[01:05] <lamont> how do I tell growisofs to erase the dvd, I wonder?
[01:18] <slangasek> lamont: -Z ?
[01:19] <lamont> slangasek: it bitched and whined at me... I'll see if it the proceeds to DTRT
[01:19] <slangasek> well, -Z is how it's meant to happen anyway
[01:19] <lamont> ok
[01:20] <lamont> :-[ RESERVE TRACK failed with SK=5h/ASC=21h/ACQ=00h]: No space left on device
[01:20] <lamont> and then proceeded to write on the disk anyway
[02:38] <lamont> slangasek: now I remember... the burn finishes, but I have a coaster.
[02:48] <superm1_> slangasek, if you can release the mythtv build at some point when you're around, i'd appreciate it
[02:49] <Hobbsee> bah.  who actually needs releases?
[02:50] <superm1_> Hobbsee, you have magic keyboard keys though too don't you?
[02:50] <superm1_> can you release it :)
[02:51] <Hobbsee> superm1_: nope.  ENODRESCHER.
[02:51] <Hobbsee> superm1_: i don't work for canonical, so i don't have teh switches.
[02:51] <superm1_> boo :(
[02:51] <Hobbsee> superm1_: i don't even have the proxy switches anymore, although that was fun.
[02:51] <superm1_> what happened to those?
[02:51] <Hobbsee> superm1_: they found a RM
[02:52] <superm1_> oh i didn't realize they were temporary for that purpose
[02:52]  * Hobbsee never had DC access
[02:52] <Hobbsee> hi spam
[02:57] <emgent> heya
[03:22] <slangasek> lamont: ummmm, are these rewritable DVDs?  Or else why are you telling growisofs to zero out a DVD?
[03:23] <slangasek> Hobbsee: hrm?  You should have access to the unapproved queue through LP, has this changed?
[03:23] <Hobbsee> slangasek: i do for that.  i think it even works.
[03:23] <Hobbsee> slangasek: but i can't do releases, etc
[03:23] <slangasek> ok, I'm pretty sure that's what superm1_ was referring to... :)
[03:24] <Hobbsee> oh, i thougth he meant doing the mythtv release, and setting those cds as final
[03:52] <superm1_> Hobbsee, yeah i was meaning to pull 'mythtv' through the queue since its in multiverse and affected by beta freeze.  'mythbuntu' alternate disks are something completely different
[04:03] <jdong> ok it's official
[04:04] <jdong> using a simple bash loop to convert rcS.d to upstart does NOT work.
[04:05]  * jdong wonders if keybuk is having nightmares right now because of how I've been abusing upstart
[04:08] <ScottK2> Laughing more likely.
[04:08] <jdong> yeah, udev not starting is pretty painfully funny.
[04:09] <jdong> particularly on a macbook where the keyboard is USB and activated by udev.
[04:09] <jdong> I'll have to catch him sometime and pick his brain about this
[04:18] <ion_> Heh
[04:19] <ion_> Probably better wait for Upstart 0.5 before actually making the whole system boot on it.
[06:12] <fabbione> cody-somerville: pong
[06:48] <pitti> Good morning
[06:53] <nixternal> mornin' pitti
[06:53] <LaserJock> hiya pitti
[06:54] <LaserJock> nixternal: yo yo homeskillet
[06:54] <LaserJock> what the blazes are you doing up?
[06:54] <nixternal> not even 02:00 yet
[06:54] <nixternal> I will prolly go upstairs in a few and watch the boobtube
[06:55] <LaserJock> man, I'm toast
[06:55] <nixternal> hrmm, after spelling that, that doesn't look like a really PC way of saying television does it?
[06:55] <LaserJock> I got the "bright idea" that maybe I should rewrite my data fitting program for school in C++ rather than Python
[06:55] <Fujitsu> Hi LaserJock, nixternal, pitti.
[06:55] <nixternal> hiya Fujitsu
[06:55] <nixternal> LaserJock: that's a smart man :)
[06:56] <LaserJock> nixternal: well, I had a look at the library I need and am thinking that might not be the best idea ;-)
[06:56] <nixternal> which library?
[06:57] <LaserJock> GNU Scientific Library
[06:57] <nixternal> boost has a nice scientific library :)
[06:57] <nixternal> boost might be better documented and more recent too
[06:57] <LaserJock> well, it's specifically the "it's written in C" part that's getting me
[06:57] <LaserJock> I'm not sure I feel like really learning C/C++ in 2 days
[06:58] <nixternal> int main(int argc, char* argv[]) { // go to town!!! } :p
[06:58] <nixternal> I have been trying to come up with an app so I can better learn Python myself
[06:59] <LaserJock> python's great for learning
[06:59] <nixternal> for some reason, I found C/C++ and Java easier to learn
[06:59] <LaserJock> but I am trying to do some data analysis and C/Fortran are the tools for the job for the most part
[07:00] <nixternal> F77 baby! :p
[07:00] <LaserJock> yeah, that's what my advisor would use
[07:00] <LaserJock> he hasn't quite gotten to F90
[07:01] <LaserJock> but C is pretty much just as good for what I've got to do and a bit more ubiquitous
[07:02] <nixternal> that it is
[07:02] <LaserJock> my program is < 500 lines in python though, I hate to know how much more it'd be in C :/
[07:03] <nixternal> someone in #kubuntu-devel said to change Ubuntu 8.04 LTS to Ubuntu 8.04 as there will not be an LTS until the point release...rumor or fact? and if rumor who is spreading it besides the guy in #kubuntu-devel?
[07:03] <dholbach> jamesh:
[07:03] <nixternal> hola dholbach
[07:03] <jamesh> dholbach: hmm?
[07:03] <Fujitsu> I heard something similar a while back, but I don't recall it being from a reputable source.
[07:04] <nixternal> ahh, so there is a rumor mill then...first I heard it was an hour ago and he never responded to tell me where he heard it
[07:04] <dholbach> jamesh: oops, sorry - nevermind
[07:04] <dholbach> good morning :)
[07:04] <dholbach> hi nixternal
[07:24] <saivann> pitty : Yes, it was an oversight (I was sure that I verified it twice). At that point, we will have to fix it with another upload I guess?
[07:51] <dwa> hi, if i click the new message button in evolution my cpu goes nuts untill i close the whole program. Top show evolution and gconfd-2 eating up my cpu. Does anyone know how to fix this?
[08:22] <pitti> saivann: right
[08:27] <seb128> hello pitti
[08:27] <pitti> hi seb128
[08:33] <pitti> dholbach: http://daniel.holba.ch/5-a-day-stats/> oh, so adding more than 5 per day is ok? :)
[08:33] <dholbach> of course :)
[08:35]  * pitti reads about five-a-day-applet ... awesome!
[08:35]  * pitti hugs dholbach
[08:35] <dholbach> pitti: thekorn's work :)
[08:35]  * pitti hugs thekorn
[08:38] <thekorn> :)
[08:47] <pitti> mdke: just answered your u-docs problem
[09:16] <seb128> pitti: the retracer seems to just untag some bugs recently, I've tried to look to it but without luck, apport-retrace -s bug_number wants a cookie for some reason, it's not supposed to, right?
[09:17] <pitti> seb128: wants a cookie> p-lp-bug
[09:17] <pitti> seb128: hm, just untag? does the log say anything? an exception or so?
[09:20] <seb128> pitti:
[09:20] <seb128> "  File "/usr/lib/python2.5/site-packages/apport/report.py", line 429, in add_gdb_info^M
[09:20] <seb128>     assert os.path.exists(self.get('InterpreterPath', self['ExecutablePath']))^M
[09:20] <seb128> AssertionError^M
[09:20] <seb128> Exception exceptions.ImportError: ImportError('No module named shutil',) in <bound method __AptDpkgPackageInfo.__del__ of <\
[09:20] <seb128> apport.packaging_impl.__AptDpkgPackageInfo instance at 0x839f8ac>> ignored^M
[09:20] <seb128> 03/16/08 23:30:12: retracing #203017 exit status: 0^M
[09:20] <seb128> "
[09:20] <seb128> pitti: that's on example of bug which has just been untagged
[09:20] <pitti> oh, ugh
[09:20] <pitti> seb128: can you please stop it for now?
[09:20] <pitti> seems it does not install the affected Package: properly
[09:20] <seb128> pitti: stopped, i386 and amd64
[09:26] <iwkse> hi all, I was searching for a duplicate of a bug i'm experiencing but i could not find. In few words, livecd can't boot and give the (initramfs) prompt when is used a usb-cdrom device. The problem seems to be due to initramfs which mount too fastly the new root before the cdrom is probed. It's a known bug?
[09:38] <MacSlow> Greetings everybody!
[09:41] <[[thufir]]> where does the Make and Model:		Samsung ML-2550, 1.0  driver come from which Ubuntu installs?
[09:53] <blackjack> hi
[09:53] <[[thufir]]> hi
[09:54] <[[thufir]]> I'm just trying to figure out what ubuntu does differently to get this printer working.  some sort of special driver?
[10:28] <asac> superm1_: http://bazaar.launchpad.net/%7Eubuntu-core-dev/network-manager/ubuntu.0.6.x ... please test the ath_pci tweaks
[10:28] <asac> superm1_: (revno 100)
[10:32] <twi_> hey pitti, lool, thanks for your work re #186647
[10:32] <asac> if anyone else here uses ath_pci and has problems with nm in hardy, please test the above branch as well :)
[10:32] <pitti> twi_: you're welcome
[11:49] <soren> ogra_cmpc: I managed to redo the grub installation stuff, by the way. Now grub does all the hard work for me. Do you still need something like that?
[11:52] <ogra_cmpc> soren, for intrepid, i resorted to syslinux
[11:53] <ogra_cmpc> and for hardy i wont change that anymore
[11:55] <soren> ogra_cmpc: Ok. It turned out to be quite simple.
[11:55] <ogra_cmpc> right, but it would require me to change the whole design
[11:56] <soren> Sure. Well, if you want to change back for intrepid, give me a shout.
[11:56] <ogra_cmpc> being bound to vfat forces certain things on me for the installer image i build (i.e. an initramfs that uses images instead of partitions for cow)
[11:56] <ogra_cmpc> i will,  for sure
[11:57] <cristi> hello
[11:57] <ogra_cmpc> even though intrepid might be the grub2.0 relese, who knows :)
[11:57] <cristi> i would like to talk to somebody about an openal replacement
[11:57] <cristi> more specifically, something that would solve https://bugs.launchpad.net/ubuntu/+source/openal/+bug/194919
[11:57] <ogra_cmpc> that would solve such probs at the base
[11:57] <ubotu> Launchpad bug 194919 in openal "libopenal needs replacement" [Undecided,New]
[11:58]  * ogra_cmpc twiddles thumbs looking at 
[11:58] <ogra_cmpc> Setting up scrollkeeper (0.3.14-15ubuntu1) ...
[11:58] <ogra_cmpc> Rebuilding the database. This may take some time.
[11:59] <ogra_cmpc> sigh ... that thig makes up nearly a quater of the overall image buildtime
[12:00] <cjwatson> ogra_cmpc: do you divert it during image build?
[12:00] <cjwatson> (i.e. does it run more than once, or just from scrollkeeper.postinst?)
[12:00] <ogra_cmpc> the cron entry
[12:01] <ogra_cmpc> but beyond that i dont ... i didnt have the balls to make it inexecutable from initramfs :)
[12:14] <pitti> dendrobates, soren: btw, is anyone in the server team looking after dhcp3 bugs? if not, I think we need to find half-a-maintainer for it
[12:15] <ogra_cmpc> pitti, i always kept half an eye on it due to it being essential to ltsp
[12:19] <soren> pitti: Hm... For some reason ubuntu-server is not a bug contact for it. We should be.
[12:34]  * lamont would like something better than "chmod o+r /var/lib/misc/shadow.db" for making screen unlocking work on gutsy (with shadow: compat db in nsswitch.conf)
[12:48] <dendrobates> pitti: not currently, but I think that it makes sense for us to support it.
[12:57] <soren> dendrobates: Do you want me to subscribe ubuntu-server to dhcp3 bugs?
[12:57] <soren> dendrobates: I've got the page open, I just need to click "Ok.. Go!".
[12:59] <pitti> mdke: wrt bug 110983, where does ubuntu-docs get its translations from?
[12:59] <ubotu> Launchpad bug 110983 in language-pack-gnome-de "translation english-german" [Undecided,Confirmed] https://launchpad.net/bugs/110983
[12:59] <pitti> mdke: I'd like to fix it in LP translations
[13:08] <dendrobates> soren: Ok, sæt i gang!
[13:09] <soren> Done.
[13:22] <seb128> hum
[13:22] <seb128> is
[13:23] <seb128> somestruct = function(somestruct->option) something safe in C?
[13:24] <broonie> seb128: Yes.
[13:24] <seb128> hum, ok
[13:24] <seb128> I'm wondering what's going on there
[13:25] <seb128> I've a case where it crashes when using -O2 or -O1 but it works using -O0
[13:26] <seb128> and the function parameter changes between the call and the function apparently
[13:26] <seb128> that is really weird
[13:28] <ScottK> doko: When you dropped python-xml to suggests for python-reportlab did you consider that solves the issue of python-xml removal for that package or is there more work that needs to be done?
[13:29] <cjwatson> ogra_cmpc: hmm. Does anything in Edubuntu care about tasks any more? That is, would it matter if Task: edubuntu-desktop because Task: edubuntu-desktop-addon?
[13:29] <cjwatson> since that's what the seeds currently want it to be
[13:31] <cjwatson> seb128: that works (see the C FAQ section 2.7), but note that data in the structure is copied by reference, not by value, so if anything in the structure was filled in using pointers in the stack in that function then its value will be corrupt
[13:31] <cjwatson> s/will/may/
[13:32] <seb128> cjwatson: http://paste.ubuntu.com/5772/ is basically the concerned code
[13:32] <cjwatson> seb128: e.g. struct foo func(void) { struct foo bar; char baz[32]; ...; foo.baz = baz; return foo } is broken
[13:33] <cjwatson> seb128: oh, that's returning a pointer, not the whole struct. Should be perfectly safe.
[13:33] <seb128> alright
[13:33] <seb128> I hate this bug
[13:33] <cjwatson> unless RROutput is a struct?
[13:33] <seb128> output->info changes between the call and the function
[13:33] <cjwatson> oh
[13:33] <cjwatson> you have no sequence point in there
[13:34] <cjwatson> *and* you redefine output in a nested scope
[13:34] <cjwatson> asking for trouble
[13:34] <cjwatson> seb128: what is "output->info" in the function arguments? more relevantly, what is "output"?
[13:36] <seb128> cjwatson: output is a RWOutput struct
[13:36] <cjwatson> seb128: trying to figure out how to explain this
[13:36] <seb128> cjwatson: if you want to see the code, apt-get source gnome-desktop; debian/rules apply-patches; editor libgnome-desktop/randrwrap.c
[13:36] <cjwatson> seb128: no, it's clear what's wrong from the paste
[13:37] <cjwatson> seb128: "output" must be declared in some outer scope, otherwise that code wouldn't work
[13:37] <seb128> cjwatson: that's the infamous "gnome-settings-daemon crashes on some cards" bug which has a zillion duplicates in the code bryce brought for the xrandr1.2 capplet
[13:38] <seb128> let me read carrefully what you wrote and think a sec about it ;-)
[13:38] <cjwatson> seb128: it would probably work a lot better if you had "RWOutput *rw_output = rw_output_by_id (output->info, info->clones[i]);"
[13:38] <cjwatson> seb128: does that clarify it?
[13:38] <seb128> yes, I did that change in fact
[13:38] <seb128> and that fixes the bug
[13:38] <seb128> I was trying to figure why now
[13:39] <seb128> alright, makes sense after some thinking
[13:39] <cjwatson> I'm betting that it isn't defined which scope output->info is resolved in
[13:39] <cjwatson> (after some thought, sequence points aren't relevant here)
[13:39] <seb128> cjwatson: https://bugs.edge.launchpad.net/ubuntu/+source/gnome-control-center/+bug/198951 current comment is the valgrind error
[13:39] <ubotu> Launchpad bug 198951 in gnome-control-center "gnome-display-properties crashed with SIGSEGV in screen_info_new() - i855, fglrx, radeonhd" [High,In progress]
[13:40] <jdong> stupid upstart question: is "stop on shutdown" assumed to happen or must be explicitly stated?
[13:40] <jdong> i.e. will a post-stop script atuomatically run on shutdown?
[13:40] <seb128> l606 is "	RWOutput *output = rw_output_by_id (output->info, info->clones[i]);"
[13:43] <cjwatson> seb128: annoyingly, I've been unable to find chapter and verse, but it's clear to me that at the very least renaming the inner-scope "output" will make it a lot easier to debug
[13:43] <cjwatson> maybe to clone_output rather than rw_output, looking at the context
[13:44] <cjwatson> in theory you might also want to drop the "RWOutput *" from the start of the line (the exact meaning is ambiguous), but from context I don't think that would be right
[13:45] <seb128> cjwatson: so using
[13:45] <seb128> RWOutput *new_output;
[13:45] <seb128> new_output = rw_output_by_id (output->info, info->clones[i]);
[13:45] <seb128> ?
[13:47] <slytherin> pitti: May I suggest a small correction to MIR for elisa? You don't have to manually edit config files for adding folders of music. D&D from nautilus is supported.
[13:48] <cjwatson> seb128: it can be "RWOutput *new_output = rw_..." if you like; using a different variable name should be enough
[13:48] <cjwatson> seb128: (and, obviously, use new_output rather than output for the rest of that block)
[13:49] <seb128> cjwatson: yes, right, that's what I tried before asking, that does the trick, I was just not sure of why because the function doesn't do any change, it just iterates through the structure and return a pointer which seemed to be fine to me
[13:50] <seb128> cjwatson: thanks
[13:54] <slytherin> I have one question. Synaptic does not yet display the 'Homepage:' field from control file. It is becoming increasingly difficult to know the homepage of an app. Is this known already? Or should I file a bug?
[13:56] <cjwatson> seb128: I think the reason that fails is that output->info refers to the output you've just declared, which is uninitialised
[13:57] <ScottK> pitti: I was wondering if you'd have time to look into fixing Bug #198183.  Becuase of it I'm afraid I'm missing lots of guidance crash bugs.  I've added several patches recently and this has me concerned.  Additionally, my experience with Guidance is that if I can get it to not crash it will eventually DTRT, so finding these bugs could make for a big difference at release.
[13:57] <ubotu> Launchpad bug 198183 in apport "Apport dies when it tries to catch displayconfig bug" [Medium,Confirmed] https://launchpad.net/bugs/198183
[13:57] <cjwatson> seb128: exactly what's in that memory might depend on the optimisation level, hence hard-to-reproduce failure
[13:57] <cjwatson> s
[13:59] <seb128> cjwatson: right, I though the declaration and assignation were supposed to be done after the function return but that seems to not be the case when using optimization
[14:01] <mvo_> slytherin: thanks for the reminder, its sort of known, but a bugreport is probably a good idea
[14:01] <cjwatson> seb128: I think you're deep in undefined behaviour territory
[14:01] <slytherin> mvo_: will do after about 2 hours. :-)
[14:01] <cjwatson> seb128: C99 is not terribly explicit, but I think the scope actually begins with the first mention of the identifier in the scope
[14:02] <cjwatson> seb128: under optimisation, it might reuse the same register for it or something
[14:02] <cjwatson> seb128: which would mean you'd get a pointer to the old structure by dumb luck
[14:02] <seb128> cjwatson: alright, anyway I think that was rather an error than upstream picked the same variable name
[14:03] <seb128> I've fixed that locally, testing in a second and uploading
[14:03] <seb128> nice to get this one fixed ;-)
[14:03] <siretart> do we really want to have /etc/init.d/checkfs.sh fail that badly if fsck is unable to resolve an UUID? (filed as bug #203165)
[14:03] <ubotu> Bug 203165 on http://launchpad.net/bugs/203165 is private
[14:05] <siretart> err, that bug doesn't seem private to me.. hmrpf
[14:05] <cjwatson> seb128: I'm told by somebody with a newer revision of C99 than me that it says "Any other identifier has scope that begins just after the completion of its declarator" which is pretty clear
[14:05] <cjwatson> ("other" means "not a label" in context)
[14:06] <seb128> alright
[14:06] <seb128> so the code is confusing and wrong ;-)
[14:06] <cjwatson> apparently so :-)
[14:07] <cjwatson> interesting digression into C99 innards though
[14:07] <seb128> indeed
[14:08] <soren> Could someone please give back virt-manager on i386 and amd64, please?
[14:10] <soren> cjwatson: Thanks for looking into the JeOS thing. *Very* much appreciated.
[14:54] <superm1_> asac, that branch fixes my card perfectly
[14:54] <superm1_> i'll add a note to the bug in case you miss my response here
[14:57] <emgent> heya people
[14:58] <asac> superm1: thanks
[14:58] <asac> superm1: can you test suspend/resume?
[14:59] <asac> superm1: and if possible please test if it can connect to hidden AP
[15:05] <hellboy195> asac: news about nspluginwrapper?
[15:06] <asac> hellboy195: didn'T i sign it of?
[15:06] <hellboy195> asac: nooo :)
[15:06] <asac> hellboy195: bug id?
[15:07] <hellboy195> asac: bug #196540
[15:07] <ubotu> Launchpad bug 196540 in nspluginwrapper "Merge nspluginwrapper 0.9.91.5-2 from Debian(Unstable)" [Wishlist,Confirmed] https://launchpad.net/bugs/196540
[15:07] <asac> hellboy195: ok i am doing a lunch break now. will look once i return
[15:08] <hellboy195> asac: k ,thank you
[15:17] <soc> hi
[15:18] <soc> currently on gutsy the fx3 builds have crappy font rendering because they don't use the system's cairo lb but their own ... will this fixed atleast in hardy?
[15:29] <pitti> ScottK: ugh, WTH? sys.argv not existing? weird
[15:30] <seb128> anybody wanting to approve my nautilus, gnome-desktop-, pygobject uploads?
[15:30] <seb128> the nautilus one is a change to the "connect to server" dialog to make it actually mount locations, it's limited to the dialog and comes from upstream svn
[15:31] <seb128> the gnome-desktop one fixes gnome-settings-daemon crashing on startup in xrandr code which is a crasher recently lot of users have recently
[15:31] <seb128> the pygobject change is a new revision from debian fixing the -dbg build
[15:43] <pitti> seb128: looking (you weren't online when I saw your q on -desktop)
[15:45] <pitti> seb128: pygobject isn't un unapproved; wasn't that a sync?
[15:47] <seb128> pitti: I did a sync yes, is that an issue, ie does it bypass the freeze?
[15:47] <pitti> seb128: yes, syncs go straight through
[15:47] <pitti> other two accepted
[15:47] <seb128> pitti: ok, sorry about that, thanks
[15:52] <jwendell> seb128, Hi, have you seen my comment in a tsclient bug?
[15:53] <seb128> jwendell: the comment is in my bugmails box, it lacked context to be useful and I didn't open the corresponding webpage yet
[15:53] <seb128> jwendell: ie, it's on my unread stack
[15:53] <jwendell> seb128, it's about a crasher bug fixed on upstream, which just needs a backport...
[15:54] <seb128> jwendell: will do, thanks
[16:00] <ScottK> pitti: All I know about that apport issue is that I found lots of Guidance bugs with that problem.
[16:00] <mario_limonciell> Keybuk, your changes for fingerprint reader didn't make it into the pam that got recently uploaded it looks.  Still going to come in post beta?
[16:02] <Keybuk> mario_limonciell: there are some security implications of the fingerprint reader support that make it difficult to justify it being enabled by default without a lot more work
[16:02] <bryce> morning
[16:03] <Keybuk> specifically it is possible to change the registered fingerprint as long as you have access to a user's account
[16:03] <Keybuk> (it is not possible to change a user's password in the same manner)
[16:05] <mario_limonciell> Keybuk, ah i see
[16:06] <mario_limonciell> Keybuk, would a debconf question to pam be more feasible?
[16:06] <mario_limonciell> to turn it on and off then
[16:06] <Keybuk> mario_limonciell: the PAM change is just one line added to a config file, it will definitely be possible to document that in a HOWTO (along with the security implications)
[16:08] <mario_limonciell> Keybuk, but if debconf could/would be possible, this change could then be at least preseeded..
[16:09] <Keybuk> mario_limonciell: I think that would have component implications though
[16:09] <Keybuk> since pam is required, whereas debconf is merely important
[16:09] <Keybuk> I'd have to check with slangasek about that
[16:11] <mario_limonciell> ah i see.  the interdependencies of those really low underlying sorts of things do end up getting forgotten (minimal,standard,desktop).  Perhaps in the postinst checking for the presence of debconf, and if it's available to do that $stuff, but otherwise normal postinst
[16:11] <warren_> hi
[16:11] <jpatrick> hi warren_
[16:11] <warren_> can someone look at this very important bug report i've just created: https://bugs.launchpad.net/ubuntu/+source/gdebi/+bug/203199 ?
[16:11] <ubotu> Launchpad bug 203199 in gdebi "Gdebi(-Kde) can't install any package!" [Undecided,New]
[16:11] <warren_> it would be nice to fix for hardy!
[16:12] <jpatrick> warren_: better place would be #kubuntu-devel and it's known...
[16:12] <warren_> ok
[16:12] <warren_> are they fixing it?
[16:12] <Keybuk> mario_limonciell: the amount of testing that would require, and the very core level of that package, makes me somewhat nervous
[16:12] <Keybuk> especially post-beta :-/
[16:12] <mario_limonciell> :(
[16:13] <Keybuk> mario_limonciell: my plan instead was to ship a binary in thinkfinger-tools that turned it on (making the PAM change)
[16:13] <mario_limonciell> Keybuk, well that's very acceptable
[16:13] <Keybuk> (well, it'd be a very simple script in fact :p)
[16:14] <mario_limonciell> did you have a plan for when you were going to implement that?
[16:14] <Keybuk> mario_limonciell: already done, will be uploaded post-freeze
[16:14] <Nafallo> next cycle is a crap cycle, no? :-)
[16:15] <mario_limonciell> Keybuk, thinkfinger-tools is universe though is it not?  could go in through the freeze?
[16:15] <Keybuk> mario_limonciell: it has an MIR open
[16:15] <Keybuk> so it's "on the way to main"
[16:15] <mario_limonciell> ah
[16:16] <mario_limonciell> well we'll look forward to seeing that post beta then.  Thanks :)
[16:17] <\sh> who must I bribe to let the last upload of lighttpd flow into our archives? ;)
[16:17] <cjwatson> Keybuk: debconf is Priority: required, whatever its control file might say
[16:18] <Keybuk> d'oh
[16:18] <cjwatson> joeyh doesn't like it much, so it's never been reflected in debian/control, but it has to be required to make debootstrap work
[16:18] <cjwatson> (given the current implementation)
[16:18] <Keybuk> you'd think that I, of all people, would know not to confuse the output of dpkg -s with reality
[16:19] <cjwatson> interesting, that's one case where I'm pretty sure it's been like that since Ubuntu started
[16:19] <infinity> At least, yes.
[16:19] <cjwatson> does dpkg grab that from control in the .deb, then?
[16:28] <Keybuk> cjwatson: yeah
[16:29] <Keybuk> dpkg -s famously doesn't include details of overrides ;)
[16:44] <Riddell> calc: what makes openoffice chose its widget style?  I'm wondering if the kde 3 style will be chosen when running under kde 4
[16:53] <calc> Riddell: not completely sure, i'll have to look around at the code to see, there is a known bug in the fallback for icons though that I need to fix
[16:53] <Riddell> calc: I don't suppose you've heard of anyone doing an oxygen openoffice theme?
[16:55] <cjwatson> calc: I happened across the relevant code last week; it looks for various KDE variables in the environment, I believe
[16:55] <\sh> slangasek: could you approve lighttpd 1.4.19-0ubuntu2? :)
[16:55] <calc> Riddell: no, but then I don't use KDE so haven't been actively looking for one either
[16:55] <calc> cjwatson: ah ok
[16:56] <cjwatson> and it only had one case for KDE
[16:58] <cjwatson> calc: see vcl/unx/source/plugadapt/salplug.cxx; it looks for KDE_FULL_SESSION in the environment, and also checks whether the netwm atom is "KWin"
[16:58] <cjwatson> then the theme is in vcl/unx/kde/
[16:59] <calc> cjwatson: thanks :)
[16:59] <calc> i am a bit over halfway done sending all the bugs upstream :)
[17:00] <calc> then need to make a second pass to deal with the 36 remaining new bugs
[17:22] <calc> most of the bugs i have sent upstream are marked to be fixed for ooo 3.0, whee :-)
[17:32] <Amaranth> calc: it seems like _every_ bug sent upstream is marked for OOo 3 :P
[17:41] <slangasek> bryce, tjaalton: why is xserver-xorg-video-i740 a dependency of -video-all on i386, but not on amd64?
[17:44] <bryce> slangasek: hmm, not too familiar with i740, let me take a peek
[17:47] <slangasek> bryce: not that I should complain too loudly, currently this is one of the packages that make the difference between viable amd64 desktop CDs and oversized i386 CDs ;)
[17:47] <bryce> heh
[17:48] <slangasek> calc: fix for the uninstallable openoffice.org-l10n-en-us is still in the pipe?
[17:48] <bryce> well it supports such cards as the famed "Spacewalker Hot-158" video card, how could we neglect that?!?  ;-)
[17:49] <bryce> slangasek: anyway, it says it supports 'any' architecture, so I've not yet spotted a reason why it shouldn't be included for amd64
[17:50] <bryce> it looks like it's a driver for a variety of bargain bin video cards which presumably would work on either
[17:51] <bryce> I do wonder though if -intel supersedes i740 like it did for i810, but I've never heard talk to that effect
[17:51] <bryce> I can dig further if you'd like?
[17:51] <slangasek> I know nothing about i740, I was just looking at the diff and that one seemed odd
[17:52] <bryce> it looks like at least for edgy, amd64 was targeted:  https://launchpad.net/ubuntu/edgy/amd64/xserver-xorg-video-i740
[17:52] <slangasek> maybe the i740 was only ever used as an integrated chipset on 32-bit boards
[17:53] <bryce> hmm, possibly, but from the card names in the readme "Joymedia Apollo 7400", "Machspeed Raptor i740", "Winfast S900 i740", etc. they sound a tad more like discrete consumer boards
[17:54] <Amaranth> is there a reason usplash isn't using uvesafb?
[17:54] <Amaranth> isn't the i740 one of the original 3D accelerator cards?
[17:54] <Amaranth> like, from the quake 1 days
[17:55] <bryce> the driver's readme is dated 1999 so it seems to come from that era
[17:56] <Amaranth> i suppose the reasoning there is "why would you put one of these in an amd64 box?"
[17:56] <Amaranth> if the amd64 box would even have a slot for it
[17:56] <bryce> heh
[17:57] <Amaranth> iirc, carmack really loved that card :)
[18:00] <_MMA_> Since we're on graphic issues here and Amaranth mentioned Usplash, https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2008-March/003499.html Any ideas?
[18:07] <jdstrand> mathiaz: thinking about the apparmor/upgrade/complain issue
[18:07] <jdstrand> mathiaz: here is my thinking so far:
[18:09] <jdstrand> mathiaz: "$2" lt <first version of app where a profile existed> == complain
[18:09] <jdstrand> 'profile existed' means either in the package or apparmor-profiles
[18:09] <jdstrand> that is the easy case-- use complain when upgrading froma version where we didn't have a profile for it
[18:10] <jdong> upstart officially boots me now
[18:10] <jdong> Keybuk: pre-poke, I'll need consultation in a bit
[18:10] <jdstrand> mathiaz: the other easy case-- don't do anything special for version higher than the first version that shipped the profile
[18:10] <jdstrand> (eg hardy -> intrepid)
[18:11] <jdstrand> mathiaz: the harder case is what to do if going from apparmor-profiles to profile in the package
[18:11] <jdstrand> mathiaz: eg bind9 upgrade from gutsy -> hardy
[18:11] <jdong> Keybuk: what I wanted to ask was, if I define states using only a pre-start script and the service keyword, startup is still parallelized, right?
[18:12] <jdstrand> mathiaz: checking for dpkg-old seems error prone, as dpkg-old could be really quite old
[18:13] <calc> slangasek: yea still in the pipe, i'll be uploading it in the next week or so
[18:13] <calc> Amaranth: a lot of bugs in upstream bts are either not confirmed at all or marked later (as in they will never get fixed) :-\
[18:13] <jdstrand> mathiaz: I could stat it, but that seems hacky
[18:14] <mathiaz> jdstrand: what would you do in the case? put in complain or in enforce mode ?
[18:14] <jdstrand> mathiaz: that is what I am thinking about
[18:14] <Amaranth> jdong: whoa, you've got a full upstart boot?
[18:14] <jdstrand> mathiaz: I'd like to peek inside the old version, then make a decision
[18:15] <Keybuk> jdong: -> #upstart
[18:15] <jdstrand> mathiaz: by old version, I mean the existing profile that apparmor was using prior to the upgrade
[18:16] <jdstrand> mathiaz: but that won't get me what I want either
[18:16] <jdstrand> mathiaz: I could just always put these into complain mode
[18:17] <jdstrand> mathiaz: that is safest from a usability point of view
[18:17] <cjwatson> pitti: bug 199129 hurts my brain. This looks like hal. Are we supposed to call hal-disable-polling on each drive to get it to go away?
[18:17] <ubotu> Launchpad bug 199129 in ubiquity "Auto-resize install fails to mount drive" [Undecided,New] https://launchpad.net/bugs/199129
[18:17] <mathiaz> jdstrand: yes - I think the main reason we're doing this is to not break existing configuration
[18:18] <jdstrand> mathiaz: so I could just do "$2" lt <first version an enforcing profile was shipped> == complain
[18:18] <jdstrand> that logic is simple, and should not break anything
[18:19] <jdstrand> mathiaz: it will turn off a previously enforcing profile though
[18:19] <mathiaz> jdstrand: yes
[18:19] <mathiaz> jdstrand: which may not be good
[18:20] <mathiaz> jdstrand: could you grep in the profile for enforce mode ?
[18:20] <cjwatson> pitti: although hal-lock looks sort of promising too
[18:21] <pitti> cjwatson: hi
[18:21] <jdstrand> mathiaz: so really, what I want is "$2" lt <first version an enforcing profile was shipped && <user said to keep the existing profile" == noop
[18:21] <pitti> cjwatson: disable-polling? that's just for CD-ROM drives
[18:21]  * pitti looks at bug
[18:21] <jdstrand> mathiaz: or flip that and say ... && <user said to use the new profile> == complain
[18:22] <pitti> cjwatson: right, the old automount bug
[18:22] <pitti> cjwatson: until gutsy, gnome-volume-manager does the automounting, now this is triggered by nautilus
[18:22] <cjwatson> pitti: it sets media_check_enabled to false, which seems to be checked in the fdi that invokes hal-storage-mount
[18:22] <jdstrand> mathiaz: I could grep for enforce mode (which is what I was thinking), but if they had enforcing, and get the new profile, something could break
[18:22] <cjwatson> we're already setting the relevant nautilus gconf settings
[18:22] <mathiaz> jdstrand: I'm not convinced you need to use first version an *enforcing* profile was shipped
[18:23] <pitti> cjwatson: with the recent fix in casper to not ask for PK passwords for 'ubuntu' any more, you won't even get the dialog any more
[18:23] <mathiaz> jdstrand: right - but they would be prompted for a configuration file change.
[18:23] <pitti> cjwatson: hm, weird; we entirely ripped out mounting from g-v-m
[18:23] <mathiaz> jdstrand: it's up to them to merge correctly then.
[18:23] <pitti> and aside from nautilus, nothing else does automount
[18:23] <jdstrand> that is a good point
[18:23] <mathiaz> jdstrand: that doesn't change from any other configuration file change
[18:24]  * jdstrand nods
[18:24] <cjwatson> pitti: the password isn't the bug
[18:24] <pitti> cjwatson: right, I know
[18:24] <cjwatson> pitti: I think hal's doing the automount all by itself, without nautilus intervention
[18:24] <pitti> cjwatson: no, hal never automounts
[18:25] <mathiaz> jdstrand: I think you need: $2 lt <first version where an apparmor profile was available> then complain
[18:25] <pitti> cjwatson: however, I think running ubiquity under hal-lock is the canonical way to block these things
[18:25] <pitti> instead of writing fake fdi files, meddling with gconf, etc.
[18:25] <cjwatson> or maybe just the partitioning bit. Yes, that was sort of what I was looking at
[18:26] <jdstrand> mathiaz: yes, that was the first case I mentioned
[18:26] <jdstrand> eg dapper - hardy
[18:26] <jdstrand> I absolutely agree
[18:26] <cjwatson> actually, the whole of ubiquity would probably be fine; I don't want it bailing out if it can't start the partition manager ...
[18:26] <jdstrand> then nothing for the other cases-- up to user to merge the changes?
[18:26] <pitti> cjwatson: at least this prevents mounting at a more fundamental level than gnome, thus also applies to KDE, and should be more stable against changes in gnome
[18:26] <jdstrand> mathiaz: ^^
[18:29] <jdstrand> mathiaz: that seems sane
[18:29] <jdstrand> mathiaz: if the profile exists in this scenario, then the user is prompted
[18:29] <cjwatson> pitti: ok, thanks, I've asked Henrik to see if he can reproduce this under hal-lock
[18:29] <mathiaz> jdstrand: hmm.. Considering that apparmor-profiles wasn't installed by default in gutsy, upgrading to hardy would prompt for a profile change for named
[18:29] <pitti> cjwatson: sudo hal-lock --interface org.freedesktop.Hal.Device.Storage --exclusive --run gedit
[18:29] <pitti> cjwatson: that works for me
[18:29] <pitti> cjwatson: without sudo it doesn't
[18:29] <pitti> (quite understandably)
[18:30] <jdstrand> mathiaz: if the profile existed, then yes.  and that is good
[18:30] <mathiaz> jdstrand: we could assume that the user knows a bit about apparmor since he installed apparmor-profiles
[18:31] <cjwatson> hal-lock's argument passing conventions are irritating
[18:31] <cjwatson> you have to pass command and arguments as a single argument
[18:31] <cjwatson> unlike sensible adverbial commands
[18:32] <jdstrand> mathiaz: I can say for the non-installed apparmor-profiles but the profile exists case, this is exactly what we want
[18:32] <jdstrand> mathiaz: prompting on a profile change
[18:32] <mathiaz> jdstrand: correct
[18:32] <mathiaz> jdstrand: however, if the system as a default profile upgrade should be transparent
[18:34] <jdstrand> mathiaz: really what we are worried about is the case where:
[18:34] <jdstrand> 1. apparmor-profiles is installed
[18:34] <jdstrand> 2. the profile migrated from apparmor-profiles to the package
[18:35] <jdstrand> mathiaz: so we need to decide if in this case, we want to enforcing mode
[18:35] <jdstrand> s/to//
[18:36] <mathiaz> jdstrand: well - if the profile is the default, we should enforce the new one. If the existing profile has been modified, we shouldn't enforce it
[18:36] <mathiaz> jdstrand: and leave it to the sysadmin to decide
[18:37] <jdstrand> mathiaz: I don't think we can do that.  the default apparmor-profiles package is in complain mode.  therefore the user could have done anything with her configuration, and it was working.  we flip to enforce and boom
[18:37] <mathiaz> jdstrand: good point.
[18:38] <mathiaz> jdstrand: so that means if apparmor-profiles was installed, we should put the profile in complain mode always
[18:39] <jdstrand> mathiaz: except that gets us back full circle to when the user has apparmor-profiles installed, *and* enabled enforcing mode
[18:39] <jdstrand> mathiaz: we just turned it off
[18:40] <jdstrand> mathiaz: but, the user would have been prompted
[18:40] <jdstrand> for the config change
[18:40] <jdstrand> conffile change
[18:40] <keescook> mvo: hi! which piece of software is responsible for the "Preconfiguring packages ..." text?
[18:40] <jdstrand> mathiaz: that seems sane
[18:41] <jdstrand> mathiaz: it doesn't break anything, and the admin was alerted to a change in the profile-- up to her to check it out
[18:41] <mathiaz> jdstrand: hmm. So we shouldn't change anything in that case.
[18:41] <mathiaz> jdstrand: no enforce or no complain mode
[18:41] <mathiaz> jdstrand: just leave it as it is.
[18:42] <mvo> keescook: debconf usually - why?
[18:42] <jdstrand> mathiaz: oh, so if apparmor-profiles is installed, just let dpkg and the admin handle it
[18:42] <jdstrand> mathiaz: ?
[18:42] <mathiaz> jdstrand: yes -
[18:42] <jdstrand> mathiaz: let me think about that
[18:42] <mathiaz> jdstrand: don't try to put the profile in complain mode.
[18:43] <JanC> anybody know if there is a reason why python-wxgtk2.6 in hardy now suddenly depends on python2.4 instead of python2.5 before, or should I file a bug?  ;)
[18:43] <mathiaz> jdstrand: the issue I see with your proposal about a complain/ directory is that there will be two places in the system where you can set a profile in complain mode
[18:43] <mathiaz> jdstrand: that will be... confusing
[18:43] <ScottK> JanC: Is it instead of 2.5 or in addition to 2.5?
[18:44] <jdstrand> mathiaz: yes, but the tools are modified to make it look consitent
[18:44] <jdstrand> mathiaz: eg aa-enforce and aa-complain will manage the symlinks
[18:44] <mathiaz> jdstrand: hm.. but if you edit a profile by hand, and set it to enforce, it can still be in complain mode
[18:45] <keescook> mvo: okay, was just looking at bug 203262 and saw the /tmp filename, wanted to double-check /tmp-usage-sanity
[18:45] <ubotu> Launchpad bug 203262 in hwtest "hwtest fails preconfigure" [Undecided,New] https://launchpad.net/bugs/203262
[18:45] <jdstrand> mathiaz: true enough, but our documentation says to use aa-enforce and aa-complain to switch between these modes
[18:45] <mathiaz> jdstrand: correct.
[18:45] <jdstrand> mathiaz: I had initially considered calling complain/ force-complain/
[18:46] <JanC> ScottK: it depended on "a python >= 2.4 and << 2.6" and "python2.5" in the past, and it depends on "a python >= 2.4 and << 2.6" and "python2.4" now
[18:46] <jdstrand> but thought that had the potential to be even more confusing
[18:46] <ScottK> JanC: Please file a bug then.
[18:46] <mathiaz> jdstrand: may be we should add a sanity check - if there is an inconsistency between the profile and the state in complain/ a warning message should be emitted
[18:47] <ScottK> JanC: Assign it to Adri2000 since he did the last upload.
[18:47] <JanC> k
[18:48] <jdstrand> mathiaz: eg 'Warning: found /etc/apparmor.d/complain/foo, using complain mode'
[18:48] <jdstrand> mathiaz: ad like you said-- *only* if inconsistent
[18:49] <mathiaz> jdstrand: yes
[18:49] <jdstrand> mathiaz: I like that
[18:49] <JanC> aah, someone beat me by 5 minutes to file this bug  ;)
[18:49] <jdstrand> keescook: have you uploaded the new apparmor yet?
[18:49] <jdstrand> keescook: with my patch?
[18:52] <qense> are there hal developers on? I think they might find bug 203094 interesting
[18:52] <ubotu> Launchpad bug 203094 in libsmbios "hal cannot set brighness on Dell notebook computers if a BIOS password is set" [Undecided,New] https://launchpad.net/bugs/203094
[18:52] <qense> (btw, maybe there should be a function added to this channel that you can poke a group)
[18:53] <keescook> jdstrand: no, but was going to soon, should I wait for updates?
[18:53] <jdstrand> keescook: mathiaz suggested I add a sanity check warning message in the case that complain/foo exists but the profile is in enforcing mode
[18:54] <keescook> jdstrand: yeah, good idea
[18:54]  * jdstrand nods
[18:54] <jdstrand> keescook: so I'll do that and ping you
[18:54] <keescook> okay
[18:54] <LaserJock> qense: the problem would be that there isn't really such groups to poke
[18:55] <qense> maybe it would be good if people could subscribe to specific keywords
[18:55] <keescook> mvo: what creates the files in /tmp prior to debconf running?  while I know the number isn't a PID, the path still makes me nervous: /tmp/hwtest.config.155713: ...
[18:55] <JanC> so I assigned https://bugs.launchpad.net/ubuntu/+source/wxwidgets2.6/+bug/203266 to Adri2000  ツ
[18:55] <ubotu> Launchpad bug 203266 in wxwidgets2.6 "python bindings of wxwidgets2.6 depends on python-2.4 instead of 2.4 and later" [Undecided,New]
[18:56] <LaserJock> qense: again, that requires such people being available I think
[18:56] <qense> yes
[18:56] <mvo> keescook: I suspect tat this file created by debconf, but its just a educated guess currently
[18:56] <qense> but it could be client side
[18:56] <LaserJock> qense: how do you mean?
[18:56] <keescook> mvo: it seems to read files from STDIN
[18:57]  * mvo checks
[18:57] <qense> a plugin for the client
[18:57] <qense> wait, that's not the work of the channel maintainers
[18:57] <Amaranth> bryce: is it possible to make bulletproofx not wipe out the xorg log file?
[18:57] <mvo> keescook: have you checked /usr/sbin/dpkg-preconfigure ?
[18:57] <bryce> Amaranth: you can turn it off in /etc/gdm/gdm.conf via the failsafe* options
[18:57] <mario_limonciell> Amaranth, it just saves it to /var/log/Xorg.$DISPLAY.log.old
[18:57] <mvo> keescook: eh, hold on
[18:58] <mario_limonciell> doesn't it?
[18:58] <LaserJock> qense: what I'm saying is that most of the time we don't necessarily have set people working on specific packages, or it's just one person perhaps. Filing a bug is much more effective than trying to ping people in IRC
[18:58] <keescook> mvo: I haven't, I assume it's something in dpkg, and I'm nearly positive it's using /tmp safely, but the all-numeric suffix bothered me.  I would expect letters if it's using mkstemp()
[18:58] <qense> ok
[18:59] <keescook> jdstrand: I pushed your changes-so-far to the primary apparmor bzr tree, if you want to merge to that
[18:59]  * jdstrand nods
[18:59] <LaserJock> qense: so when you ask for hal people, well there aren't any specific hal people that I know of
[18:59] <Amaranth> mario_limonciell: yeah, someone is telling me on their system that gets wiped out too
[18:59] <Amaranth> *shrug*
[18:59] <LaserJock> qense: but somebody will take a look at a bug report
[18:59] <qense> aha
[19:00] <qense> can I confirm the bug or should the reporter give more information?
[19:00] <qense> that was what I actually wanted to ask
[19:00] <qense> I think the largest part of the information is given
[19:00] <qense> and the dev can decide which specific information (s)he needs
[19:00] <mathiaz> keescook: bug 141295 is still relevant ?
[19:00] <ubotu> Launchpad bug 141295 in linux-ubuntu-modules-2.6.22 "apparmor: kernel fix for missing audit type" [High,Incomplete] https://launchpad.net/bugs/141295
[19:01] <mathiaz> keescook: I've seen you've submitted a couple of kernel patches for apparmor
[19:01] <LaserJock> qense: yeah, if you can confirm it go ahead. Adding information is always encouraged
[19:01] <qense> not that way ;)
[19:01] <qense> I'm a member of BugSquad
[19:02] <qense> I wanted to set it to confirmed(status) so the dev can decide what is needed more
[19:02] <qense> since things like this can be very system specifix
[19:02] <qense> c
[19:02] <LaserJock> I don't think setting it as Confirmed will do much
[19:02] <qense> ok
[19:03] <LaserJock> if you can confirm it or get somebody else to do so then set it as Confirmed, otherwise I'd say leave it as is
[19:03] <psyke83> hi, I'm the author of this bug report, and I'm concerned that a fix or workaround will not make it into Hardy: https://bugs.launchpad.net/bugs/195929
[19:03] <ubotu> Launchpad bug 195929 in gtk2-engines-murrine "Cosmetic bug: rectangular white outline surrounding rounded buttons (dup-of: 180962)" [Low,Fix committed]
[19:03] <ubotu> Launchpad bug 180962 in xulrunner "new GTK widgets could look better" [Unknown,Confirmed]
[19:04] <psyke83> I have already posted a fix for buttons, but GtkEntry boxes are still not rendering properly... this issue impacts Firefox 3, and since it's going to be the default browser, IMHO we should have some eyes look over the code
[19:05] <mario_limonciell> Amaranth, that's a bit surprising.  I've always had that log for failing back to
[19:06] <qense> psyke83: post your patch again in the main report
[19:06] <qense> (the one your original report was marked as a duplice of)
[19:06] <qense> or maybe even post the fix at the mozilla bug tracker
[19:07] <qense> because that's where it finally has to be fixed
[19:07] <psyke83> qense, damn, I didn't see it had been marked as duplicate.. IMHO my bug has more useful information
[19:07] <qense> I think they did that because your report is a bit 'untidy' ;)
[19:08] <psyke83> qense, the fix is only for buttons, not GtkEntry/GtkCombo boxes, and the fix is actually in the GTK2 engine code
[19:08] <qense> well, if I were you I'd post the information again in the main report and/or at moz
[19:09] <keescook> mathiaz: oh, yeah, I couldn't find that bug
[19:10] <keescook> (I opened a new one, bug 202888 I think)
[19:10] <ubotu> Launchpad bug 202888 in linux "aa-logprof unable to process logs without "type"" [Undecided,Triaged] https://launchpad.net/bugs/202888
[19:11] <mathiaz> keescook: I'll mark 141295 as Won't fix
[19:12] <keescook> mathiaz: well, I intend to fix it -- I have the patch in my hardy kernel git tree.
[19:12] <mathiaz> keescook: right - but the bug is against linux-ubuntu-modules-2.6.22
[19:13] <mathiaz> keescook: I could also mark it as a duplicate of bug 202888
[19:13] <ubotu> Launchpad bug 202888 in linux "aa-logprof unable to process logs without "type"" [Undecided,Triaged] https://launchpad.net/bugs/202888
[19:13] <keescook> ah, right
[19:14] <mathiaz> keescook: now I wonder - this patch had to be applied otherwise the genprof doesn't work in gutsy
[19:20] <LaserJock> man UUIDs can be a pain
[19:25] <ion_> laserjock: Huh?
[19:25] <LaserJock> they seem to change a lot on me, especially swap
[19:26] <ion_> Well, that seems strange.
[19:26] <LaserJock> I was trying to figure out why my computer wouldn't hibernate, then a look around and I've got no swap
[19:28] <LaserJock> and sure enough the UUID had changed
[19:28] <ion_> I can't think of what could cause that.
[19:29] <LaserJock> well, this time it happened after I booted into a different OS
[19:30] <LaserJock> it's usually related to an install or another OS
[19:30] <LaserJock> but it's rather aggravating
[19:31] <keescook> LaserJock: swap changing UUID is a bug :(  bdmurray might know more about it
[19:31] <LaserJock> yeah, I think I know the bug
[19:32] <LaserJock> I've been having this problem since I think edgy or so
[19:32]  * keescook nods
[19:32] <keescook> I haven't been able to reproduce it, but I also don't multiboot
[19:32] <LaserJock> initially I made the mistake of using the same swap partition for multiple OSs
[19:32] <LaserJock> which makes sense that hibernation might cause problems
[19:33] <LaserJock> so now I have a swap partition for each OS
[19:33] <keescook> well, it seems like there is an OS that just mkswap's for fun.
[19:33] <LaserJock> which wastes a lot of space
[19:34] <cjwatson> keescook: it's apt-extracttemplates
[19:34] <keescook> cjwatson: yeah, mvo and I switched to privmsg, all is fine.  :)
[19:34] <cjwatson> keescook: the code looks scary on the face of it, but I think it is just about safe
[19:34] <cjwatson> ok, cool
[19:34] <LaserJock> but the thing for me is that it's just a problem in using UUIDs, if I use the actual name of the partition it's all fine
[19:34] <keescook> cjwatson: yup, I figured it must be, but the tmp name didn't feel like mkstemp so I wanted to see for myself.  :)
[19:34] <emgent> heya people
[19:35] <keescook> heya emgent
[19:35] <LaserJock> it seems that while UUID is handy it's also fragile
[19:35] <keescook> LaserJock: well, generally only in this weird case.  you can use other names too
[19:36] <LaserJock> yeah, I would hope that UUIDs of non-swap partitions would stay the same :-)
[19:39] <cjwatson> LaserJock: you're not the first person to report this for swap, by a long shot; I'd love to know what's causing it
[19:40] <LaserJock> cjwatson: this time it was a tad strange
[19:40] <LaserJock> what I did was install Fedora 8 on another partition
[19:40] <LaserJock> made sure it wasn't using the Ubuntu swap
[19:40] <LaserJock> but I swear I've rebooted a couple times into Ubuntu without a problem before today
[19:41] <LaserJock> anymore the first thing I check after doing an install is that swap still works
[19:42] <LaserJock> perhaps Fedora does a mkswap anyway
[19:42] <LaserJock> even if I tell it I don't want to use that partition for swap
[19:55] <keescook> cjwatson, bryce: what are your thoughts on bug 190370?  this seems like an improper "escape" from bullet-proof mode?
[19:55] <ubotu> Launchpad bug 190370 in xorg "privilege escalation when canceling the low-graphics warning prompt" [High,Confirmed] https://launchpad.net/bugs/190370
[19:56] <cjwatson> I commented on it already, I definitely think it's a bug and seems like a possible vulnerability
[19:56] <cjwatson> well, maybe not, since you have to have physical access
[19:56] <keescook> yeah, has me worried -- though I'm a bit on the fence due to the physical access part
[19:57] <bryce> keescook: is that actually bulletproof-x mode?
[19:57] <keescook> bryce: not sure, I haven't had time to reproduce it yet.
[19:57] <cjwatson> bryce: it's unclear, but it does seem vaguely like the failsafe mode
[19:57] <cjwatson> they don't quote exact text, but displayconfig-gtk does mutter about low graphics mode
[19:58] <cjwatson> I suspect it's more likely to be a gdm bug than anything else
[19:58] <cjwatson> or at least that that's a better place to start than xorg
[19:58] <cjwatson> but didn't want to reassign it, not being Bryce or Timo
[19:59] <bryce> ted said he reproduced it; perhaps he can give more details
[20:05] <ogra_cmpc> soren, could i get a look form a virtualization expert on bug #134865 ?
[20:05] <ubotu> Launchpad bug 134865 in ltsp "x11-common segfaults in ltsp chroot in xen dom0" [Undecided,Fix committed] https://launchpad.net/bugs/134865
[20:05] <ogra_cmpc> does that fix sound sane to you ?
[20:10] <mdke> pitti: we download the translations from translations.lp.net/ubuntu/hardy/+source/ubuntu-docs - you can see the relevant template for each document
[20:13] <soren> ogra_cmpc: The fix being that it installs libc6-xen?
[20:13] <soren> ogra_cmpc: Yes, that sounds sane.
[20:14] <ogra_cmpc> it installs libc6-xen if thats installed on the server
[20:14] <ogra_cmpc> ok, fine then, i'll commit it after he fixed it up a bit
[20:17] <mdke> pitti: reference is at https://wiki.ubuntu.com/DocumentationTeam/Translation which is always worth a read if you haven't translated docbook xml before
[20:42] <murlidhar> http://sourceforge.jp/forum/forum.php?thread_id=7399&forum_id=5768
[20:42] <murlidhar> can anybody port cabos application to linux
[20:42] <murlidhar> please
[20:46] <wasabi> I suspect you'll have more luck doing it yourself.
[20:47] <cristi> hello
[20:47] <cristi> is there any developer here
[20:47] <cristi> with whom I could discuss some openal issues?
[20:49] <cristi> more exactly https://bugs.launchpad.net/ubuntu/+source/openal/+bug/194919
[20:49] <ubotu> Launchpad bug 194919 in openal "libopenal needs replacement" [Undecided,New]
[20:52] <cristi> have I perhaps asked the wrong room?
[20:54] <_MMA_> People are most likely busy of not working atm.
[20:54] <cristi> thanks
[20:54] <cristi> but is this the right room?
[20:58] <keescook> asac: I've filed bug 203306 -- you'd asked me to do so a while back but it had fallen down my TODO list until now
[20:58] <ubotu> Launchpad bug 203306 in firefox "when upgrading to FF3, some about:configs are reset" [Undecided,New] https://launchpad.net/bugs/203306
[21:08] <asac> keescook: those are booleans, right?
[21:10] <keescook> asac: yeah
[21:11] <LaserJock> oops, I seem to have killed the hwtest app :/
[21:15] <asac> keescook: ok i attached a patch that still applies cleanly :). don't ask me if it still works though.
[21:15] <keescook> hehe
[21:16] <\sh> slangasek: please approve php-xdebug (universe) thx :)
[21:17] <\sh> s/slangasek/whoever has the mighty powers to approve stuff/
[21:38] <warp10> Hi all!
[22:14] <alex-weej> bryce: hey. any idea what's going on with xorgs tendency to detect screen dimensions incorrectly lately?
[22:16] <bryce> alex-weej: not sure what you mean.  Hardy does rely more on the automatic screen detection rather than the debconf heuristics, so there is possibilities that if your screen had been set up correctly before, it may work differently now, but that change went through a few months ago
[22:16] <alex-weej> maybe i just hadn't noticed it as gnome has my DPI set in gconf at 96
[22:16] <alex-weej> have you seen my bug?
[22:17] <seb128> slangasek: bug #202193 is somewhat noticable and you might to milestone it
[22:17] <ubotu> Launchpad bug 202193 in ubuntu-gdm-themes ""incorrect user name" errormessage displays in a wrong place" [High,Confirmed] https://launchpad.net/bugs/202193
[22:18] <bryce> alex-weej: perhaps one of the most common issues people have run into is if they have a tv-out; on some cards X ends up using that output's resolution (typically 1280x1024), instead of what's expected.  In these cases, we generally can just quirk your card to ignore the tv-out
[22:18] <bryce> alex-weej: I've likely seen your bug, which # is it?
[22:18] <alex-weej> i'm trying to find it, 2 secs
[22:19] <alex-weej> https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/201491
[22:19] <ubotu> Launchpad bug 201491 in xorg "Screen DPI detected incorrectly" [Undecided,New]
[22:19] <alex-weej> basically xdpyinfo gets the physical dims wrong
[22:20] <alex-weej> but ddcprobe gets it right
[22:20] <alex-weej> where is xdpyinfo getting its numbers from?
[22:20] <alex-weej> also there is this sort of related issue with Firefox getting its DPI settings directly from X without checking GNOME's configuration
[22:20] <alex-weej> https://bugs.launchpad.net/ubuntu/+source/firefox-3.0/+bug/201487
[22:20] <ubotu> Launchpad bug 201487 in firefox-3.0 "Firefox uses the wrong DPI" [Undecided,Confirmed]
[22:20] <alex-weej> which some people have also added more information on
[22:20] <alex-weej> but they are two separate problems
[22:36] <slangasek> bryce: if displayconfig-gtk is no longer to be included in the menus, should it also be unseeded?
[22:40] <cjwatson> keescook: I have a security vulnerability in gfxboot for you
[22:44] <keescook> cjwatson: seriously?
[22:44] <cjwatson> keescook: </troll> elmo reckoned you might help if I claimed that ;-)
[22:44] <keescook> hah
[22:44] <cjwatson> (super-weird memory corruption in the middle of assembly code, though :-/ )
[22:45] <keescook> yerg.  bug #?  or some pointers?
[22:45] <cjwatson> ah, I'm on it, I've just been on it for two hours and needed light relief
[22:45] <keescook> heheh
[22:45] <cjwatson> it's breaking the Kubuntu and Xubuntu boot processes at the moment
[22:45] <cjwatson> as in, completely; CD boot menu hangs
[22:46] <keescook> whoa.  but only [kx]ubuntu?
[22:46] <keescook> what's different with them beyond theme?
[22:46] <slangasek> the number of letters in the name, for one :)
[22:46] <cjwatson> right
[22:47] <keescook> I get enough of this literalness from soren.  ;)
[22:47] <cjwatson> no, he's serious
[22:47] <cjwatson> I've isolated it to that
[22:47] <keescook> *lol*
[22:47] <keescook> but mythbuntu doesn't hang?
[22:48] <cjwatson> haven't tried Mythbuntu, but Ubuntu and Edubuntu are just fine
[22:48] <cjwatson> (I know Edubuntu doesn't have a bootable CD, but I was just changing the labels in isolinux.cfg on the fly)
[22:48] <slangasek> seb128: 202193 milestoned for release, thanks. (anyone working on it yet?)
[22:48] <_MMA_> cjwatson: What days image. Ill try Ubuntu Studio.
[22:48] <_MMA_> *image?
[22:48] <nxvl> where is that i request for givin-backs
[22:48] <cjwatson> _MMA_: today's, but I don't need more tests - it's at a hideously low level
[22:49] <cjwatson> and, as I say, I was doing this just by editing isolinux.cfg
[22:49] <cjwatson> not using whole different images
[22:49] <_MMA_> cjwatson: Ok
[22:49] <keescook> cjwatson: at what point does it freeze?  is it arch-specific?
[22:49] <seb128> slangasek: (no idea, that's an artwork issue)
[22:50] <cjwatson> keescook: it freezes in the middle of a 'savescreen' op in the xmenu.init function in gfxboot-theme-ubuntu/xmenu.inc
[22:51] <cjwatson> savescreen is defined in gfxboot/bincode.asm
[22:52] <bryce> slangasek: no, it's still needed for the bulletproof-x mode (no getting away from that via xrandr - this use case requires patching up the xorg.conf, and only dcg does that now)
[22:52] <cjwatson> I would love advice, but on the other hand I'm aware that anyone else will need a while just to get up to speed on this code
[22:52] <slangasek> bryce: ok
[22:52] <keescook> I can read asm quickly.  gfxspamasm I'll need to grok a bit longer.  ;)
[22:53] <slangasek> bryce: btw, I had occasion to be grumpy with the new, not-recorded-anywhere-in-xorg xrandr handling when my first attempt to play with the new GUI crashed X, and then afterwards my display was coming up sideways after login... :)
[22:53] <keescook> cjwatson: okay, reading savescreen, one sec
[22:54] <cjwatson> alloc_fb ends up in a malloc implementation
[22:55] <keescook> I assume "prim_" prefix just means its callable?
[22:56] <cjwatson> corresponds to ops in the interpreted language, pretty much
[22:56] <cjwatson> primitive
[22:56] <keescook> what's your methodology for debugging gfxboot?
[22:56] <bryce> slangasek: yeah
[22:56] <cjwatson> prayer
[22:56] <keescook> hahah
[22:56] <keescook> so, no gdb under it, etc?
[22:56] <cjwatson> you can stick 'dtrace' ops wherever you like, and then you get a debug console
[22:56] <cjwatson> shows you the stacks, and you can single-step
[22:56] <slangasek> bryce: but I guess that's just an instance of the one bug you're already tracking, isn't it ;)
[22:56] <cjwatson> you can't do any more serious inspection, but it's a good start
[22:57] <cjwatson> you can only debug at the interpreted language layer, though
[22:57] <keescook> can I run it on a local console sanely?
[22:57] <bryce> slangasek: bug 197673 (second in importance only to bug 197645 which I'm finalizing a (hopeful) fix for)
[22:57] <cjwatson> kvm
[22:57] <ubotu> Launchpad bug 197673 in gnome-control-center "gnome-display-properties should revert change automatically if not acknowledged" [High,Confirmed] https://launchpad.net/bugs/197673
[22:57] <ubotu> Launchpad bug 197645 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in rw_screen_list_outputs()" [Medium,New] https://launchpad.net/bugs/197645
[22:57] <cjwatson> you can't run it under Linux - it's bootloader hooks
[22:57] <cjwatson> s/Linux/Linux directly/
[22:58] <keescook> okay, yeah, it's doing direct BIOS stuff, I assume.
[22:58] <bryce> slangasek: I share your xrandr gui grumpiness
[22:58] <cjwatson> keescook: I have a vague suspicion here that snprintf might be writing off the end of a memory region or something
[22:59] <bryce> slangasek: but these are all top priorities for me (unfortunately I was sick most of last week - was hoping to have more of them resolved by now)
[23:02] <keescook> cjwatson: you're suspecting that an sprintf is corrupting soemthing that prim_savescreen, save_bg, or alloc_fb use?
[23:03] <keescook> cjwatson: are there any async things that happen in gfxboot?  any timers, interupt handlers, etc that might be firing?
[23:05] <cjwatson> there are, but none of those seem relevant; I've reproduced this entirely consistently
[23:05] <cjwatson> I'm suspecting that snprintf is corrupting the malloc arena in pretty much the usual way, albeit transferred to this environment
[23:07] <keescook> right... so, if the diff between ubuntu and kubuntu is litterally 1 letter, can you reproduce by modifying the otherwise-working ubuntu theme and making it fubuntu or something?
[23:07]  * keescook reads prim_snprintf
[23:07] <cjwatson> keescook: I considered Kubuntu and Xubuntu good enough proof
[23:07] <cjwatson> I might just be *using* snprintf wrongly, conceivably
[23:08] <keescook> the themes are just read from the .cfg file?
[23:08] <cjwatson> the terminology is misleading
[23:08] <cjwatson> gfxboot-theme-ubuntu is actually the implementation for all flavours
[23:08] <keescook> I guess I want to ask, "How do I reproduce this?"  :)
[23:09] <cjwatson> should be able to do it by taking an Ubuntu CD, editing isolinux/isolinux.cfg, and making the menu label entries that mention Ubuntu say Kubuntu instead
[23:09] <cjwatson> changing it to "Fubuntu" has the same effect
[23:10] <keescook> more stupid questions: and this does not happen with the 3.3.28-0ubuntu2 (gutsy) version?
[23:10] <LaserJock> can I get an archive admin to give back mysql-gui-tools please?
[23:11] <seb128> LaserJock: no, you need a buildd admin there
[23:11] <LaserJock> ah
[23:12] <cjwatson> keescook: string lengths were different in the configuration associated with that
[23:12] <cjwatson> I haven't tried otherwise
[23:12] <keescook> okay
[23:12] <LaserJock> I thought archive admins could do give-backs
[23:12] <cjwatson> it's worth noting that we hardly ever call malloc directly - it's almost always through a 'string' wrapper that deals with adding 1 to the length
[23:13] <keescook> cjwatson: yeah, not reading through the prim_length and other things in bincode
[23:13] <keescook> er s/not/now/
[23:14] <keescook> i'm uncomfortable with the memory "type"ing being done...
[23:15] <cjwatson> seems to work pretty well in general
[23:15] <keescook> yeah, I'm sure it's okay -- it's just hurting my poor eyes
[23:16] <keescook> (this is the first time I've actually tried to seriously read the gfxboot code)
[23:17] <cjwatson> I rarely have to try to read the assembly, fortunately
[23:17] <cjwatson> ok, it's specifically the "^Install Kubuntu" string that it hates
[23:18] <keescook> you'll have this fixed by the time I'm understanding the primitives ;)
[23:19] <cjwatson> don't hold your breath, it's been three hours now
[23:19] <cjwatson> oh, if you want documentation rather than grokking it from first principles, run make in doc/
[23:27] <Amaranth> so gfxboot hates kubuntu? good to know ;)
[23:27] <seb128> slangasek: any reason you didn't accept the gnome-power-manager update?
[23:27] <keescook> cjwatson: any docs on the memory management design?  it seems to make distinctions between memory locations for doing memory sizing
[23:27] <keescook> cjwatson: I'm trying to convince myself that the "sprintf" definition is safe.  :)
[23:28] <cjwatson> I'm wrong, it's the "^Try Kubuntu without any change to your computer" string
[23:28] <keescook> the  dup cvp length  seems right, unless "length" malfunctions
[23:28] <cjwatson> keescook: the stuff in doc/ is all there is
[23:28] <keescook> cjwatson: I suspected.  ;)
[23:28] <cjwatson> oh, and the comments in bincode.asm ...
[23:29] <slangasek> seb128: because I did?
[23:29] <seb128> slangasek: hum, just received the -changes mail, you can ignore the question
[23:29] <slangasek> :)
[23:29] <keescook> cjwatson: yeah, that's what I've been reading.  seems like a fun tool -- just insanely difficult to debug.
[23:30] <cjwatson> I have to say that it's normally been slightly easier than this
[23:30] <cjwatson> provided you can wrap your head round an RPN language, but gfxboot is hardly the only such
[23:30] <Kopfgeldjaeger> n8
[23:31]  * slangasek whimpers in pain at the amount of time it takes to get a bzr co of ~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy
[23:31] <keescook> cjwatson: yeah, I'm cool with the RPN -- it's just reading what each prim is doing that's taking me time.
[23:32] <keescook> I'm now trying to convince myself that the sprintf definition doesn't need a "1 sub" to the length before calling snprintf
[23:32] <LaserJock> slangasek: hehe,  imagine how long it took to push it
[23:33] <LaserJock> slangasek: I think it was something like 5 days
[23:33] <seb128> slangasek: how restrictive are you on changes right now? I've a patch to make libgnomeui uses gio for thumbnails which is in fedora for some time and has been commited upstream now, it fixes bug #203151
[23:33] <ubotu> Launchpad bug 203151 in libgnomeui "Image thumbnails are not shown in Trash" [Low,Fix committed] https://launchpad.net/bugs/203151
[23:34] <seb128> slangasek: the change would be nice to have but there is no hurry, I'm fine delaying to after beta
[23:36] <slangasek> seb128: I think I'd prefer that one to wait until after beta
[23:36] <slangasek> LaserJock: what format is this branch in?
[23:36] <slangasek> maybe it needs some bazaar guru lovin'?
[23:37] <LaserJock> slangasek: don't know actually
[23:37] <LaserJock> slangasek: well, I believe lifeless helped get it all set up
[23:37] <keescook> cjwatson: *hah* prim_snprintf aims the gfx buffer at the alloc'd region and calls "printf"! hahaha. I love reading asm.
[23:37] <cjwatson> keescook: I think I've convinced myself of that after staring at it: in get_length_40, LOOPNZ decrements ECX before the test, so an empty string will end up with ECX == -1, and NOT ECX => 0
[23:37] <slangasek> LaserJock: hrm, ok.  OTOH, how long ago did he do that, and how much has the state-of-the-art moved since then :)
[23:37] <LaserJock> slangasek: you might want a lightweight check out
[23:38] <seb128> slangasek: ok
[23:38] <keescook> cjwatson: ewww
[23:38] <LaserJock> slangasek: very true, it has been a while
[23:38] <cjwatson> NOT -2 => 1, etc.
[23:38] <cjwatson> ones-complement negation FTW
[23:39] <slangasek> LaserJock: right, for what I'm doing with it that's probably a good idea
[23:40] <keescook> cjwatson: should be easy to test?  have it print out the length of various strings?
[23:41] <keescook> cjwatson: these func_NN names keep making me think I'm reading BASIC.  ;)
[23:48] <superm1> asac, suspend/resume was finicky before, but i'll try it
[23:57] <cjwatson> keescook: ok, this is seriously freaky. This seems to work around it (with other problems, but still): http://paste.ubuntu.com/5804/
[23:58] <cjwatson> so copying the characters one-by-one breaks, but strcpy succeeds?