toresbeI have a quick question about a bug I just submitted which I'm not sure got submitted right.00:52
toresbehttps://bugs.launchpad.net/ubuntu/+source/linux/+bug/194196 - under "Affects"00:53
ubotuLaunchpad bug 194196 in linux "Fails to insert nVidia SATA disk modules on boot causing boot fail" [Undecided,New]00:53
toresbedoes that mean the bug is being forwarded to the Kernel.org dudes? 'cause I think this is a kernel packaging issue00:53
LaserJocktoresbe: I think the Ubuntu Kernel team will be looking at it00:55
toresbeLaserJock: great, that's my "target audience". :) thanks00:55
* toresbe sticks around in case someone says LaserJock is wrong00:55
LaserJockyou might also ask #ubuntu-kernel00:56
toresbeah, right, thanks.00:56
=== milli_ is now known as milli
toresbeLaserJock: solved, thank you00:59
toresbe(my confusion, not the bug :))00:59
toresbeI'm seeing some very strange behavior in the tracker system, but I'm not sure where to report it.01:11
toresbeWell, launchpad, yes, but I don't know why this bug is occurring or anything. it's a rather strange bug.01:12
blueyedtoresbe: post a bug in launchpad itself: https://help.launchpad.net/Feedback01:13
toresbesorry, not *that* tracker system, but "tracker" the deb package :)01:14
toresbeI'd like to ask a tracker upstream developer about it... y'know where they might be found?01:14
blueyedtoresbe: no. I would look at the homepage/google or just try #tracker01:15
toresbe#tracker's empty01:15
toresbeand "tracker" is quite difficult to google!01:16
blueyedvia https://launchpad.net/tracker => Homepage01:16
blueyedIRC: GimpNet, channel #tracker01:16
toresbeaha! Thanks. :)01:17
blueyedoften the mailing list might be better though :)01:17
blueyedgood luck and good night.01:17
toresbeHehe. I really like this bug.01:38
toresbeSo - here's what's happening. tracker is searching along, and encounters a graphics file. It wants to thumbnail it, so it passes it along to convert. Convert creates a temporary file. Tracker notices that a file has been created. It notices it's an image file, so it wants to thumbnail it. It calls convert.01:38
=== jelmer_ is now known as jelmer
* Hobbsee wonders why stuff from firefox is only playing out of one speaker.02:16
lifelessbecause fox is singular ;)02:18
=== superm1 is now known as tgm4883_
=== tgm4883_ is now known as superm1
=== calc_ is now known as calc
=== XSource_ is now known as XSource
pittiGood morning07:20
pittiKeybuk: I am now07:22
=== hunger_t is now known as hunger
=== calc__ is now known as calc
z5000manI've got an issue with gutsy installation, can someone lend a hand, as i have been in like 10 different channels where noone talks07:52
=== doko_ is now known as doko
warp10Good morning!08:12
pittihey warp1008:15
brycez5000man: sorry to hear of the troubles; this probably isn't the right channel either, but maybe try asking via answers.launchpad.net/ubuntu08:24
evandAgostino Russo and I are trying to determine whether a ntfs-3g panic occurs outside of virtual iron.  If someone would be so kind as to try installing Ubuntu 64-bit in Windows using Wubi from the latest daily live CDs, I would very much appreciate it.08:31
evandThe panic occurs post-install, on the first boot.08:31
dholbachgood morning08:42
mdkemorning dholbach08:48
dholbachheya mdke08:48
Mithrandiris it just for me that sudo -u $USER -i (i386, hardy) segfaults?08:54
Mithrandirhm, ok.08:56
toresbetoresbe@mjollnir:~$ sudo -u postgres -i08:56
toresbe[sudo] password for toresbe:08:56
Mithrandirtoresbe: that's something else than what I asked about08:57
MithrandirI wonder if it's related to me adding myself to a group just before.08:57
toresbeoh, $USER as in, the actual shell variable $USER, and not "whatever user". :)08:57
cjwatsonoh, I misunderstood that too08:58
toresbethat works, too.08:58
Mithrandirit's related to pam_mount, it seems08:59
=== \sh_away is now known as \sh
evandRiddell: Kubuntu Wubi works...sort of.  I need to fix the calling of kwin/dcop as they currently crash and leave you with just an X cursor, but the install goes through just fine.09:05
evandYou just have to sit and stare at a mostly black screen and wait for it to reboot09:05
Riddellevand: exciting.  so I should be able to test it on my girlfriend's dual boot laptop without it wiping her windows partition?09:10
evandRiddell: indeed09:11
evandRiddell: Just pop in one of today's daily live CDs, then select "Install inside Windows"09:12
evandit'll create a C:\Ubuntu folder and add Kubuntu to the NT bootloader, both of which can be undone by running the Wubi uninstaller in Add / Remove programs09:13
Riddellevand: does today's daily CD include 20080221.1?09:20
RiddellI suppose it does in your timezone09:20
Riddellmade at 21:4109:20
=== pbn_ is now known as pbn
evandheh, indeed it does09:21
evandreally any Kubuntu CD after the new livefs landed (I think that was day before yesterday, no?) will work09:21
davmor2out of curiosity will the language selector at the beginning of the cd look better than it does at the moment?09:25
evanddavmor2: can you elaborate?09:26
Riddellhe means the gfxboot menu that now pops up by default09:28
Riddellit's useful to non-english speakers but looks very out of place to those of us who don't need it09:28
davmor2evand: currently at the boot up of any *buntu cd you get the language selector over the top of the menu.  It's grey and blocky and looks cheap and tacky09:28
evandcjwatson: ^09:28
RiddellI suspect it would take a lot of gfxboot hacking to improve it notably09:29
evandSuSE seems to have done it with theirs, which I quite like, but I agree.  gfxboot is scary, and it would probably take quite a bit of work.09:29
evandor maybe that's a false memory, I'd have to check again09:30
evandRiddell: is there going to be a separate Kubuntu Alpha 5 announcement?09:30
davmor2Riddell: not even if you don't need it everything else on the cd looks nicely integrated and smooth but this looks like an after thought at the moment.  I can see that it is useful just ugly.09:30
Riddellevand: not if it's released at the same time09:31
evandRiddell: ah, well in case you end up needing them, I took some screenshots of the Kubuntu umenu and Wubi: http://evalicious.com/kubuntu/09:34
* evand -> bed09:35
Riddellevand: excellent, thanks09:36
cjwatsondavmor2: yes, it is kind of an afterthought. We'll try to improve the UI. In the meantime please accept that this is known (I think that this is the second time you've mentioned it)09:39
davmor2cjwatson: no first the other language thing is latter on in ubiquity09:40
pitti\o/ NEW queue == 009:45
sorenpitti: really?!? Wow.09:47
* soren hugs pitti09:47
* dholbach hugs pitti too09:48
pittiasac: can you please modify the firefox seeds in -desktop for firefox-3.0?10:01
asacpitti: why? all is fine imo10:02
=== pbn_ is now known as Pbn
=== Pbn is now known as _pbn_
=== _pbn_ is now known as pbn
Q-FUNKdid the PPA builders go belly up?10:12
Q-FUNKthey fail a build becuase they try to include a non-existant sources.list item10:13
cjwatsonQ-FUNK: URL to build log?10:13
FujitsuQ-FUNK: That's a #launchpad problem. Did you just activate your PPA now?10:14
cjwatsonQ-FUNK: odd; presumably that *would* exist if it managed to build and publish anything10:14
cjwatsonbut yes, as Fujitsu says, better in #launchpad10:14
FujitsuI can see why that would happen as of 1.2.2, though.10:14
Q-FUNKFujitsu: no, I've been using it for a while already10:14
FujitsuQ-FUNK: It was empty, though?10:14
Q-FUNKalright then10:14
Q-FUNKit wasn't empty10:15
cjwatsonerr, ok, that's very weird, that URL *does* exist10:15
Q-FUNKmy gutsy PPA was laready in use10:15
Q-FUNKhardy was just activated with a 1st package today10:15
FujitsuAh, that's it, then.10:15
saispoany hibernate expert in the room ?10:15
FujitsuBefore 1.2.2, builds didn't start for up to 45 minutes after the source was uploaded.10:15
Q-FUNKbut the gutsy package failed the same way10:15
FujitsuAs sources are published every 20 minutes, they were already published before the builder tried to update its sources.10:16
FujitsuNow builds start immediately.10:16
Q-FUNKthe builder produces different output than pbuilder does10:16
FujitsuQ-FUNK: Those builds failed normally.10:17
FujitsuOnly hardy has that problem.10:17
FujitsuGutsy fails for legitimate reasons.10:17
Q-FUNKI'm not so sure.10:17
Q-FUNKher,e build works fine using pbuilder10:17
FujitsuIt's entirely unrelated, at any rate.10:17
FujitsuIt's a problem in the package, but it might not be picked up in pbuilder.10:18
Q-FUNKcould be10:19
Q-FUNKis the PPA using sbuild?10:19
FujitsuA horribly mangled sbuild, but yes.10:19
Q-FUNKhorribly mangled might be the key.10:20
FujitsuI don't find it likely that it could produce that kind of error.10:20
Q-FUNKit builds fine on debian10:21
FujitsuDebian's archive isn't identical.10:21
FujitsuA lot of things fail to build here.10:22
Q-FUNKbuilders should be10:22
FujitsuEr, why?10:22
FujitsuWe have completely independent archive and builder infrastructure.10:22
Q-FUNKthey both are debian packages.  there's no reason for a complete fial on one distro and not on the other.10:22
FujitsuTell that to all of our FTBFS bugs.10:22
cjwatsonthere are plenty of reasons why that sort of thing might happen10:23
Q-FUNKhmm. ok10:23
Q-FUNKanyhow. thanks for the explanation10:23
cjwatsonin this case your CFLAGS seems not to be taking10:23
pittiasac: hm, but package 'firefox' is transitional, isn't it?10:43
asacpitti: nope ... meta10:44
pittioh, argh, that10:44
pitti   firefox | 3.0~b3+nobinonly-0ubuntu2 | hardy/universe | all10:44
pittithen we need to promote it10:45
asacwhy is it in universe?10:45
pittibecause I demoted source+binaries of firefox, and apparently soyuz got confused10:45
asacah ok10:45
asacpitti: promote all firefox binaries: firefox + firefox-gnome-support + firefox-libthai (that will appear again in next upload) and firefox-dom-inspector10:46
asacpitti: well firefox-libthai is the only transitional package10:46
asacbecause the feature is now in core-xulrunner10:46
davmor2mjg59: ping11:33
mjg59davmor2: Hi11:37
davmor2mjg59: Riddell just asked me to have a word about an unusual effect in Kubuntu.  This display fades to black once you've logged in.11:38
mjg59Hm. Sounds like it may not be my problem :)11:39
davmor2Riddell: is thinking it might be a hal issue don't ask me why.11:40
mjg59Can't see how it would be a hal issue11:40
mjg59But it'd be nice to have some more background11:40
Riddellwell I've no idea, but it doesn't seem to be kmilo or guidance which are the parts that touch brightness in kde11:40
mjg59hal will never alter anything without being told to11:40
ion_And sometimes when told to, he answers “I’m afraid I can’t do that, Dave”.11:41
davmor2Riddell: don't forget that guidance doesn't seem to be running.11:41
davmor2slangasek: cjwatson:  Ubiquity issue is sorted now run straight through to 100% :)12:00
cjwatsongood stuff12:01
Keybukjdstrand: around?12:27
jdstrandKeybuk: I'm a while for a while12:44
jdstrandwow, I just woke up12:44
jdstrandI'm around for a while12:45
Keybukjdstrand: pitti and I were discussing PAM12:45
Keybukand he mentioned you'd been working on a way to automatically change /etc/pam,d/common-auth for libpam-ldap12:45
jdstrandKeybuk: it's called auth-client-config.  it is profile based, and can switch pam and nsswitch.conf12:46
jdstrandit's in main12:46
pittiright, that was it12:46
dredgit works pretty well, especially in conjunction with things like puppet12:47
jdstrandbasically if an admin knows what pam/nss settings he/she wants, then the admin can put the profile into the auth-client-config datase, run a command and *poof* the machine is updated12:47
jdstrandit doesn't do pam/nss configuration (and stacking) however-- the admin has to know what he/she wants12:47
Keybukso it can't be used in postinst?12:49
jdstrandKeybuk: yes12:50
Keybukyes it can't or yes it can? :p12:50
jdstrandKeybuk: it was designed to have the necessary flags in place to work safely in this regard12:50
jdstrandoh heh-- still sleppy12:50
jdstrandKeybuk: AKAIK, no package is using it for that12:51
jdstrandKeybuk: but /usr/share/doc/auth-client-config/README hasa blurb on it12:52
jdstrandKeybuk: it is largely untested in this regard, and ideally we would need to develop a policy to ensure safety12:52
jdstrandKeybuk: but I believe everything is there to do it properly12:52
jdstrandsleepy even12:53
Keybukright, what would that policy look like?12:53
jdstrandKeybuk: postrm is easy: auth-client-config -p profile_name -a -r-- this will revert pam and nss for profile_name only if the current system settings match profile_name12:56
Keybukwhat about install?12:56
jdstrandit reverts pam/nss to pre-auth-client-config settings12:56
=== cjwatson_ is now known as cjwatson
asacdholbach: can i participate in five-a-day by just reporting to the log?12:57
dholbachasac: sure12:57
jdstrandKeybuk: thinking...12:58
asacdholbach: what bugs do qualify for that list? every bug touched? or only fix release/invalid/wontfix?12:58
dholbachasac: I think that if you contribution helps to get it fixed, that qualifies it :)12:59
jdstrandKeybuk: I think just a debconf question asking whether or not to activate the profile would be good enough, then use 'auth-client-config -a -p profile_name'13:01
jdstrandKeybuk: well, that may not be good enough13:01
dredgi think you're looking for something like redhat's authconfig :)13:02
jdstrandKeybuk: no, if it was a shared debconf key, I think that might do it13:02
cjwatsondredg: IME Red Hat tools are just about as difficult to port to Ubuntu/Debian as our tools would be in the opposite direction13:04
dredgcjwatson: very much so13:04
cjwatsondistributions tend to build up a lot of infrastructure on which this sort of thing is built, and we're pretty divergent in those terms13:04
jdstranddredg: redhat doesn't have a packaging policy like we do, so they can take shortcuts13:05
dredgwhile you can preseed some things, like libnss-ldap, you can't easily preconfigure pam. if you could preseed auth-client-config with a profile (or profile location it could slurp) and tell it to apply a profile13:05
Keybukjdstrand: a-c-c has no knowledge of the default config, right?13:06
jdstrandKeybuk: no. but it can reset the system to pre-a-c-c values13:07
jdstrandKeybuk: when you apply a profile, it will comment out the old settiings.13:07
Keybukdoes anything use a-c-c at the moment?13:09
jdstrandKeybuk: I had initially thought that pam and base-files might want to use auth-client config, along with ldap, kerberos, etc13:09
jdstrandbut currently no packages use it that I am aware of13:09
jdstrandKeybuk: slangasek wasn't too keen on base-files and pam using it13:10
jdstrandKeybuk: he has a valid argument that pam stacking configuration should be handled by the packages13:10
jdstrandKeybuk: the problem is that is incredibly difficult once you get past a couple of pam mosules13:11
jdstrandKeybuk: and in so doing, packaging makes assumptions about what the admin wants13:11
dredgplus pam config is very site specific13:11
jdstrandKeybuk: this conveniently bypasses all that, as long as the admin knows what she wants13:11
=== Vegar_ is now known as Vegar
jdstrandnot to mention-- this pam stacking configuration code for the packages doesn't exist yet :)13:12
jdstranddredg: exactly13:12
MithrandirI disagree about the site specific bit.  If you are using pam_ldap, you most likely want one type of config.  Ditto if you're using pam_mount.13:13
jdstrandMithrandir: but no kerberos, for example13:13
jdstrandMithrandir: and if you throw disconnected users in, there are various ways to do caching13:14
jdstrandnot kerberos13:14
Mithrandirjdstrand: kerberos generally consists of putting pam_krb5.so into the local login files, not much work and very much stock13:15
Mithrandiras for caching, why can't we choose a sane default and stick with it?13:15
jdstrandKeybuk: a-c-c doesn't scale either for large numbers of profiles, but then, I on't think that is a big concern13:15
Mithrandirnobody is suggesting the files shouldn't be configuration files, just that we have a sane way to update the config programatically.13:15
dredgputting pam_krb5.so in, depends if you want pam-heimdal or mit pam-krb5. then you need to configure the module - do you want the config in krb5.conf or in the pam file?13:16
jdstrandMithrandir: a-c-c will allow for programatically updating these13:16
dredgit's site specific.13:16
cjwatsonwhat would base-files have to do with PAM configuration?13:16
jdstrandcjwatson: base-files provides nsswitch.conf13:17
cjwatsonah, ok13:17
jdstranda-c-c supports profiles for just changing nsswitch.conf, or just changing pam as well13:18
jdstrand(ie, a profile doesn't have to do both-- there is a profile for using cracklib for example)13:18
pittiyay, I killed 50 hal bugs by triaging; does that count as 5-a-day? :)13:19
dholbachpitti: absolutely :)13:20
* seb128 hugs pitti13:20
jdstrandKeybuk: I think there is some potential here for simplifying large deployments, as well as providing users with a good sane profile on install of a pam package-- but the packaging scripts policy need more thought13:20
* seb128 hugs dholbach13:20
jdstrandI am happy to update a-c-c as needed for this of course13:21
* dholbach hugs seb128 back13:21
* Hobbsee hugs seb128 and dholbach13:23
* dholbach hugs Hobbsee too13:23
dholbachis it a Hug Day and I missed it? :-)13:23
Hobbseedholbach: what does this mean?13:23
Hobbsee - Discussion of the process for requesting someone participation13:23
Hobbsee(temporarily or permanently). Agreement to review past activity and13:23
Hobbseedocument process to smooth such occurrences if they were to happen in13:23
Hobbseethe future.13:24
* seb128 hugs Hobbsee13:24
dholbachHobbsee: there has been long mailing list threads and we thought it'd help to massage this into a concise wiki document13:25
Hobbseedholbach: i don't understand.  the first is not valid english, and words may be missed, and i don't see the relevance of the second statement to the first.13:25
dholbachugh, you're right13:26
dholbach"cease" is the missing word13:26
Hobbseeright, that makes more sense then13:26
dholbachI'll ... err ... follow up :)13:27
Hobbseei didn't think i was that dense.13:27
* lamont finally notices the bind9 upload to hardy from last week, decides he maybe better fetch it and deal with adding that13:39
lamontjdstrand: did that get coordinated with a change in debian AppArmor?  or did we just fork bind9 while I wasn't looking?13:40
jdstrandlamont: ubuntu specific-- but changes are minimal13:41
lamontmoving ownership of a file is not _MINIMAL_13:42
jdstrandlamont: you mean usr.sbin.named?13:42
lamont  * debian/control: Replaces apparmor-profiles << 2.1+1075-0ubuntu4 as we13:42
lamont    should now take control13:42
* jdstrand nods13:43
lamontwhich requires coordination with the owner of apparmor13:43
jdstrandlamont: I meant simply in terms of the size of the diff13:43
lamontso that it _STOPS_ delivering it13:43
jdstrandlamont: I coordinated that with ubuntu's apparmor13:43
lamontright.  meaning that the package is now _FORKED_ until someone works with the debian apparmor maintianer13:44
lamontwhere it used to be one that we could just sync.13:44
* jdstrand again nods13:44
jdstrandlamont: this was a server team FF directive13:45
lamontso now I get to decide if it's worth generating debian/control as part of the build (which makes me vomit), or just let y'all keep merging13:45
lamonthas anyone even _started_ the discussion with debian's apparmor?13:45
jdstrandlamont: not that I am aware off13:45
mathiazlamont: there isn't any apparmor package in debian.13:45
lamontmathiaz: that does help some...13:46
lamontis hardy going to let one choose between selinux and apparmor? or are we still stuck apparmor crap?13:46
mathiazlamont: there is work done on the selinux front too.13:47
jdstrandlamont: selinux is in universe, but it will be much better supported by the system13:47
jdstrandlamont: with a kernel boot option, selinux can be turned on and apparmor disabled13:47
lamontjdstrand: selinux has always struck me as the better of the two...13:47
jdstrandlamont: it is in terms of security-- no doubt13:48
jdstrandlamont: the problem is usability for end-users13:48
ScottKlamont: Enabled apparmor is better than disabled SE Linux.13:48
lamontScottK: I'm not 100% convinced of that.13:48
mathiazlamont: see https://wiki.ubuntu.com/HardySELinux for the state of selinux in Hardy13:48
jdstrandlamont: apparmor provides much better security than not having either, but is easier for people to get into13:48
lamontreally good locks on a cheap door are not better than no locks on a solid door: they lead the owner to _THINK_ that he's secure, when he's not.13:49
jdstrandlamont: while I understand your point, I disagree.  apparmor is actually quite good with network daemons-- less so for local confinement (eg firefox)13:49
* lamont didn't say apparmor was worthless... just that it's not the panacea it's sold as.13:50
jdstrandbut still better than nothing for local confinement (every hurdle is better than nothing)13:50
lamontjdstrand: so long as the user understands that it's just a hurdle13:51
jdstrandjdstrand: everything is just a hurdle-- until you disconnect the network :)13:52
jdstrandand for local confinement, the power cord13:52
jdstrand(nice that I responded to myself, eh)13:52
lamontjdstrand: is it safe to assume that the patchset isn't in git anywhere?13:53
jdstrandlamont: not git for bind913:54
jdstrand(apparmor is in bzr though, but I doubt you care about that part at this point)13:54
lamontXS-Vcs-Git: git://git.debian.org/~lamont/bind9.git13:54
lamontI was noticing that I hadn't actually uploaded postfix to debian last night, did that, and was tossing a couple more changes into bind9 before uploading that... which is how I noticed that someone had uploaded a new bind9 to ubuntu13:55
lamontso now this evening after work, I get to do more stuff, or I could save that for you.. :)13:55
jdstrandlamont: as debian doesn't have apparmor, I assumed the apparmor portion would just be a diff ubuntu would carry13:56
jdstrandI can do the git bit, but not until next week (I am technically on vacation right now)13:57
lamontjdstrand: why?  the package hasn't diverged in the past, other than security uploads for versions that debian wasn't shipping13:57
lamontiz OK.  I'll smack it around this evening13:57
jdstrandlamont: why what? why apparmor at all?13:57
lamont"why" == "why have an ubuntu diff and all the pain that brings"13:58
lamontadmittedly, BIND is not postfix...13:59
jdstrandlamont: apparently my assumption was wrong wrt to git and the debian packaging13:59
lamont(postfix discovers which distro it's building for at build time and does the right thing)13:59
Riddellsiretart: libxine1 is brining libmad onto the CDs13:59
asachow can i force mutt to reply to all instead of Mail-Follow-Up: ?14:00
lamontjdstrand: generally, if the debian maintainer is an active ubuntu devel (esp core team), one can expect that they're trying to do it all in one branch, unless that stops making sense...14:00
lamontand "stops making sense" goes a little beyond  some other definitions....14:00
lamontasac: if 'g' doesn't do it, then use cut-n-paste, your editor, and your firm confidence that you know better than the sender.14:01
seb128lamont: you can't expect everybody to know about that and try to get changes to debian first though14:01
jdstrandlamont: ok. well, like I said, if you'd like me to do the git stuff, I can next week.14:01
lamontseb128: I can too.  It's just not reasonable.14:01
seb128lamont: agreed that's better, when that doesn't happen the debian maintainer can also grab changes14:01
lamontseb128: yeah.  I'm just being grumbly this morning14:02
lamontwhile I'm grumbling....14:02
lamontis network-manager supposed to auto-discover a wireless net and connect to it?  I nuked it and reinstalled it on my gutsy laptop, and now I just have manual mode...14:03
lamontOTOH, gutsy is the first time that it was getting sane enough for me to consider leaving it on..14:03
lamontjdstrand: and pardon my being cranky14:04
ScottKlamont: You have to manually edit /etc/network/interfaces back to auto mode14:04
* lamont had assumed that apparmor was in debian as well, which was wrong of me.14:04
jdstrandlamont: np :)14:04
lamontScottK: and how does one know it's in "auto mode"?14:04
lamontand back to work for me14:04
=== ScottK3 is now known as ScottK2
ScottK2lamont: When /etc/network/interfaces says auto14:06
jdongintriguing. My macbook uses nearly 2W less power after a suspend than before.14:06
jdong(I meant after waking up from a suspend vs a cold boot)14:06
jdongbryce: oh btw, the xv debacle gets even weirder....14:08
lamontScottK2: lew;/14:09
lamontkewl. even14:09
jdongbryce: it seems like only shortly after a VT switch into X, xv causes a hang. If I wait 10 secs or more after Compiz initializes or coming out of a suspend, it works just fine.14:09
jdongbryce: could that be some kind of AIGLX initialization race condition?14:09
siretartRiddell: how that?14:21
\shsiretart: libxine1: Recommends: libxine1-ffmpeg and suddenly we have Recommends installed by default? ,-)14:26
Riddellsiretart: because it depends on libxine1-plugins which depends on libxine1-ffmpeg14:26
siretartRiddell: it installs recommended universe packages?14:27
Riddellit's a depends.14:28
siretartAFAIUI, recommends to universe packages should be installed by default14:28
\shasac: looks like that ff3 and flash-nonfree doesn't honour my sound settings from gnome...the sound is still coming from the main speakers instead of the headset (other apps are working as expecting and giving me sound on the headset)14:28
cjwatsonrecommends-by-default isn't on yet14:28
cjwatsonwith the exception of metapackages14:28
\shsiretart: forget what I said...what riddell said is more correct14:28
asac_\sh: you use sound server?14:29
\shasac_: well, what are we using by default= sound setting tells me: USB Audio  (Sennheiser)  so I think no sound server14:29
siretartRiddell: -plugins is a convenience package which drags in all xine packages. I didn't expect it to be an installation candidate on the live cd.14:30
slytherinjamesh: Can you please tell me (or publish somewhere) what is the format of entry to be added in rhythmdb.xml for the FM radio plugin (using tuner card)?14:30
cjwatsonsiretart: did you intend libxine1-misc-plugins to be used instead?14:30
asac_\sh: can you see if gnash works better?14:30
asac_\sh: (sound wise)14:30
\shasac_: if gnash plays youtube, I can test it14:30
cjwatsonsiretart: if so, you should reverse the order of the |-ed dependency in libxine114:30
asac_\sh: it should play a few14:31
\shasac_: can I tell ff3 to use gnash exclusively without uninstalling adobe?14:32
asac_\sh: you could try to link the gnash.so to the .mozilla/plugins/ directory14:34
asac_maybe that will prefer it then14:34
asac_otherwise update the alternative14:34
siretartcjwatson: well, we used to ship all xine modules in earlier releases. now the modules are split across smaller packages. I had an argument with the debian release team, and vorlon explained to me that not depending on all plugin packages would break partial upgrades14:36
siretartI understand that this affects dapper->hardy upgrades14:37
\shasac_: I updated the alternative...but no gnash14:37
siretartTBH, I'd prefer to just drop the dependency on -plugins14:38
\shasac_: no ways ... it prefers adobe...purguing it14:39
=== asac_ is now known as asac
peciskwhich package provides live cd grub menu at the begining? I need to translate it but I can't find which package it is14:41
asac\sh: the alternative is named xulrunner-addons-flashplugin14:42
asacanyway purging should help too14:43
asacbe sure that you have no flashplugin in your .mozilla/plugins dir14:43
\shasac: yep14:43
asacwho is eclipse/swf package liason for ubuntu14:43
cjwatsonpecisk: that isn't grub14:46
cjwatsonpecisk: gfxboot-theme-ubuntu is the one you're looking for14:46
peciskok, then what it is?14:46
peciskthanks ;)14:46
\shasac: tells me all the time, that I have to install a flash plugin...gnashplugin is installed, but it gives me all the time this...(youtube, video.google. etc.)14:47
asac\sh: oh i think the link is still missing :)14:47
asac\sh: do ln -s /usr/lib/gnash/libgnashplugin.so $HOME/.mozilla/plugins/14:48
\shasac: ok...now...youtube works somehow, but no sound14:50
\shasac: video.google no ways to get it to work14:50
xivulonIf anyone has a 64bit machine + windows + 5GB free, please test Wubi on latest daily ISO14:51
asac\sh: interesting ... maybe the the fluendo mp3 plugin for gstreamer works better14:51
=== cr3_ is now known as cr3
xivulonwe have mixed reporst on that, and some clarification would be nice in light of the coming release14:51
\shasac:  15:52:03: ERROR: Unimplemented: MovieClip.attachAudio()14:52
asac\sh: try a different youtube video ... there are a few players in use that use different swf functions afaik14:54
\shasac: well, I have tried a  lot of...even our own app (which is using flash normally) is not working ;)14:55
ScottKasac: re eclipse I don't think we have one.  It's generally whoever touched it last (which thankfully isn't me any more).14:55
xivulonWhether wubi works (or not) please report it to myself or evand in #ubuntu-installer14:56
asacScottK: someone claimed that latest milestone supports xulrunner 1.914:56
\shasac: totem <foo>.flv (downloaded with youtube-dl) works like a charm :)14:57
brettaltonCan I get help on compiling a package into .deb format?14:57
brettaltonam I in the right place?14:57
ScottKasac: It's going to stay not me who touched it last.  Sorry.  No time to look into it.14:57
asac\sh: ok then gnash would work as well i guess14:57
\shasac: gnash doesn't play .flv just swfs14:57
\shasac: but gnash failes when opening a youtube url, (where the player comes from)..15:01
brettaltoncan someone help me compile ProjectPier and/or KohanaPHP in .deb format?15:02
brettaltonKohanaPHP should be similar to phpMyAdmin where it installs to /usr/share/phpmyadmin15:02
=== dholbach_ is now known as dholbach
\shasac: any possibility to set the adobe plugin to a more "noisy" mode? or how would you debug ff3 + flash plugin to find out what it does15:05
jakob__Hi! The infamous "High frequency of load/unload cycles on some hard disks may shorten lifetime"-Bug:15:07
jakob__What does it take to get the Debian Patch into Hardy?15:07
jakob__And then as a SRU into Gutsy.15:08
jakob__Patch is here already: http://launchpadlibrarian.net/12024340/debian-hdparm-fix.debdiff15:08
jdongjakob__: we do not do that by default, and I assume anyone who turns on that acpi-support option wants aggressive power savings15:09
jdongwhich doesn't come free15:09
Ngthat patch is not a fix, it's a workaround at best and harmful at worst15:10
jdong128 still lowers disk life just to a lesser degree15:10
jdongit also saves power to a much lesser degree15:10
jdongwe should do a FFT on the disk activity history to predict when it's optimal to park the disk!!!15:11
jakob__Well, i didn't turn on acpi-support.15:11
jakob__It's the HD default.15:12
jdongjakob__: what package is that against? hdparm?15:12
jakob__The patch is a workaround, yes.15:12
jakob__No, against acpi-support.15:12
jakob__It's installed by default.15:12
jdongacpi-support is installed by default, but we do not set any hdparm disk modes by default15:13
jakob__Laptop HDs have crazy power managemant defaults.15:13
jdongI don't see why we should be forcing a particular mode (128)15:13
cjwatsonjakob__: http://mjg59.livejournal.com/77672.html if you haven't read it already15:13
jakob__Seen it already.15:14
jdongjakob__: well I don't see why Ubuntu should blindly apply such workarounds to all systems when those affected can do it themselves15:14
jdongjakob__: I don't make Ubuntu send my laptop into suspend and resume in rc2.d15:14
jakob__Yeah, but the workaround is quite difficult.15:14
jdongeven though my particular laptop requires that :)15:14
jdongmy hard drive defaults to -B 1 (disabled). Setting -B 128 will lower my disk life by that argument.15:15
Keybukjakob__: I've yet to see conclusive proof of anybody with a laptop hard drive killed by their BIOS's default power management settings15:15
jakob__Probably the workaround should become a independent package?15:15
jdongjakob__: not to mention I have one laptop harddrive with 15 million park cycles15:15
jakob__You mean, the excessive parking is not harmful?15:15
jdongand it is still working15:15
jdongjakob__: there's no conclusive proof of it, no15:15
jakob__Ok, i see.15:15
Keybukjakob__: on all of the laptops I've looked at, the amount of parking was consistent with the manufacturers recommendations and expected life of the drive15:16
jdongjakob__: the load cycle limit spec on the disk is not exactly a ticking timebomb15:16
Keybukit hasn't helped that most people pasting SMART values have misunderstood them and actually not pasted the important values!15:16
jakob__Ok, so what to do about the angry users (lots of them out there)?15:16
Keybukthe load cycle limit specifies when the drive will be considered "old" by the manufacturer15:16
jdongjakob__: what do you propose we do? Set the wallpaper to a large bold message explaining their blogposts were false?15:16
KeybukUbuntu increases this at a rate that would roughly coincide with the time factor of when it would be considered "old" as well15:17
cjwatsonjakob__: Matthew already posted a clear response and it's been largely ignored or misread; what do you propose we do?15:17
jakob__Package the workaround?15:17
Keybukjakob__: users feelings have been stirred up by the media, without any technical basis for the problem15:17
jakob__To make it easy to apply.15:17
Keybukthe workaround isn't working around anything though15:17
jdongjakob__: no other OS does something this silly15:17
jdongjakob__: and your workaround penalizes (parks more often) properly working hard drives15:18
cjwatsonwe package workarounds to real problems, not injured feelings15:18
Keybukit's a workaround for bad publicity15:18
Keybuknot a workaround for technical problems15:18
jakob__Well, Debian does it...15:18
cjwatsonif it is demonstrated to be a real problem, then it becomes a different matter15:18
jakob__Ok. Then the Bug shouldn't be critical?15:18
\shjakob__: tell those users to buy a sun x4500 with 48 500g sata drives...and that they have to renew them every year at least once...even if the manufacture of the drives are not saying they are old...what you have in your desktop is more crappy then that what you have in your laptop...regard the fact, you are travelling with your laptop in suspend/switched off mode a lot more then you do with your desktop/server15:18
jdongjakob__: the only thing that I would do in Ubuntu is in the comments of /etc/default/acpi-support, warn that ENABLE_LAPTOP_MODE sets the hard disk to aggressive mgmt by default15:19
Keybukjakob__: the bug is marked critical by a user15:20
jakob__I see.15:20
jdongif your disk firmware or BIOS is setting insane settings.... that's not really the business of the OS to resolve15:20
\shcjwatson: the old toshiba r200 used on windows a tool which parked the drive head every time you touched the laptop a bit harder...just because to protect the user for drive damage...it harmed the user more to park the head, then to disable this tool in general15:20
Keybukjdong: arguably the OS should help15:20
Keybuk(we used to fix insane DSDTs)15:21
=== Ubulette_ is now known as Ubulette
Keybukbut we've yet to see evidence of alleged insane settings15:21
jdongKeybuk: but help in a more sane manner than blindly forcing -B 12815:21
jdongKeybuk: monitoring how SMART values change over time could be useful for a Hardy+1 spec15:21
Keybukisn't that exactly what smartd does? :p15:22
jdongKeybuk: isn't that what we don't offer by default or have a nice dbus notify bubble frontend for? :)15:22
\shcjwatson: especially this little drive is working after the laptop felt on the floor...display dead, drive alive15:22
Keybukthat's a different matter :)15:22
jakob__Probably make the pm setting GUI-configurable?15:23
Keybukjakob__: e.g.15:23
Keybuk193 Load_Cycle_Count        0x0032   059   059   000    Old_age   Always       -       41291615:23
jdongKeybuk: oh btw, do I cry to you if sysfs power supply sucks at detecting that my ac cord is unplugged?15:23
Keybukfrom my own laptop15:23
Keybukit self-claims that the drive has 59% of its load cycle count left15:23
peciskin which package I can find graphical installer translations?15:23
Keybukuntil it is considered old15:23
jakob__Hm those values are seldum linear15:24
Keybukit's about 18 months old, and the drive is manufacturer rated for three years15:24
Keybukjakob__: the entire point of the three digit ones is that they are normalised15:24
Keybukit's the last number that's never linear15:24
cjwatsonpecisk: debian-installer15:25
cjwatsonpecisk: (I know it's a little weird to look there, but all the installer translations are collected there)15:25
Keybuk(and in most of the hysteria, it's only the last number people have looked at - and compared it to a number somebody once plucked out of the air)15:25
peciskthanks guys15:26
jakob__Retracting while on AC power is quite useless, the drive can't know about that.15:27
jakob__Ok, the dev position is that it doesn't harm either... Probably, yes.15:28
Keybukquite useless ?15:28
Keybuk * Scott uses his laptop in bed while plugged into AC power, he'd like it to not overly burn him or have head crashes15:29
jakob__Dropping it with a power cord plugged in is unusual15:29
ScottKActually my most common laptop drop scenario is tripped over the power cord.15:29
KeybukI think that the only time I've ever "dropped" a laptop has been *because* of the AC cord!15:29
jakob__Have you applied the workaround?15:30
jakob__The "workaround"15:30
KeybukI have no reason to15:30
* ScottK just leaves it all in it's default configuration.15:30
Keybuk193 Load_Cycle_Count        0xc720   088   214   032    Old_age   Offline      -       22278870891100815:31
Keybuk(pretty much proof that the load cycle count doesn't kill hard drives)15:32
Keybukthat one still works just fine15:33
jakob__Looks like corrupt data15:33
Keybukit came out of a Sky+ box15:33
Keybukthey have *really* aggressive power management by default15:33
Keybuka few years15:33
Keybukless than 515:33
jakob__That's 807380 load cycles.15:34
jakob__PER SECOND.15:34
Keybukerr, more than 515:35
jakob__+/- few years doesn't make a difference here. That's corrupt data.15:36
Keybukthey pretty much do "load mpeg frame", "park", "unpack", "load mpeg frame", etc.15:36
jakob__Yeah, but nut 100000 times a second.15:36
Keybuksince it counts load and unload15:38
Keybuk0.2ms response, that's about right15:38
Keybukbut pretty much proves that15:39
Keybuk a) you can't trust that number for anything15:39
Keybuk b) its value is meaningless15:39
Keybuk c) hard-drive manufacturers hate you15:39
KeybukI went digging last time this came up15:40
cjwatsoneven if it is corrupt: if it can be corrupt in one drive, it can be corrupt in another, even if the resulting number is smaller15:40
Keybukand proved that according to the manufacturer, you can't operate a hard drive within the usual temperatures for a desktop computer15:40
cjwatsonfor evidence you need an actual scientific controlled experiment15:40
jakob__Keybuk: You can't what?15:41
Keybukjakob__: the drive in my desktop is usually about 31°C iirc15:43
Keybuk(I put a thermometer on it)15:43
peciskin which package I can find network-manager GUI strings? I can't find it :(15:43
Keybukit's only rated by the manufacturer up to 28°C15:43
jakob__Really? Strange drive...15:43
seb128pecisk: network-manager-gnome15:44
jakob__Another point is that the load-unload click is15:44
jakob__1) annoying15:44
jakob__2) raises latency15:44
jakob__That's why i applied the workaround.15:44
=== pedro__ is now known as pedro_
peciskseb128: https://translations.launchpad.net/ubuntu/hardy/+source/network-manager-gnome/+pots/network-manager-gnome/lv/+translate15:45
peciskit says it can't find it15:45
jakob__I don't know if "annoys people" is enough to do something about it. Is it?15:45
seb128pecisk: the source network-manager-applet15:48
peciskok, thanks15:48
seb128you are welcome15:49
cjwatson"annoys people" is not enough if the workaround causes damage to other people's systems15:49
cjwatsonwhich has been alleged here15:49
jakob__Would a patch that makes the pm setting configurable through some /etc file make it through? I'd probably write that then.15:49
cjwatsonsurely it is already configurable in /etc/hdparm.conf15:49
jakob__Not after resume?15:50
jakob__Because of that users download strange stuff from forums and drop it anywhere in their systems.15:51
jdonghook restart hdparm into resume?15:51
jdong*shrug* sounds possibly reasonable15:51
jakob__Yes, something like that.15:51
cjwatsonI'm sorry, but we cannot prevent users downloading strange stuff from forums and dropping it into their systems15:51
cjwatsonI know that the ubuntuforums guys try to keep the level of that sort of thing under control15:52
Keybukcjwatson: perhaps its time to resurrect that idea of not giving users root on their own machines? <g>15:52
jakob__You can. By making it configurable by default.15:52
jdongyeah, we do, but again users tend to come up with interesting workarounds faster than we can read them :)15:52
cjwatsonjakob__: that won't actually help15:52
cjwatsonpeople will continue to fail to find the configurability for one thing or another, and instead do something else15:53
jakob__Yeah, of course.15:53
cjwatsonmaking it configurable on the desktop might help15:53
cjwatsonsomething in /etc probably won't make a whole lot of difference15:53
jakob__But in this case, the configurability doesn't exist.15:53
cjwatsonthough for things that are configurable on the desktop, we need to be very sure that we completely understand the issues and can explain them clearly in the UI15:54
cjwatsonotherwise it's:15:54
cjwatsonMake my computer work one way <-----------> Make my computer work a different way15:54
cjwatsonpossibly s/work/break/g15:54
jakob__Yeah, of course.15:54
jakob__But hdparm should be hooked into resume.15:55
jakob__Because the hdparm.conf is quite pointless otherwise,15:55
cjwatsonI think that would be reasonable15:55
jakob__(when on a laptop)15:55
jakob__But it doesn't know about AC or batt then15:58
jakob__Just leave that to laptop-mode..?15:59
=== warp10 is now known as warp10_
pittiseb128: sorry, have class now; I think I fixed the retracer, but we lost some bugs due to untagging and crashing16:00
pittiseb128: do you think you can take a look at the log, re-tag some bugs, and check whether it works now?16:00
seb128pitti: will do, thanks16:01
KeybukI've always mused how hardware events get dealt with during suspend16:01
Keybukif you unplug a device while suspended16:01
=== warp10_ is now known as warp10
Keybukhow does it know?16:01
LaserJockmy computer knows eeeeeverything16:02
jakob__Hooking hdparm into resume... There is a kernel argument "nohdparm" that has to be obeyed.16:07
jakob__So it's not _that_ trivial.16:07
Keybuksurely you just /etc/init.d/hdparm restart ?16:08
jakob__Ah. Right. That handles the kernelarg.16:09
jakob__"Feature freeze" means that can't get into hardy, right?16:10
jakob__Hm, won't that overwrite laptop-mode settings at resume?16:10
Keybuksodomy non sapiens16:13
jakob__Anyway. Does somebody think about writing a patch for this?16:14
jakob__If not i'd probably try.16:14
jakob__If somebody would probably commit it then.16:14
Keybukjakob__: PPA :-)16:16
jakob__Keybuk: Uhmm, which of these? http://en.wikipedia.org/wiki/PPA16:17
KeybukPro-pedophile activism16:18
Keybukoh man16:18
Keybukjakob__: this one https://help.launchpad.net/PPAQuickStart16:18
jakob__Thx :)16:19
mrcheeksCan I ask about a hardy related bug here?16:19
jakob__I got to go now, thanks for the discussion, i'll do something about the hdparm-resume-stuff next week.16:21
Keybukmrcheeks: if you have a patch for it, sure16:22
Keybukmrcheeks: if you have more information, it's better to just add it directly to the bug16:23
mrcheeksKeybuk:I don't know how to fix it, so I don't have the patch... I am not able to login from gdm or a terminal, I guess it is a bash issue16:24
Keybukmrcheeks: #ubuntu would be more able to help you16:24
\shsoren: is it possible to have sound output in one or more xen instance?16:32
=== \sh is now known as \sh_away
pittiseb128: ok, class is over17:03
pittiseb128: taking a look now17:03
geserpitti: Please give-back delo. It should build now with the fixed pkg-create-dbgsym. Thanks.17:12
pittiurllib2.HTTPError: HTTP Error 403: Forbidden17:23
pittigeser: done17:24
DktrKranz2pitti: stani confirmed Feisty is not affected by bug 124896. Can you delete 0.8.2a+repack-1ubuntu0.7.04 from feisty-proposed?17:28
ubotuLaunchpad bug 124896 in spe "spe 0.8.2a+repack-1 fails with python-wxgtk2.8" [Medium,In progress] https://launchpad.net/bugs/12489617:28
pittiseb128: meh, hardy apport chroot is all crash-o-rama due to p-lp-bugs17:31
pittiseb128: I guess I will install the upstream trunk patch on Monday17:31
pittiI just need to leave soon17:31
pittiso if you want to give it a shot, please go ahead17:32
seb128ok, I might have a look17:32
pittiseb128: use bug 194406 as test case17:32
ubotuLaunchpad bug 194406 in gnome-utils "gnome-dictionary crashed with SIGSEGV" [Undecided,New] https://launchpad.net/bugs/19440617:32
pittiI opened it17:32
seb128would you suggest rather picking changes or trying to use trunk?17:32
pittiit needs apport-chroot login (see history), edit python files, save, re-tagbug, run crash-digger17:32
pittiwash rinse repeat17:33
pittiseb128: I honestly don't know17:33
seb128ok, I'll manage17:33
seb128thanks for the efforts you already put into it17:33
pittiseb128: hm, now that I opened it you should actually be able to use "apport-retrace -s 199406" in the chroot17:33
seb128I should be able to do those changes17:33
pittithat's much quicker than re-compressing the chroot everytime and calling the digger17:33
seb128ok, cool17:33
* seb128 hugs pitti17:33
pittiso, edit the python, call apport-retrace -s 199406 until it works17:34
* pitti hugs seb128, thanks17:34
dholbachMOTU Q&A Session in #ubuntu-classroom in a bit17:57
pochuHow's alpha 5 coming?18:20
davmor2pochu: getting there :)18:21
pochudavmor2: cool. It's targetted for today, right?18:21
davmor2as far as I know.  Most of the cd's are tested a few bugs but nothing really major anymore18:22
slangasekyes, it's still targetted for today18:25
slangaseksooner if someone wants to help test ubuntu studio, xubuntu images18:26
davmor2slangasek: burning Xub now is it just the desktop?18:27
Keybukslangasek: so, one piece remains18:27
slangasekjust desktop this time, yes; alternates were oversized and I got no response from xubuntu devs to my pings about it18:27
slangasekKeybuk: ?18:27
Keybukchange the default common-auth in libpam-runtime, this would still seem to require postinst fu to change the existing common-auth since it won't do it othewise18:28
Keybukapply sed-fu in libpam-thinkfinger postinst to modify common-auth when it's installed18:28
slangasekKeybuk: well, there's existing pam postinst fu for upgrading to a new default common-auth18:28
Keybukuse auth-client-config to change it18:29
Keybukslangasek: execpt the existing fu doesn't get used except on install18:29
slangaseker, not true unless someone broke it?18:29
slangasekif [ -z "$2" ] || dpkg --compare-versions "$2" lt 0.76-1718:29
Keybukah, so I could just increase that version?18:29
Keybukwouldn't that break the "respect the user deleting the conffile" check though?18:30
Keybuksince it'd put them all back18:30
slangasekand stow the md5sum of the current config file in the list18:30
slangasekummm... I suppose, but I'm generally uninterested in use cases that involve users willfully breaking their setup :)18:31
KeybukI just mention it, since the comment in that file implies the maintainer wanted to preserve that18:31
Keybukand since the maintainer is either you18:32
Keybukor a prior incarnation of you18:32
Keybuk... :)18:32
davmor2does anyone know the current state of gvfs' obex-ftp module?18:32
slangasekKeybuk: well, since I failed to implement my comment correctly, feel free to ignore it too :-)18:33
Keybukslangasek: :-)18:33
slangasekwe can fix it up later, low priority18:33
Keybukslangasek: so do you think that libpam-runtime is the right place to do this?18:33
jdongpitti: may I try to plea the case to you that the two-batteries-in-gpm issue was fixed wrongly?18:33
Keybukif so, should that recommend the other package?18:33
jdongpitti: i.e. sure "a battery" is not shown anymore but IMO the wrong one is being shown18:34
slangasekKeybuk: yes to the first, for the second I have no opinion18:34
jdongpitti: or do you feel this should be a bug filed against the kernel or hal for sysfs battery readouts sucking?18:34
Keybukpam is ok if a module mentioned in common-auth doesn't actually exist?18:34
slangasekthe config does need to be tested to make sure it works when the module doesn't need to exist, but AFAIK it's /supposed/ to work18:35
Keybuk(I assume so, since pam_foreground mentioned in common-session doesn't exist)18:35
slangasekyeah, should still be tested to make sure there are no differences wrt "optional" vs. "sufficient" modules18:36
Keybukslangasek: so why isn't the current common-auth md5sum in the common-auth.md5sums file?18:37
slangasekKeybuk: because we don't bother adding it there until the file itself has been superseded18:37
Keybukah right18:37
=== afflux_ is now known as afflux
Keybukslangasek: http://people.ubuntu.com/~scott/pam_0.99.7.1-5ubuntu5.debdiff18:41
Keybukslangasek: does that looks reasonable to you?18:41
slangasekKeybuk: is 'try_first_pass' really appropriate here? is pam_thinkfinger ever going to prompt for a password?18:46
slangasekdoesn't hurt much if it doesn't, and would be consistent with a config generated by the new pam framework stuff, so no big deal18:47
slangasekotherwise, looks good18:47
Keybukslangasek: it does prompt, yes18:47
Keybuksince you obviously want to be able to swipe your finger *or* type a password18:47
slangasekthat's true, but what does pam_thinkfinger think it's going to do with the password when it has it? :P18:48
slangasekI'd call that a bug in pam_thinkfinger for prompting for things that are none of its business18:49
Keybuknothing, it just passes on to the next authentication18:49
Keybukdoes PAM give you any way to prompt for something that's not a text entry field? :)18:49
slangasekthere's a Linux-PAM extension for binary prompts, yes...18:49
Keybukcan you have multiple PAM modules active at once18:49
slangasekit's basically a poor reimagining of SASL :-P18:50
Keybuksimultaneously have pam_thinkfinger and pam_unix waiting for input18:50
Keybukwith one cancelling the other?18:50
Keybukso it wouldn't work your way ;)18:50
Keybukyou'd have to swipe your finger, or press some key to move on to trying a password, no?18:50
slangasekyeah, with that constraint, that's valid18:50
slangasekunfortunately, it means that if your next module in the stack is something like, say, pam_opie, you don't get a useful prompt18:51
slangasekone-time-pass module18:52
KeybukI think upstream basically just want to bypass PAM for the biometric stuff18:52
Keybuksince it doesn't quite fit the model18:52
slangasek(hmm, worse would be trying to stack with SecureID... :)18:53
=== zachr_ is now known as zmr5r
Keybukit's not so much a stack, as an alternative18:54
slangasekfor PAM values of "stack"... :)18:54
Keybuke.g. at a login prompt, you shouldn't even need to ask for a username18:55
slangasekdavmor2: hrm, I definitely haven't noticed bug #194007 in my testing of Ubuntu, I think I would've noticed if my keyboard had suddenly switched out of Dvorak... was this Kubuntu-specific, or maybe keymap-specific?18:55
ubotuLaunchpad bug 194007 in ubiquity "Ubiquity keyboard is not being setup after world map." [Undecided,Confirmed] https://launchpad.net/bugs/19400718:55
evandslangasek: it's just not picking a proper default.  So you'll end up on USA - USA every time.18:57
slangasekKeybuk: so what's the prompt that pam_thinkfinger gives the user as the alternative?  Does it ask for the username first, or the password?18:57
=== \sh_away is now known as \sh
slangasekevand: oh, so it should be hinting a keyboard preference and it isn't?  Fair enough18:58
evandslangasek: indeed, I've updated the bug title to reflect that18:58
slangasekI don't think I'll bother release-noting that one18:58
davmor2slangasek: No across the board.  It does actually change your keyboard.  But I for example set my language as english and time zone as London.  With the old ubiquity this set my keyboard to Uk English.  Now however it stays on Us English until you change it.  This used to be automatic.18:58
slangasekdavmor2: right, understood18:58
davmor2s/does/doesn't sorry18:59
Keybukslangasek: it doesn't ask for a username yet18:59
Keybukright now you get "Password or swipe finger"18:59
slangasekKeybuk: hrm.  unfortunately there'll be cases where PAM already knows the username anyway, so you don't want to prompt for it unconditionally... otherwise, a username prompt would let you handily bypass prompting for passwords that don't belong to it19:00
davmor2wine is a bit bust seg faulting19:02
=== cjwatson_ is now known as cjwatson
cjwatsonevand,davmor2: weird, I'm sure it selected UK for me19:02
evandcjwatson: I'd be very surprised if it did.19:03
evandI can't for the life of me figure out why this is happening though.  There haven't been any console-setup changes between the version that worked and this one.19:03
davmor2cjwatson: It's happened on every live install so far 64 bit and 32 bit on Ubuntu and Kubuntu19:03
cjwatsonoh, I might have forced it in the bootloader19:04
cjwatsonmy suspicion would be timezone widget changes19:04
Keybukslangasek: this is all, it has to be said, "future versions"19:04
evandcjwatson: I thought so too at first, but it's basically the same code for that.19:04
slangasekKeybuk: yes :)19:04
davmor2cjwatson: Kubuntu's has though has it?19:04
evanddavmor2: no, kubuntu uses its own timezone widget19:04
Keybukslangasek: they want to get to the point where you can walk up to a computer and just swipe your finger19:05
davmor2same error on kubuntu it's where I first noticed it19:05
Keybukit'll do a user switch and login there and then19:05
slangasekKeybuk: aieee19:05
davmor2evand: Now you see that makes a lot more sense :)19:08
davmor2sorry back to my question earlier.  Gnome-vfs-obexftp got upgraded to main the day after the whole backend got switched to gvfs/gio which means that the bluetooth applet does next to nothing again is this likely to get fixed before hardy is released or is it more likely to be a Intrepid.19:11
davmor2slangasek: Xubuntu isn't booting I'll try a reburn but it could be dead :(19:12
slangasekdavmor2: mm, at what point does it fail to boot?19:12
davmor2pass but I did the check disk for errors and it's shown one up hence the reburn.19:13
dholbachLibrary Packagin Session in #ubuntu-classroom19:14
=== ion___ is now known as ion_
slangasekoh bother, how did I manage to get my UI in Spanish on my last login19:26
dholbachslangasek: it must have been the end of a very long night :)19:27
slangasekdholbach: it doesn't bother me too much, except when I'm trying to navigate mutt and all the keybindings are changed19:29
dholbachhehehe, I can imagine :)19:29
dholbachor should I say jéjéjé (in a good impersonification of pedro_)19:30
* pedro_ hugs dholbach19:31
pedro_it should be the default language ;-)19:31
* dholbach hugs pedro_ back19:31
dholbachsure... like seb128 would vote for french :)19:31
pedro_why? he's german19:31
dholbachAlsatian :)19:32
dholbachanyway... I call it a day - see you on monday and have a great weekend :)19:32
ogra_cmpcciao dholbach19:32
dholbachbye ogra19:32
dholbachand give sistpoty some love in  #ubuntu-classroom !19:33
slangasekdavmor2: any luck with the xubuntu reburn?19:37
davmor2slangasek: trying it now19:37
ogra_cmpccjwatson, will you kill me if i add a edubuntu-desktop.postinst to -meta ? i think it needs a notification to re-login now that its always installed as addon19:41
Kunisorry to bring this here, but over at +1 we were wondering what the status is on alpha 5. Could someone clarify please?19:42
ogra_cmpc(the artwork changes wont take effect without logging out, apps appear in the menu indeed)19:42
slangasekKuni: http://iso.qa.ubuntu.com/qatracker/build/all shows the testing status of the images that are being prepared for the release; when we have adequate test coverage to ensure we're not releasing broken images, we release19:45
Kunisweet, thanks19:45
davmor2slangasek: still issues I'm just check all the md5sums so far they check out I'm just checking the iso against the cd now19:46
slangasekwhich image is this, precisely?19:47
davmor2slangasek: cced43ba9e39f1853c06f890ebf4212e *hardy-desktop-amd64.iso19:48
davmor2slangasek: I think it's the re-writeable disc at fault19:52
davmor2so I'm going to write it to a standard disc19:53
davmor2slangasek: that's craked it :)19:58
davmor2slangasek: printing from FF3B3 doesn't work :(20:05
slangasekdavmor2: crash in texttops, or are you seeing a different problem than the one I just tested?20:08
* Keybuk wonders what tracker is indexing20:09
davmor2slangasek: stgraber: http://ubuntu.pastebin.com/f6308f46f that's stgrabers cups issue20:10
slangasekok, so it seems to be a cups problem of some sort?20:11
slangasekoh, 'empyt print file', hmm20:11
davmor2it only happens with FF320:11
slangasekok, is this being filed as a bug report?20:12
davmor2doing it now20:13
slangasekok, please give me the bug # when you have it for release-noting20:15
Keybukwhat the fuck is ~/xmp ?20:18
Keybukactually ~/xmp:-20:19
ogra_cmpclsof ?20:19
davmor2slangasek: how permanent is the pastebin can I link to it or should I copy the info across to the bug report?20:19
Keybuk"convert" apparently20:19
slangasekdavmor2: copy it across20:20
ogra_cmpcKeybuk, a freaked out nautilus thumbnailer ?20:20
davmor2slangasek: bug 19448620:21
ubotuLaunchpad bug 194486 in firefox-3.0 "printing in Firefox 3 Beta 3 is broken" [Undecided,New] https://launchpad.net/bugs/19448620:21
KeybukI have a feeling it might be a freaked out tracker indexer20:22
Keybuksomething like20:22
slangasekdavmor2: thanks20:22
ogra_cmpcuding convert ?20:22
Keybuktracker sees file, runs convert, convert freaks and dumps to ~/xmp:-, tracker sees file change, runs convert, convert freaks...20:22
Keybukit is tracker too20:23
davmor2slangasek: Xubuntu is downloading the language packs in ubiquity and taking forever over it.20:23
Keybuk \_ tracker-extract /home/scott/xmp:- image/gif20:23
ogra_cmpci thought that was from imagemagic20:23
Keybuk     \_ convert /home/scott/xmp:- xmp:-20:23
ogra_cmpcwhy does tracker convert images ???20:23
ogra_cmpcindexing must be fun on kwwii's disk with that20:24
LaserJockogra: my guess would be to provide the thumbnails in the search results20:24
ogra_cmpcoh, fun20:24
Keybukbasically, from what I can see20:25
Keybuktracker is chasing its own tail20:25
ogra_cmpcshould we rename it to beagle ?20:25
slangasekTheMuso, persia: any chance of getting some testing of the ubuntustudio images?20:31
TheMusoslangasek: Are they ready to test? If so, I can do a bit of testing.20:32
_MMA_I can also.20:32
* TheMuso syncs.20:33
davmor2slangasek: I feel sorry for anyone on dial up it going to be another 25 minutes ish for these lang packs20:33
slangasekTheMuso: yes, though I see I've got the wrong ones linked yet from the tracker, sorry about that20:34
slangasekTheMuso: please be sure you grab 20080221 rather than 2008022020:34
TheMusoslangasek: IF ITS THE CURRENT ONE< I'm getting it.20:34
slangasekhrm, ubuntustudio images are supposed to be DVD-sized then, not CD-sized?20:35
slangasekthey're > 700MB, but not marked as oversize :)20:36
TheMusoYes they are DVD size.20:36
_MMA_slangasek: The page itself says DVD. Though, some people still have a hard time burning it in windows. Even trying to use a DVD.20:39
davmor2slangasek: Xubuntu is okay but should it download the language packs?21:15
TheMusoHrm. I pulled the -es langpack from our seeds yesterday, yet they are still showing up in our CD report as uninstallable.21:16
TheMusoslangasek: Are the -es issues supposed to be fixed in the archive now?21:16
slangasekTheMuso: they are now fixed in the archive; I didn't bother rerolling the images for that issue, do you think it's important enough to do so?21:17
slangasekdavmor2: if you chose a language that requires downloading them from elsewhere, I imagine so?21:17
TheMusoluisbg: Do you want -es on the latest alpha?21:17
TheMusoslangasek: I don' but others on our team, for example luisbg might.21:18
davmor2slangasek: I choose english?21:18
slangasekdavmor2: well, right :)21:18
davmor2slangasek: I think though that it pulled down all of the language packs considering it took 25 minutes on a 20 meg connection21:19
luisbgI'm spanish as a lot of this planet21:19
luisbgit's the 3rd biggest language AFAIR21:19
slangasekdavmor2: er, odd, ok21:19
luisbgI don't mind not having it in alpha21:19
luisbgbut having it soon is a MUST21:19
davmor2slangasek: other than that It's okay21:19
_MMA_TheMuso: slangasek: Well I'd like to think more -es users were testing but I think its ok to go ahead as is.21:19
TheMuso_MMA_: same21:20
luisbgThere is a huge test base in Spain and spanish speaking countries21:20
luisbgGNU/Linex, Guadalinex, Bardinux, are a few of Ubuntu Derivatives made by spanish governments21:21
luisbgthey hire companies to make local themed or local purpose derivatives21:21
slangasekif you think not having all of the -es langpack packages being installable from the disk will break the ability to get good feedback, I'm happy to reroll21:21
_MMA_luisbg: I'd like to hear from them about Ubuntu Studio then. :P FOr now, I think its fine.21:21
slangasekotherwise, none of the images have been respun for this particular issue, and I think the fallback is just that it has to go to the network to grab the ispanish package...21:22
luisbgslangasek, in a test build with possible installation from network21:22
luisbgit's ok21:22
TheMusoGah still syncing  here.21:23
zdzichuBGtjaalton: xf86-video-intel-2.2.121:29
zdzichuBGtjaalton: was released, with fix for https://bugs.launchpad.net/xserver-xorg-video-intel/+bug/178505 ;  could you please package it?21:30
ubotuLaunchpad bug 178505 in xserver-xorg-video-intel "TV-out missing from xrandr output on Thinkpad z61t with Intel" [Medium,Confirmed]21:30
cjwatsonogra_cmpc: I don't mind, but it sounds like maybe it would be more appropriate in the artwork package21:35
tjaaltonzdzichuBG: next week, unless someone beats me to it21:46
zdzichuBGok, thanks.21:47
ogra_cmpccjwatson, hmm, probably, tes21:48
slangasekTheMuso: what news on ubuntustudio?22:07
mathiazslangasek: have you send an email to test ubuntu-server iso from iso.qa.u.c ?22:12
slangasekluisbg: la instalación de alpha-5 desktop-amd64 en español rula bien22:12
slangasekmathiaz: no, I don't typically coordinate ISO testing directly except when we're missing test coverage and I need to chase people down22:13
mathiazslangasek: hum.. So how can I be notified that a new ubuntu-server iso is availabe and should be tested ?22:14
davmor2slangasek: 64 server is tested there's only 32bit left server wise22:14
davmor2oh and sparc anyone with a spare sparc box ?22:15
slangasekmathiaz: a) monitor http://iso.qa.ubuntu.com/qatracker/build/all, b) ask stgraber who normally coordinates this, c) lurk on #ubuntu-testing around milestone time?22:15
mathiazslangasek: ok. I'll keep that in mind.22:15
slangasekmathiaz: I'd be happy to have a better process for notifying testers of new ISOs, that means fewer tests I need to do myself :)22:15
davmor2slangasek: iirc at wednesday's meet there was talk of automating this process.  Also anyone who has signed up to do tests can get emailed as a reminder that the images are up and need testing (unfortunately this time round stgraber broke it :) )22:19
slangasekdavmor2: heh, that would be a welcome improvement22:19
mathiazwell - that's what I was waiting for.22:20
davmor2slangasek: The email thingy has been in the tracker for a while but it needs to be switched on from the profile page.22:20
mathiazUsually I start testing the ubuntu-server isos when I receive the email from the iso tracker22:20
davmor2mathiaz 32bit and sparc need testing.22:21
slangasekI'm actually already testing 32bit22:21
slangasek(next wishlist bug: some way to indicate what tests are in progress...)22:23
davmor2slangasek: on the tracker test in progress22:23
=== emgent is now known as carlo__
=== carlo__ is now known as emgent
davmor2slangasek: if you go to the tracker page there is a link for tests in progress22:25
slangasekdavmor2: that doesn't tell me what I want to know22:26
davmor2Oh right :)22:26
slangasek"in progress" here only means "some parts have been reported as tested"22:26
davmor2ah true22:27
slangasekwhere I want to know "someone over there has started downloading image $foo, so others shouldn't duplicate that"22:27
TheMusoslangasek: COnnection kinda choppy here, only just got it synced, but am heading off for the weekend in a bit. _MMA_ is doing a test install/test now.22:29
davmor2slangasek: mind you it's not too bad to have more than one pair of eyes viewing the image to be fair.22:29
_MMA_(94% complete)22:29
mathiazdavmor2: right - but if we could assign the test to different people, we'd get more coverage.22:30
mathiazdavmor2: but that depends on the test cases.22:30
mathiazdavmor2: the one for ubuntu-server are well defined and targeted for a different configuration each time.22:31
mathiazdavmor2: I'm not interested in knowing that three people have sucessfully tested that openssh is working correctly.22:31
mathiazdavmor2: I'd rather now that openssh, mysql and postfix are all working correctly.22:31
mathiazbut for more generic test, such as boot the livecd and poke around - then having more people doing the same test case is usefull.22:33
slangasek_MMA_: 94% of the download, or the install?22:35
slangasekdavmor2: it is bad if you only have enough testers to get 1 test each, and they lose time duplicating efforts; and if you have more testers, knowing who's downloading what already helps you still distribute the tests more evenly22:36
_MMA_slangasek: Rebooting into install now.22:36
slangasek_MMA_: ok22:36
davmor2slangasek: mathiaz: both true.  On the tracker though there is a big list of who's meant to test what.  Heno swapped most of my tests around because some area's where covered really well while others sucked.  So as people sign up who is testing what can be monitored.22:39
slangasekthat would be sufficient if all the people signed up for testing were available 24h a day and were notified of new ISOs in a timely manner :)22:40
slangasekit doesn't help if you have opportunistic testers and need ad hoc coordination of testing of the untested images22:41
luisbgslangasek, nice spanish, where did you got that from?22:41
_MMA_slangasek: I grabbed and installed the 32-bit disk from 20080220. Looks good here.22:43
davmor2slangasek: I have a small rsync script which used to download 22 images.  However with the edubuntu cd's being all over the place now I need to modify it.  So quite often I test images that aren't tested :)22:44
* _MMA_ wishes he could do a 63-bit one but just cant nuke his main desktop yet.22:44
davmor2_MMA_:  64bit one what?22:46
_MMA_Ubuntu Studio22:46
stgraberA good way to know what needs testing is : http://iso.qa.ubuntu.com/qatracker/testcases22:50
stgraberSo you can make sure every install options of debian-installer or ubiquity have been tested at least once22:50
Amaranthbryce: did you see https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/175774/comments/25 ?22:54
ubotuLaunchpad bug 175774 in dell "[hardy] Enabling "Normal" effects produces badly drawn window shadows." [High,Fix committed]22:55
bryceAmaranth: hah!  :-)22:55
stgraberslangasek: Can I remove the two Edubuntu Desktop builds ? (IIRC we don't have Edubuntu Desktop but only the Add-on CD)23:03
slangasekstgraber: the ones that are disabled?  yes, they can be removed23:10
slangasekedubuntu desktop exists, but isn't in a state to include in this alpha23:11
stgraberslangasek: done. I'm now downloading Ubuntu Alternate and will do a LTSP + LVM + encrypted FS and then install the Edubuntu Add-on23:15
slangasekstgraber: ok, thanks23:16
slangasek_MMA_: umm... you know 20080220 wasn't current, right?23:17
slangasekluisbg: many years of practice. :)23:17
luisbgslangasek, how so?23:18
slangasek_MMA_: did you really test 20080220, or 20080221 which is current?23:18
_MMA_Sorry. 21. :P23:18
slangasekluisbg: I have a degree in Spanish23:18
luisbgslangasek, nice23:18
slangasek_MMA_: awesome, I'll go pull the lever then :)23:19
slangasekstgraber: the edubuntu add-on test is the only one outstanding, then; do you have an idea how long it'll take?23:21
slangasekmathiaz: what bit of the postgresql install failed for you?23:22
mathiazslangasek: postgres doesn't start23:23
mathiazslangasek: sudo -u postgres psql -l fails with unable to connect to socket23:24
mathiazslangasek: alpha4 already had the problem.23:24
mathiazslangasek: but I hadn't had time to investigate the problem.23:24
mathiazslangasek: I need to file a bug and milestone for beta at least23:24
stgraberslangasek: ltsp can take a while to install, I would expect a result in 30-40mins23:24
slangasekmathiaz: right, yes please :)23:29
slangasekstgraber: ok, thanks23:29
davmor2slangasek: Does this mean that Alpha 5 is ready for release?23:46
davmor2Right I'm off Bye23:56

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!