[00:01] kirkland: pam 1.0.1-3ubuntu3 uploaded [00:01] kirkland: now supports --remove [00:02] slangasek: yo [00:02] slangasek: is that what I should use in the postrm? [00:02] kirkland: in the prerm [00:02] kirkland: I replied to your bug; getting your mail? [00:03] slangasek: sifting through now [00:03] k [00:03] slangasek: just got back to the hotel [00:03] where are you this week that involves all this hotelling? :) [00:04] slangasek: london [00:04] oh-ho [00:04] slangasek: managers doing their thing, though i had a few work items that involved stealing cjwatson's time [00:05] slangasek: seeing as the installer is tightly guarded magic ;-) [00:05] heh [00:05] (just kidding for anyone else out there (or reading scrollback) [00:06] slangasek: okay, i'll move the call to prerm [00:06] slangasek: how did you feel about my postinst logic? [00:06] slangasek: changed drastically from the md5sum stuff you saw last night [00:07] hmm, perhaps I need to look clos [00:07] er [00:07] I don't think I re-reviewed that part yet [00:07] slangasek: i subscribed you to the bug [00:07] slangasek: let me grab the # [00:07] * slangasek nods [00:07] I have it [00:08] slangasek: i decided that the md5sum of those files depended too much on the precise wording of the pam configs you spit out [00:09] slangasek: actually, what i (selfishly) care about is whether pam_ecryptfs is in the stack or not [00:09] actually, i'm in London, so "whether pam_ecryptfs is in the stack or naught" [00:10] kirkland: ah. no, I'm not happy with that as a solution. [00:11] slangasek: i figured as much... [00:11] slangasek: okay, so lets work this to a compromise.......... [00:11] kirkland: we only want to use --force when we can *verify* that the config files currently on disk are the ones output previously by auth-client-config [00:12] slangasek: almost.... [00:12] kirkland: if they weren't output by auth-client-config, then AFAICS it's always out of scope for this package to force an overwrite of the user's config [00:13] slangasek: what about when a user has modified their own pam config, said "libpam-client, leave me alone", and then installed ecryptfs-utils [00:13] that's a bit harsh.... [00:13] then it's absolutely not the place for this package to override that [00:13] agreed. [00:13] because --force will overwrite the config with *no* prompting [00:13] see: harsh. [00:14] did you happen to see my proposed postinst from last night? http://pastebin.ubuntu.com/39257/ [00:15] if you can show that there's a case where this doesn't DTRT, that we can fix without clobbering users' config in some cases, then I'm happy to accept that [00:15] slangasek: i did, however, this involves a lot of trust that the giant blob of comments in those files will not change [00:16] but --force is only meant to be used when the calling package knows that the configs are not modified by the user [00:16] er, howso? [00:16] slangasek: let's say that we ship with an ecyrptfs-utils with a hardcoded md5sum [00:17] we only need to use --force to handle upgrades from a version that pre-dates pam-auth-config [00:17] slangasek: IIUC, that md5sum should be a default Ubuntu pam config, where a user has told libpam-client, "yes, please, manage my pam" [00:17] slangasek: so i get hardcode that md5sum for each of the 3 pam configs i care about [00:17] the md5sum is the md5sum for the legacy auth-client-config version [00:18] we're saying "if the user has the config installed by auth-client-config, we know it's safe to overwrite it without prompting" [00:18] slangasek: ah, that's where, perhaps, we're differing in approaches [00:18] slangasek: i don't care that much about those individuals [00:19] slangasek: if someone told libpam-client, "no, please, don't mess with my pam", and they had auth-client-config'd pam_ecryptfs, they still have what I need in the stack [00:19] sorry, what's "libpam-client" here? [00:19] libpam-runtime? [00:19] slangasek: libpam-runtime [00:19] ok [00:19] slangasek: pam-client-update/pam-auth-update [00:20] :-P [00:20] pam-client-config [00:20] asdfasdhqw[eriqwlefnsz;ldhqw;elrnaw;eohjqw;erihjqwelih [00:20] sorry [00:20] if the user said "no, don't mess with my pam" (which is the default) and they had auth-client-config'd pam_ecryptfs, then they have a set of files that ecryptfs-utils, but not libpam-runtime itself, can detect it's safe to overwrite [00:20] bye guys [00:21] if the user said "no, don't mess with my pam" and they /don't/ have pam_ecryptfs in their config, then it's not appropriate for pam_ecryptfs to force itself in there by ignoring the admin's choice [00:22] if the user /didn't/ say "no, don't mess with my pam" then --force is a no-op [00:22] slangasek: okay, that seems fair [00:22] slangasek: what state of a file, then, should I md5sum as the "--force"able state? [00:22] hello guys [00:23] and, in the case that they do have a config populated by auth-client-config, we want to upgrade it to the new system so that they get the full n-way module integration going forward [00:23] slangasek: i drew a flowchart on the whiteboard today, i wish i would have photographed........ [00:23] kirkland: if all of /etc/pam.d/common-{auth,password,session} match what auth-client-config spits out for ecryptfs by default, and /etc/pam.d/common-account matches what libpam-runtime shipped in hardy [00:23] slangasek: absolutely, the user will have to upgrade libpam-runtime first [00:24] that's the only case I see where we can say: yes, we know what's in these configs and the user didn't touch them, so we can overwrite without having to ask the user for permission [00:24] slangasek: that user will have to tell libpam-runtime "yes", or "no" [00:25] kirkland: I would encourage you to test the proposed postinst in http://pastebin.ubuntu.com/39257/ and see if you can find a case where it doesn't do what's expected :) [00:26] slangasek: okay, so the case for --force is when /etc/pam.d/common-[auth|password|session] is that which was configured by auth-client-config precisely (as defined by md5sum) [00:26] * slangasek nods [00:26] which implies that the answer to libpam-client was "NooOOOOOOO" [00:26] if it doesn't match precisely, we really can't know what it is that we'd be overwriting [00:26] right [00:26] slangasek: okay, gimme a few to fan that out over a few kvms [00:26] ok :) [00:27] slangasek: you around for a few? [00:27] yep [00:27] slangasek: i'd love to kill this one and have a good nights sleep :-) [00:27] * slangasek grins [00:27] * kirkland makes himself busy [00:27] you mean you won't be like me, sitting at your computer for hours at a time and giddily flipping the switches on pam to watch what happens? [00:31] kirkland, what are you trying to kill? [00:31] (I'm currently killing the ada issues in hardy, but I can take a break from it) === superm1|away is now known as superm1 === asac_ is now known as asac === Kopfgeldjaeger is now known as Kopfi|offline [00:50] NCommander: thanks for the offer [00:50] * NCommander waits for the but [00:51] NCommander: i think i've convinced myself yet again that slangasek is the brilliance he expects him to be [00:51] s/he/i/ [00:52] s/brilliance/eerie dpkg state machine/ [00:52] slangasek's brilliance can't be measured in rational numbers [00:53] if it were really brilliance, it shouldn't have taken me 5 years to get from concept to implementation :P [00:53] My brilliance is entirely imaginary, too. [00:53] slangasek, of what specifically? === superm1 is now known as superm1|away [00:53] NCommander: pam-auth-update [00:54] it could be worse [00:54] (augh, I almost called it pam-auth-config again, I really /did/ name it wrong) [00:54] You could have ended up coding something like RPM/Vista [00:54] slangasek: okay, the trick now is finding a kvm of mine that hasn't been maligned yet with curse-ed intrepid updates, with which to take pristine md5sums :-) [00:54] kirkland: hmm, I thought we already knew what the md5sums were/ [00:55] * slangasek gesticulates in the direction of his pastebin url [00:56] slangasek: i had them, a pastebin from last night, if that is to be trusted [00:56] slangasek: i want to confirm that [00:56] ok [00:57] slangasek: for better or worse, i've upgraded almost everything of mine to your new magic [00:57] * slangasek grins [00:57] * NCommander roots for slangasek's new magic [00:59] slangasek: ha ha.... fortunately, i have 40G of virtual machines :-) [01:00] lol [01:09] slangasek: okay, tell me what you think of http://pastebin.ubuntu.com/39565/ [01:12] kirkland: I think you should also guard against the possibility that /etc/pam.d/common-account has been modified; and I think you do need to have a version check; and I think my code was prettier ;) [01:12] slangasek: indulge me.... what does the version check gain us? [01:13] kirkland: but the only one of these that I think is a blocker is the version check - we should skip this md5sum checks entirely if we aren't upgrading from a known non-pam-auth-update version [01:13] kirkland: it allows for the possibility that, after upgrade, a user has deliberately re-installed the old config files [01:15] kirkland: I have honestly had users do this sort of thing before, and become irate that on every package update, I would overwrite their config anew :-) [01:16] slangasek: fair enough. for whatever reason, i decided to write this from scratch before just plugging and chugging with your code as pasted http://pastebin.ubuntu.com/39257/ [01:17] kirkland: to understand it bette? :) [01:17] better [01:20] wow, /me notes the similarity in at least the two comment statements [01:22] well, you wrote them both, AFAIK ;) [01:24] slangasek: in that case, both of me agree [01:24] hehe [01:24] excellent! [01:28] slangasek: https://bugs.launchpad.net/ubuntu/+source/ecryptfs-utils/+bug/259915 [01:28] Launchpad bug 259915 in ecryptfs-utils "ecryptfs should use pam-auth-update rather than auth-client-config" [High,In progress] [01:28] slangasek: i'm satisfied with the testing of that patch in my scenarios [01:31] kirkland: rockin', let me roll that up [01:31] slangasek: does "roll that up" mean "sponsor it"? if so, I'd be delighted.... [01:31] yes [01:31] slangasek: rocking [01:57] slangasek: i didn't see an upload yet.... did you have a problem with the patch? or just more pressing issues? [01:58] slangasek: i was just about to call it a night [01:58] kirkland: samba was ahead of it in my outbound queue; it's uploaded now [01:58] \o/ [01:58] slangasek: nice, thx [02:18] kirkland: hmm, FTBFS on lpia? [02:21] slangasek: typical [02:22] slangasek: have never investigated [02:22] ok [02:22] slangasek: ie, not a regression [02:22] it appears to be a regression relative to version 50-1ubuntu1, the last one that built [02:22] but not a high-priority regression, clearly :) [04:29] So, is it safe to upgrade now? Is pam gonna bork me hard like it did last night ;) [04:29] * bytor4232 never even got a phone call [04:29] No flowers, no nothin [04:29] it's never save, until it's released. [04:29] And #ubuntu+1 is likely to contain answers. :) [04:29] cool [05:36] bug #163156 can be fixed, according to upstream and System76, with new versions of libusb and libfprint. is there any chance of that bug being closed in intrepid, or will it have to wait for jiggy jaguar to support that hardware? [05:36] Launchpad bug 163156 in linux-source-2.6.22 "UPEK TouchStrip 147e:2016 not supported at all" [Undecided,Won't fix] https://launchpad.net/bugs/163156 [05:48] maco: note the proximity of feature freeze. [05:48] "jiggy jaguar" Oh noes, I hope not. [05:49] jivin jython? [05:50] is a jython an animal? i thought it was just a made up name for the language [05:50] it is [05:52] crimsun: you already talked to me in person. i wanted to see if others agreed with you that it's "oh noes" or if someone who knew more about how libusb works could say if core functionality was affected or not [05:53] maco: I suspect it's not a matter of "oh noes" but lack of adequate testing [06:01] crimsun: Can I help with the issue re pulse by default with a fallback to dmix for alsa-lib? [06:01] In terms of patching/implementing? [06:06] TheMuso: crimsun says, "tell him that sjoerd's patch is in alsa-plugins 1.0.17 and newer and can be backported to the current one. and tell him to make sure he checks that the pulseaudio patch that's in debian/patches/ is in place that checks for the environment variable" [06:07] maco: Thanks. [06:07] ok now crimsun and i go to bed [06:07] night night guys [06:25] Yay I love ambiguity in replies in the third person. :) [08:01] good morning [08:14] morning dholbach [08:14] hi raphink [08:14] ;) [08:14] what's up? [08:17] just got through most of my mails and will dive into the sponsoring queue now! [08:17] as you all should :) [08:17] hehe :s [08:18] some time soon, I might get a job with a day for Ubuntu every week [11:04] xserver-xorg-hardcore [11:37] ion_: surely you meant xserver-xxxorg-hardcore [12:22] good morning === Kopfi|offline is now known as Kopfgeldjaeger [13:32] slangasek: any time I run pam-auth-update, it duplicates the Additional block in /etc/pam.d/common-* [13:32] slangasek: though interestingly it does not duplicate pam_unix in common-password [13:34] slangasek: this is bad because the duplicated-bits files cause PAM clients (at least gnome-screensaver and login) to segfault ... [13:38] aha [13:38] Name: libpam-runtime/profiles [13:38] Template: libpam-runtime/profiles [13:38] Value: unix, smbpasswd-migrate, ecryptfs-utils, smbpasswd-migrate, ecryptfs-utils [14:43] Is anyone up for archive-admin stuff today? [14:44] persia: i think Hobbsee is [14:44] Hobbsee: Can you NEW stuff? [14:44] but, today is pitti's turn and he is on vacation [14:44] nxvl: Hence me not poking anyone specifically :) [14:44] persia: yes. not at this time of night while i'm half asleep, though... [14:45] unless it's really simple [14:45] Hobbsee: It's really simple. ubuntu-mid-default-settings. Maybe 20 lines of code, total. [14:45] Hobbsee: @ .au you have the same tz as japan? [14:46] it's 11.45pm now here [14:46] nxvl: No, it's even later for her [14:46] not sure what japan is [14:46] yeah, one hour [14:46] * nxvl love's gnome clock [14:46] * Hobbsee looks [14:50] persia: accepted. [14:50] * persia adds another 100g to the chocolate debt tally [14:50] hmmm. i wonder if that's a launchpad bug. [14:50] OK: ubuntu-mid-default-settings(universe/(unchanged)) [14:52] yes, apparently it is. [14:52] This is a new and different LP bug, or a previous one? [14:54] apparently it's not a bug. [14:56] persia: your binaries are served. [14:56] well, they're still pending an archive run. but served enough. [14:57] Hobbsee: Thank you ever so much. Now I can go about filing the appropriate removal requests :) [14:57] :) [14:57] can't do them, i'm afriad [14:57] No worries. I'm not blocked on removals, as this means I can ignore the crud until later. [14:57] true [15:02] hi. is it possible to do bindtextdomain() with gettext without all those subdirectories ("en_US/LC_MESSAGES/")? [15:24] I am a bit annoyed : I am creating a new package (a python module) which is called pywebkitgtk (pythons bindings for the webkitgtk). According to my understanding of the python policy, the name of the package should fe python-modulename... which in my case will be python-webkit [15:24] isn't it a bit problematic ? [15:25] (I mean the loss of the gtk stuff) ? [15:25] python-webkitgtk would be appropriate i guess [15:26] like we have python-gtk2 for upstreams pygtk [15:26] (not sure why that still carries the 2 though) [15:26] but the thing is in that case gtk2 is the module name I think [15:26] and thus is respects the policy [15:27] i dont think upstream calls it pygtk2 [15:27] ok [15:27] but if your modules is called pywebkitgtk i'd call the package python-webkitgtk [15:27] *module [15:28] ogra the project is called pywebkitgtk [15:28] but the module is called webkit [15:28] well, but its not pywebkitqt :) [15:28] :) [15:28] ok [15:29] so I name it python-webkitgtk thus [15:29] thanks [15:29] james_w: ^ do you know anything about that? [15:30] huats: can you give the link to the policy where you saw this? [15:30] sure [15:30] I've seen pretty much all the alternatives I think [15:30] but let me first find it again :) [15:31] i.e. using the upstream project name, the module name, or something else that makes more sense [15:32] james_w: http://www.debian.org/doc/packaging-manuals/python-policy/ch-module_packages.html [15:32] 2.2 [15:32] Public modules should be packaged with a name of python-foo, where foo is the name of the module [15:32] yeah [15:32] (and what I explained is my understanding of that :)) [15:32] 'should' != 'must' though [15:32] ok [15:33] it's odd they called the module just "webkit" [15:33] yeah [15:33] couldn't there be pure bindings to webkit with that name? [15:33] especially if thats gtk specific [15:33] yeah [15:33] i'll report that upstream (I ve very good contacts with them) [15:33] huats: yeah, ask them [15:34] I'd advise going against policy if there's any chance that there will be a pure python-webkit, to save headaches later [15:35] and you will definately make Riddell happy in case there comes a pywebkitqt at some point [15:36] OK [15:36] so I change my package name to python-webkitgtk [15:37] and I ask upstream explaining them everything [15:37] (and pushing them toward changing the name of the module...) [15:37] :) [15:40] thanks ogra, james_w and dholbach [15:40] de rien [15:40] have a great weekend everybody! :) [15:40] :) === Kopfgeldjaeger is now known as Kopfi|offline [15:48] ogra: there is [15:48] and it uses namespaces [15:55] kees: there? [15:59] Koon: slangasek and kirkland were working on some pam update changes [16:00] mathiaz: what's up? [16:00] Koon: you may wanna ask them about how to handle pam configuration update wrt to likewise-open [16:00] mathiaz: yes, I suppose it's related. But dendrobates is handling the likewise-open update right now. I'm busy with tomcat6 and wbem [16:01] I only did a quick test on my AD test infra === Kopfi|offline is now known as Kopfgeldjaeger [16:02] kirkland: likewise-open can't join a domain, complains about the current PAM files layout. Apparently a change in /etc/pam.d in recent pam updates confuses it [16:03] Koon: that's certainly possible... what does it use? auth-client-config? [16:03] kirkland: it uses its own let-me-whack-your-pam.d functions [16:03] Koon: oh, that's about as bad as it gets then [16:04] Koon: grab the source for ecryptfs-utils [16:04] Koon: see the bits that involve pam-auth-update [16:04] Koon: requires libpam-runtime >= something [16:04] Koon: you'll need a config file for it [16:04] Koon: and to conditionally run pam-auth-update [16:05] kirkland: might not have time to do that -- but dendrobates will be interested by the pointers === superm1|away is now known as superm1 [16:10] asac: good morning :) [16:11] kees: fedora and Redhat. hahahaha [16:11] kees: good morning ... bb in 40min [16:11] ;) [16:12] kees: I'm sure they had a good laugh at us. ;) [16:12] dendrobates: lol [16:13] dendrobates: to name it 'openssh-blacklist'. it's so confusing, it's funny. :) [16:13] kees: I saw that and wondered if I'd misunderstood the vulnerability, then I realised that they just named it badly. [16:26] we have high confidence [16:26] that the intruder was not able to capture the passphrase used to secure [16:26] the Fedora package signing key [16:26] haha [16:27] LOL [16:28] if you ask me... if anybody cares about security, that key needs to be changed regardles, lol [16:29] mkrufky: guess what.. that's what they're doing [16:29] slangasek: do you know what http://cdimage.ubuntu.com/stable is, apart from a Windows executable? === superm1 is now known as superm1|away [17:00] is intrepid safe enough to use for a developer who isnt afraid to get his hands dirty (but would prefer not) [17:00] ? [17:01] mkrufky, the safe answer is no [17:04] hehe [17:04] ok [17:05] thanks, cody-somerville [17:06] * calc still isn't using intrepid either :) [17:06] mkrufky: That said, lots of people do use it every day, just be aware that it *will* break, so you may need to fix it. [17:06] i just build in a intrepid chroot [17:06] slangasek: Riddell: cjwatson: hi arch admins... kees pointed me to you, to "deNEW" the package update-motd from https://launchpad.net/ubuntu/intrepid/+queue [17:07] i'll probably upgrade to intrepid on Oct 2 (Beta) [17:07] hmm, well then maybe its a better idea for me to just install hardy on this new server [17:07] (server for home use -- not for business) [17:08] kirkland: looking [17:09] kirkland: accepted [17:09] Riddell: thank you, sir! [17:19] mvo: http://www.kubuntu.org/~jriddell/tmp/update-notifier-kde_0.1_all.deb should you want to take a keek === cjwatson_ is now known as cjwatson [17:23] kirkland, "update-motd - This package contains a generic framework for regularly running" [17:24] running what ? [17:24] doh [17:24] a sports package :) [17:25] ogra: whatever you put in /etc/update-motd.d/, concatenated together and written to your motd [17:25] ogra: is the control file broken? [17:25] kirkland: first line of description is the short description, that should be fixed [17:25] looks like [17:25] Also overly wordy: doesn't fit the guideline $(packagename) is a $(short-description) [17:25] long descriptin starts on the next line [17:25] Riddell: that should be a one liner, then? [17:25] * \sh has a new toy...acer netbook, with 120GB hd...looks very neat...now it just needs ubuntu on it, and not linspire [17:25] kirkland: yes [17:26] yeah, its trivial and not harmful [17:26] Riddell: fudge, sorry, i'll fix immediately :-) [17:26] kirkland: Also doesn't need to be a full sentence: just a predicate clause [17:27] persia: ogra: Riddell: hows "Description: Modular framework for running scripts to generate the /etc/motd" [17:27] <80 chars [17:27] works for me [17:27] Modular framework for to generate the message of the day dynamically ? [17:27] -for [17:28] ogra: that's better [17:28] kirkland: Yes, that would be the typical pattern [17:29] so did anyone make the consideration that I might actually not like running f-spot? [17:29] * persia wonders how this got through REVU, and decides that the REVUers need training again [17:29] like, back in the day, GNOME was like, "Hey you just plugged in a card containing pictures. Want to import photos?" [17:29] these days it just opens f-spot. [17:30] yea, i have to constantly kill that stupid app [17:30] so I went to system->preferences->removable media and unchecked everything, particularly "Import digital photographs when connected" for cameras [17:30] it still does it. [17:30] bluefoxicy: changelogs are the key to understanding history. [17:30] i just want to cp -a the files over [17:30] so I deleted `which f-spot` === mkrufky is now known as mkrufky-lunch [17:30] Is there a bug for this? [17:30] it says it's attempting to open the card for a few seconds now, but doesn't run fspot [17:30] what the hell? [17:30] i don't know [17:31] bluefoxicy: hmm it shouldn't be importing photos if you turn it off, i forgot to just go in and do that myself [17:31] bluefoxicy: if it doesn't care its not checked then that would be a bug [17:31] calc: can you try to reproduce? [17:31] bluefoxicy: yea i can go get my camera card and see what it does [17:31] istr my camera card not doing it but a friends starting f-spot [17:31] istr? [17:32] i seem to recall [17:32] ok got my card going to see what it does [17:33] * ogra doesnt think gnome-volume-manager is particulary useful nowadays since gvfs handles all that [17:33] <_MMA_> bluefoxicy: You also have to disable it through Nautilus now. [17:33] ogra: patch attached to https://bugs.launchpad.net/ubuntu/+bug/260441 [17:33] Launchpad bug 260441 in ubuntu "description line too long for update-motd" [Low,In progress] [17:33] right, what _MMA_ said [17:33] _MMA_: okay, wait, there's TWO places to disable the same feature? [17:33] ogra: any chance I could talk you into sponsoring? [17:33] _MMA_: so much for gnome being easy to use, lol [17:33] ogra: since it was your complaint? :-) :-) [17:33] and they're not linked [17:33] <_MMA_> Edit->Prefs-> Media tab [17:33] * calc thinks that would be HIG non-compliant [17:33] I'm going to have to call that a bug [17:33] its a half done transition [17:34] between gnome-vfs and gvfs [17:34] lol [17:34] I love the options for software [17:34] "Ask what to do" "Open autorun prompt" (the second, I assume, asks you what to do... since it's called a "prompt") [17:34] <\sh> ogra: are you coming to froscon? [17:34] \sh, froscon ? [17:35] <\sh> ogra: free and opensource conference, st. augustin ;) [17:35] kirkland, looking, i will have to wait until i can grab the source pkg off the archive (i'm lazy) but will do the sponsoring then [17:36] ogra: cool dude, thanks. [17:36] <\sh> ogra: it starts tomorrow morning ;) so I wonder if you want to visit your old area ;) and have a nice o'reilly beer ,) [17:37] \sh, hmm i saw the blog post from lydia ... thogh i thought st. augustin is somewhere in austria :P [17:37] <\sh> ogra: lol...between bonn and cgn yeah austria rotfl [17:38] 7win 7 [17:38] gah... [17:39] more fun stuff on my system [17:40] soren: are you playing a slot machine? [17:40] when i plugged my sd card in it froze my system to where sysrq couldn't even reboot it [17:40] i had to hard reset it with power [17:40] kirkland: It feels that way. [17:42] cjwatson: hrm, looks like a wayward copy of wubi... [17:43] cjwatson: duplication when running pam-auth-update> which version do you have installed? This should be fixed in the latest [17:43] slangasek: You never got back to me on the Nagios promotion thing in Hardy, did you? [17:44] I sent it on the Friday where you were probably leaving for Debconf, so timing is not on my side. [17:44] (unless, that is, I broke it again) [17:45] soren: no, I had not; though today looks promising, if I don't find that I have another pam fire I didn't notice from yesterday [17:47] mvo: is there a way to do something like markReinstall() with python-apt? [17:47] slangasek: Cool, thanks. === Kopfgeldjaeger is now known as Kopfi|offline [17:52] tseliot: yes, give me a sec [17:52] slangasek: 1.0.1-3ubuntu3 [17:53] slangasek: I filed a bug about it with the details [17:53] cjwatson: ok; I know which commit broke it then, I'll have a fixed package away shortly [17:53] slangasek: great, thanks [17:56] tseliot: cache._depcache.SetReInstall(pkg._pkg) [17:57] mvo: does cache stand for apt_pkg.GetCache or apt.Cache ? [17:57] persia, I am posting about the key team policy, I got no more responses w.r.t. to needing changes, so I'm calling the issue closed and resolved [18:03] tseliot: apt.Cache [18:03] mvo: ok, thanks a lot [18:03] cheers [18:08] cjwatson: hmm, when this /first/ happened to you, what version of libpam-runtime did you have installed? I actually can't reproduce the problem, I wonder if I've already fixed it for everything but an upgrade case that I can't pinpoint [18:14] mvo: that method complains that it requires 2 arguments: self.cacheUi._depcache.SetReInstall(pkg._pkg) --> TypeError: function takes exactly 2 arguments (1 given) [18:15] slangasek: same [18:15]