:20] <soren> ted1: I've been looking for you, by the way..
[23:20] <ted1> Uh, oh.
[23:21] <cjwatson> soren: true, but I set all the elements documented by sigaction(2), with the exceptions that (a) it says not to set both sa_handler and sa_sigaction (b) it says not to use sa_restorer
[23:21] <cjwatson> maybe you're supposed to zero sa_restorer anyway
[23:23] <soren> cjwatson: I'm not sure.
[23:23] <soren> cjwatson: It just jumped out at me and thought it might help.
[23:23] <cjwatson> looking at arch/x86/kernel/signal_32.c, I think it *might* get used even if you don't put SA_RESTORER in sa_flags
[23:24] <cjwatson> hard to be sure, but you could be right
[23:25] <cjwatson> another instance of that in pipeline_pump too
[23:26] <cjwatson> though code in gnulib doesn't bother to do the memset
[23:27] <cjwatson> I'll shove memsets in for the sake of paranoia, but unfortunately I'm not convinced :-/
[23:28] <cjwatson> dearly wish I could get an strace of the damn thing
[23:28] <soren> :)
[23:29] <Kopfgeldjaeger> n8
[23:31] <soren> cjwatson: Well, a hint as to which signal it's receiving would also be nice.
[23:32] <cjwatson> I suspect it's just as easy to get an strace as to get them to install a new version
[23:32] <cjwatson> and unfortunately stdio in signal handlers is not kosher
[23:32] <cjwatson> which makes safe debugging output tricky
[23:33] <soren> I'm not sure I realise what the problem is? At least not in a sigchld handler.
[23:34] <cjwatson> signals may be taken while stdio is in an inconsistent state. If you use stdio in a signal handler you may crash randomly depending on exactly when the signal was taken.
[23:34] <cjwatson> having just taken out code that used stdio (assert) for that reason, I'm not going to put it back in. :)
[23:36] <soren> Ah... Yes, good point.
[23:37] <cjwatson> http://www.opengroup.org/onlinepubs/009695399/functions/xsh_chap02_04.html#tag_02_04_03 lists the POSIX-defined functions safe for calling within a signal handler
[23:52] <crimsun> TheMuso: hmm, I still have system sounds (WRT 192888)
[23:54] <TheMuso> crimsun: Well I've only tested on my notebook, and when using dmix with pulse, the startup sounds doesn't play for example. Playing it from gnome-sound-properties gives a sample caching error.
[23:54] <crimsun> interesting, because that's the test I /just/ attempted :-)
[23:54] <TheMuso> I'll test on other machines I have here.
[23:55] <TheMuso> This was with the latest pulse stuff from bzr.
[23:58] <TheMuso> Let me test on my amd64 box with an hda-intel sound chip.