[00:00] about 70% of the time i get memory its bad out of the box [00:00] directhex: most likely my system can only take up to 8GB iirc [00:00] hm, UDS is at google? that's awesome [00:00] directhex: yea [00:01] calc: sounds like you are mishandling it. [00:02] or buying really crappy memory [00:02] Mithrandir: its from Fry's (US store notiorous for bad hardware) [00:02] Mithrandir: unfortunately they are the only real hardware store around the area [00:02] i used to work at a pc retailer [00:02] I've heard of this thing called the internet where you can order stuff from other places than nearby stores. [00:02] you ever heard of that? [00:02] boxed hard disks have a similar weight to a soccer ball [00:02] ;-P [00:02] generally if someone returns something they just stick it back on the shelf to sell again [00:03] calc, yea, we did that [00:03] Mithrandir: yes but then "you can't just walk to the closest shop and pick up a replacement? [00:03] with said football hard disks [00:03] that part doesn't work (bad copy/paste caused early return) [00:03] you kicked them before putting them back ? [00:03] calc: internet is even closer than the closest shop [00:03] but yes i have fairly good results from buying from newegg in the US [00:03] *shrug* [00:04] Mithrandir: but takes several days for shipping unless you pay large amount of money for next day (in the US) [00:04] in the end not much faster than just RMAing the memory which is what i am going to do [00:06] * calc thinks he'll go get food while waiting on the rest of his system to pass testing [00:14] Hi, it looks like there's a dependency issue between the mozilla and seamonkey packages in hardy-{updates,security} [00:14] Should I file a bug against both packages? [00:15] Caesar: yes, please. [00:15] Caesar: please mark it a security issue, and then flag it as public. [00:16] kees: okay [00:23] kees: https://bugs.launchpad.net/bugs/300519 [00:23] Error: This bug is private [00:23] Drat [00:23] Now it's public [00:23] Caesar: cool, thanks! [00:23] np [00:51] Salve Caesar! Morituri te salutamus! [00:53] * Caesar blinks [00:54] Caesar non Latine dictus? [00:55] am hitting https://bugs.launchpad.net/ubuntu/+bug/300162 [00:55] Launchpad bug 300162 in ubuntu "bad security update (dup-of: 300247)" [Undecided,New] [00:55] Launchpad bug 300247 in hplip "apt-get upgrade (hplip...) fails on gutsy" [Undecided,Triaged] [00:55] sounds like a bad push [00:57] ok second bug answers my question [01:07] mneptok: no [01:49] kees, poke [01:50] kees, you can't have NX without PAE === ubott2 is now known as ubottu [02:52] NCommander: and if you want PIE you have to compile with --finished_dinner=YES [02:52] lol [02:53] * mneptok is in the Clean Plate Club! [02:53] hahaa === spm changed the topic of #ubuntu-devel to: LP down at ~0315UTC for ~ 30-60mins| Archive: frozen for jaunty alpha 1, MoM running | Ubuntu 8.10 released! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs === spm changed the topic of #ubuntu-devel to: Archive: frozen for jaunty alpha 1, MoM running | Ubuntu 8.10 released! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs [05:01] I have been looking at harvest and found a package called moodbar has a patch from fedora for it. [05:01] How do I figure out what the patch does? [05:02] Here: http://cvs.fedoraproject.org/viewvc/devel/moodbar/moodbar-0.1.2-glib.patch?view=markup [05:02] I would like to learn to patch packages but am sure in the changelog you want to state why the package is being patched. [05:02] Or do I just say something like: "patched with fedora patch"? [06:30] does anybody have time for a sru? Bug 263779 [06:30] Launchpad bug 263779 in gnome-settings-daemon "Evince takes over global shortcut keys" [Undecided,In progress] https://launchpad.net/bugs/263779 [06:32] NCommander: right, I know about NX needing PAE, but I encountered non-NX _with_ PAE. finally tracked it to a bios setting. !! [06:32] o_O; [06:32] yay for crapply written bios [06:34] NCommander: yeah. I'm deeply unpleased that a) it is even _controllable_ by the BIOS since the CPU feature downgrades gracefully for OSes that don't use it, and b) Dell ships with the feature disabled. grrr [06:34] kees, VTx is the same way [06:35] NCommander: yeah. I can _almost_ understand why that's in the BIOS, but not really. [06:35] Well, its because of PAE [06:35] PAE can screw up some old operating systems, but it should be disabled out of the box :-/ [06:36] (i.e., the kernel must write the necessary control information for PAE to activate) [06:38] good morning [06:46] Any X heads here? Why is the default RenderAccel method for radeon still XAA on all cards, but the driver will spit a warning to Xorg.0.log saying that they shouldn't be using XAA? [06:47] RAOF: Because it was causing instability, IIRC. It will probably be changed in Jaunty. === doko_ is now known as doko === tkamppeter_ is now known as tkamppeter [07:44] slangasek, if you would like to reclaim some space on cdimages.ubuntu.com, you can remove the old mythbuntu/{daily,hardy,releases} directories. we've got those images on mirrors now, so the only relevant one is daily-live/ === dholbach_ is now known as dholbach [08:04] Good morning [08:15] tkamppeter: the patch which you mailed me for bug 294671 isn't for cups; is it for gutenprint? where does the patch come from? [08:15] Launchpad bug 294671 in gutenprint "Driver reports "ColorSpace is not supported" for Canon iP4500" [Undecided,Fix released] https://launchpad.net/bugs/294671 [08:26] it's friday morning. a weekend of exciting possibilities awaits for developers new and old. so... who feels like kick-starting the mono packaging transition, and uploading my 2.0.1-0ubuntu1 package? pitti? [08:27] directhex, once mono is uploaded, I'll do the universe sponsorings [08:28] directhex, i think 9.04 alpha 1 will probably need to get out the door first... [08:28] NCommander, excellent. $DEITY-willing, coordinating efforts with pkg-mono will mean things can go into experimental and be synced as per usual, in many cases [08:28] superm1, what exactly does alpha 1 aim to freeze? [08:29] directhex, you still need people to ack the sync [08:29] superm1, alpha freeze 1 is main only ATM [08:29] NCommander, aye. just sayin' syncs are best. did you see my handy dandy transition tracking page? [08:29] xxNope [08:29] http://wiki.debian.org/Teams/DebianMonoGroup/Mono20TransitionTODO [08:30] Youch [08:30] Why didn't you use an LP bug to track the status as one big group? [08:32] NCommander, yeah and last i checked mono itself was in main [08:32] directhex, get a mono freeze exception [08:32] (or wait for alpha 1 to pass) [08:32] hey mvo [08:33] hey NCommander [08:33] NCommander, how much of a wait? the idea of going with a 0ubuntu1 rather than waiting for -1 to finish debian NEW was to let people start on the transition ASAP [08:33] * NCommander is checking the release schedule [08:34] ... [08:34] * NCommander thinks it would help if I checked on wiki.ubuntu.com vs. wiki.debian.org [08:34] mvo, when do we leave alpha 1 freeze specifically? [08:35] directhex, we might be out of freeze now that I take a closer look, the schedule said November 20th [08:35] Its the 21st now [08:36] it's not clear to me what the purpose of alpha 1 freeze is - given the lenny freeze in debian, it's not as if there's any value in the automatic sid importing, this time around. and last i checked the wiki didn' explain that freeze [08:36] NCommander: don't use one bug listing lot of tasks to track transitions, that mailspam the subscribers to all the task including the ones which have been closed [08:36] fair enough [08:37] seb128, I see jaunty CD images [08:37] Does that mean we're out of freeze? [08:38] the topic doesn't indicate so and I just woke up so dunno what happened during the night [08:38] better to ask to slangasek [08:38] and the amd64 image is oversized [08:38] by less than a megabyte [08:38] \o/ [08:38] NCommander, well, the amd64 image shrinks by several meg once this transition is done! [08:39] NCommander: normally when its released, I don't except it will be much later (but #ubuntu-release will know for sure) [08:40] mvo, would you do me a favor, and once we leave freeze sponsor directhex? (and my kde4libs upload (yes, its a recurring cycle of breakage, but I testbuilt on armel with the help of distcc :-)) [08:41] NCommander: sure [08:42] directhex, there you go, mono sponsor :-P [08:42] hah. thanks to you both! [08:44] directhex: sure, where is it? [08:45] https://bugs.launchpad.net/ubuntu/+source/mono/+bug/300133 [08:45] Ubuntu bug 300133 in mono "Mono 2.0.1 package (triggers major packaging transition, please read in full)" [Undecided,New] [08:45] directhex: erm, hang on, has alpha-1 been released? if not, uploading it now would be bad [08:45] pitti, are we still frozen? [08:45] pitti, I see CD images [08:45] see /topic [08:45] pitti: see scrollback ;-) [08:45] * NCommander is just wondering if the topic is lagging [08:46] seb128: I didn't see anything which would indicate that the topic is wrong and a1 is released [08:46] pitti: ok, we were just wondering if it was any likely that new images will be rolled [08:46] freeze going over friday are no fun [08:47] * NCommander can't figure out why we're freezing for alpha 1 [08:47] seb128: don't know, I wasn't tracking a1 at all this time [08:47] ok, me neither [08:47] what does alpha 1 achieve? [08:47] NCommander: it's a soft freeze [08:47] I guess we need to ping slangasek to know [08:48] NCommander: we can upload stuff that doesn't break consistency [08:48] oh [08:48] pitti: the mono transition is a transition [08:48] while kde4libs is simply lets unbreak armel :-) [08:48] pitti: so it will break installability [08:48] seb128: exactly [08:48] pitti, after the last armel FTBFS, I actually wired up a cross-compiler this time. 12 hours later, I have confirmed my fix works [08:49] pitti: btw I didn't do syncs due to the freeze but I can do those after the freeze if you want [08:49] seb128: I'll have a look at them, no problem to do universe ones, and trivial main ones [08:49] but yes, an autosync run is too dangerous right now [08:50] pitti: ok, just letting you know that I can do that if you are busy, I've nothing special to do today [08:52] it would probably be really valuable to add a marker for main/universe to that transition page, wouldn' it [08:53] directhex, yes [08:53] doko, I have the debdiff uploaded to #299906 if you care to sponsor [08:54] er, Bug #299906 [08:54] Launchpad bug 299906 in kde4libs "armel build failure (package not yet in the archive)" [High,Triaged] https://launchpad.net/bugs/299906 [08:54] THere we go [08:55] directhex, for these transitions, what specifically needs to be done? [08:55] directhex, new upstreams, or just a rebuild bump? [08:56] NCommander, best case: reduce build-deps a little, and add a configure flag/env var [08:56] This sounds this transition is going to suck [08:57] Speaking of sucky transitions, someone needs to look at ghc6 [08:57] haskelltastic! [08:57] s/ask//g [08:57] :-) [08:58] http://wiki.debian.org/Teams/DebianMonoGroup/Mono20Transition#head-67c13a005dab7f510b0fd1ee8db7a30689e89669 is a brief guide i wrote on how to transition a package. i should probably put the simple option before the complicated option [08:59] truth be told, i did the complicated one due to lack of CDBS knowledge (i happened to pick a CDBS package for my example) [08:59] speaking of breakage, can one of the archive admins handle #298785 [09:00] bug #298785 [09:00] Launchpad bug 298785 in intrepid-backports "please backport lirc and mythbuntu-lirc-generator to Ubuntu intrepid" [Wishlist,In progress] https://launchpad.net/bugs/298785 [09:01] NCommander: any hurry that you need to ping on irc about it? [09:01] What should i do re: http://launchpadlibrarian.net/19829294/buildlog_ubuntu-jaunty-armel.wordnet_1%3A3.0-13_FAILEDTOBUILD.txt.gz [09:02] NCommander: usually when people lag a bit on archive admin work that's because they are busy so there is no need to ping on irc, just wait a few days rather ;-) [09:02] seb128, can't I just call fork() and get my own archive admin :-)? [09:02] * NCommander hears a gunshot [09:02] NCommander: nice try ;-) [09:02] stefanlsd, is this package in main? [09:03] NCommander: you don't have enough on your plates yet? ;-) [09:03] seb128, I'm determined for beta 1 that main is FTBFS free on powerpc and armel, with close to zero on the non-installability accounts [09:03] I just need to know where to place this in the queue [09:04] * NCommander notes that KDE is currently the priority since GNOME built [09:06] NCommander: universe [09:07] okay. i've still got to do libs, but: [09:07] stefanlsd, its sorta a case of patches welcome :-/. We're focusing on main ATM (KDE is a load of unhappiness). Is there a pressing reason to handle it now? [09:07] mono stack, 4/9 packages in main [09:08] NCommander: nope, other than i would just like to learn about ftbfs and how to fix it. not pressing in terms of release at all [09:08] cli apps, 5/42 packages in main [09:09] stefanlsd, for that build failure, you likely need to relibtoolize or autoreconf of the package (it likely will also FTBFS on amd64/i386 if you tried to build it there) [09:14] NCommander: do u know if there is any documenentation on doing that, or how we approach FTBFS in general? [09:14] NCommander: done [09:15] stefanlsd, In case of FTBFS, poke NCommander ;-) [09:15] * NCommander runs [09:17] hehe. NCommander poke! mm. but i think its great that your a FTBFS expert, but if you can get something up so we can get some more ncommander expert clones, that would be great (i wanna volunteer to learn about ftbfs) [09:17] cli libs, 13/41 packages in main [09:17] * hyperair wonders if anybody has time for a sru for gnome-settings-daemon [09:17] specifically Bug 263779 [09:17] so a total of 22 packages in main are affected by the mono transition, including mono itself [09:17] Launchpad bug 263779 in gnome-settings-daemon "Evince takes over global shortcut keys" [Undecided,In progress] https://launchpad.net/bugs/263779 [09:17] stefanlsd, I'm one of the few people from Debian who is a porter more than a packager. I'm used to handling FTBFS :_) [09:22] directhex: I'd really appreciate you waiting with mono; it won't take much longer [09:22] could somebody please score up hw-detect and udev on the architectures where it has not yet been built? [09:24] hyperair: not sure if that's really worth a sru, that seems a corner case, usually people use multimedia keys for such actions, anyway subscribe the sponsor team so it's on the sponsoring list [09:24] NCommander: you don't see CD images in cdimage.u.c/releases/jaunty/. Dailies aren't releases [09:24] oh [09:25] seb128: usually people use multimedia player keys to control their media players, not a pdf reader =p [09:25] I suppose it's just hppa/ia64/sparc, I'll upload debian-installer without waiting for those [09:26] hyperair: right, which means you don't have a conflict between applications [09:26] hyperair: or the bug description is not clear [09:26] seb128: no the issue is that evince is hijacking the multimedia keys from media players [09:27] hyperair: so the description is not clear and I suggest you update it [09:27] what should i change it to? [09:27] cjwatson, on the bright side, with the new build-deps, we don't anticipate needing another packaging transition any time in the next few years (i.e. not until MS release a new version of the CLR spec which isn't built on top of a previous version, which last happened in 2006, AND mono becomes compatible enough with that new release to warrant making it the default CLR version to use). so at least 5 years away [09:27] hyperair: right now it says to use ctrl-alt-letter to get the issue, which I doubt many users do [09:28] hyperair: change to a common usecase, ie: use a multimedia key, start evince, verify the multimedia key stop working [09:28] directhex: that'll be great [09:28] directhex: though not something I'm worried about right now, I'm focused on getting CDs working [09:30] zul: ping question in bug 228693 [09:30] Launchpad bug 228693 in bacula "[SRU] bacula-director-pgsql postinstall broken" [Medium,Fix released] https://launchpad.net/bugs/228693 [09:35] ok, cancel my hw-detect/udev request, but could somebody please score up debian-installer? [09:35] pitti, the patch is for the pdftoraster filter which is in debian/local/filters/pdf-filters/filter/ (Intrepid only). [09:36] cjwatson: you mean retry? [09:36] seb128: i've just modified it. is the new description okay? [09:36] cjwatson: oh, nevermind, buildd.py used the previous version for some reason [09:37] indeed, the version shown on https://edge.launchpad.net/ubuntu/+source/debian-installer is out of date [09:37] doko, are you working on kdelibs? [09:37] NCommander: no [09:37] * NCommander is starting to at the moment, but since the last upload is still broken on armel [09:37] ok [09:37] pitti: Because it's not published. [09:38] cjwatson: done (don't count on ports, though, they are horribly clogged) [09:38] building qt4 and OO.o [09:38] and kde4libs [09:39] wooo [09:39] and sparc is down to one buildd, which is OO.o'ed [09:40] cjwatson: i386 shoudl be ok, if amd64 is urgent, we need to prod the buildd gurus to forcefully kill kde4libs or openjdk [09:40] pitti, kde4libs will only take one hour, 17 minutes to build give or take five [09:40] since it started 26 minutes ago ... [09:41] hyperair: the description is better indeed, does it mean that the keys will not work in evince if rhythmbox is running though? [09:41] tkamppeter: but rasterdsp.cc doesn't even exist in cups [09:41] seb128: yes that's right, but that's intended behaviour. [09:41] seb128: when rhythmbox is closed, the keys wll work in evince [09:42] hyperair: I would expect the intended behaviour to be that whatever application having the focus should respond [09:42] and not evince working or not depending of whether some other software is running somewhere [09:42] seb128: well, these are global shortcuts. you don't see shortcuts defined in sys/pref/keyboard shortcuts reacting to focused apps [09:42] pitti: I guess I'll just wait for kde4libs [09:42] NCommander: see, this is how porting can end up blocking alpha releases [09:43] seb128: let's say i define ctrl+alt+c to start the calculator, and some other application has some menu item which reacts to ctrl+alt+c. if i focus that application and press ctrl+alt+c would you expect a calculator to pop up, or the menu item to trigger? [09:43] hyperair: not sure that the behaviour is ideal but the patch makes sense over what is happening now, just subscribe the sponsor team, I'll try to have a look to sponsoring today [09:44] hyperair: the menu item to be selected [09:44] but that's a different question than the sru change one [09:44] alright [09:44] pitti, remove the rasterdsp.cc part, only take the part acting on pdftoraster.cc and apply it to pdftoraster.cxx (edit the patch appropriately). [09:44] let's get the obvious issue fixed first and then that can be discussed upstream later if required [09:45] tkamppeter: okay; where does this patch come from? [09:45] seb128: alright. i've subscribed ubuntu-sru and ubuntu-main-sponsors [09:45] hyperair: thank you [09:45] np === thekorn_ is now known as thekorn [09:56] pitti, the patch is based on the upstream distribution of the pdftoraster filter. For the use with CUPS I did not put the full upstream distribution into the CUPS package but only the source file of the filter itself and I patched the build system of CUPS to get the filter compiled. [09:57] tkamppeter: right, I'm interested in getting a websvn link or something, for putting into the patch as an upstream reference [09:58] what's the name of xfce's window-manager? [09:59] found it, never mind [09:59] tkamppeter: oh, ignore me, I didn't upload the cups SRU yet, so I can include this one [10:00] pitti, here is the patch on the upstream web SVN: http://svn.sourceforge.jp/view/pdftoraster/trunk/src/pdftoraster.cc?root=opfc&rev=850&r1=848&r2=850 [10:02] hi all, i experienced LP#291878 and wrote a trivial patch to the kernel to add a keyboard quirk for the affected laptop model. the patch works the way it's supposed to on my laptop and solves the issue, but so far no one else affected by the bug has responded to my patch. what's the proper way to get it included into the ubuntu kernel and also getting it upstream? [10:03] rlj: try taslking to the folks in #ubuntu-kernel [10:04] thanks [10:04] pitti, then you can SRU the following now: cpdftocps (bug 299707), pstopdf (bug 282186, and perhaps also bug 293832), pdftoraster (bug 294671 and duplicates). [10:04] Launchpad bug 299707 in cups "Impossible to print in reverse order" [Undecided,In progress] https://launchpad.net/bugs/299707 [10:04] Launchpad bug 282186 in hplip "HP Photosmart 2610 top first cm not printed" [High,Confirmed] https://launchpad.net/bugs/282186 [10:04] Launchpad bug 293832 in foomatic-db "printer prints page at wrong position, page cut" [High,Incomplete] https://launchpad.net/bugs/293832 [10:04] Launchpad bug 294671 in cups "Driver reports "ColorSpace is not supported" for Canon iP4500" [Undecided,Fix released] https://launchpad.net/bugs/294671 [10:05] pitti, note that it is not confirmed whether the current pstopdf fixes bug 293832, we must ask the poster to try the -proposed package once you have uploaded it). [10:05] Launchpad bug 293832 in foomatic-db "printer prints page at wrong position, page cut" [High,Incomplete] https://launchpad.net/bugs/293832 [10:05] tkamppeter: okay [10:06] tkamppeter: so your latest trunk commit fixes bug 282186? [10:06] Launchpad bug 282186 in hplip "HP Photosmart 2610 top first cm not printed" [High,Confirmed] https://launchpad.net/bugs/282186 [10:06] tkamppeter: if so, please set the jaunty task to "fix committed" and commit the LP # to the changelog [10:06] pitti, yes. [10:07] tkamppeter: please tell me when you committed/pushed, then I can merge this into the intrepid branch [10:09] pitti: please could you process liblink-grammar4 binaries in NEW? makes abiword installable on armel [10:09] seb128: ^ can you please? [10:10] pitti: looking [10:10] doko: could you start using requestsync to file sync requests? [10:11] doko: your bugs are the only one we need to open to know if the package is in main or universe, that's extra work for no good reason [10:11] pitti, bug number committed/pushed, bug set to "Fix Committed". [10:11] doko: requestsync set nice title which have those informations and make the work easier for everybody [10:11] tkamppeter: thanks [10:12] doko: does liblink-grammar4-java need to go to main or universe? [10:12] seb128: "makes abiword installable on armel", so I assume that should be main [10:13] doko: liblink-grammar4 is also new on armel so I was wondering if that's only that one or also the java binary [10:13] doko: I don't think abiword uses java [10:14] seb128: well, usually we add language bindings into main if they are built from the same source and the language is in main [10:14] so yes, it should go to main [10:14] doko: not really, they are listed on component-mismatch if you do that, they either need to have a depends or be seeded [10:15] * directhex hands seb128 cake, udpates his transition wiki [10:15] directhex: ;-) [10:15] well, you're both right, (a) they're usually added (b) they need to be seeded [10:15] doko: newed [10:15] seb128: we seed after we have the mismatch, not if the package is in NEW [10:15] *-gcj is included automatically but *-java isn't [10:16] doko: right, just pointing that you need to have a depends or seed it [10:16] doko: there's nothing wrong with seeding when it's in NEW, actually [10:16] and it's often handy [10:39] tkamppeter: fix-the-world-cups uploaded to intrepid-proposed [10:43] pitti, great, thanks. Now the PDF filter chain should really work perfectly. [10:44] pitt, at least on Intel, on PowerPC there is bug 271350 (already reported to upstream). [10:44] Launchpad bug 271350 in cups "pdftopdf filter on PowerPC corrupts data" [Medium,Triaged] https://launchpad.net/bugs/271350 [10:46] pitti, have you put a comment into all bugs which I mentioned so that the posters know how to activate -proposed to get the new package and to try it? [10:46] tkamppeter: yes, including #293832 (the needinfo one) === enrico is now known as `ls` === `ls` is now known as enrico [12:03] tkamppeter: argh http://launchpadlibrarian.net/19831839/buildlog_ubuntu-intrepid-i386.cups_1.3.9-2ubuntu2_FAILEDTOBUILD.txt.gz [12:04] tkamppeter: a local build worked fine, because I had libcupsimage-dev installed; this probably needs an -I option to the local cups/image.h [12:13] tkamppeter: ok, fixed === veloc1ty_ is now known as veloc1ty [12:35] pitti: yeah that got slipped with things ill do that today [12:38] hi! avahi seems to be disfunctional in Intrepid [12:39] can't see any other services in the local network after reinstalling Intrepid. [13:17] pitti: bacula fixed and uploaded [13:17] cjwatson: http://launchpadlibrarian.net/19834786/buildlog_ubuntu-jaunty-armel.debian-installer_20081029ubuntu2_FAILEDTOBUILD.txt.gz [13:18] beep is in the archive, so why not beep.udeb? [13:18] doko: NCommander already prepared a MIR for this [13:18] https://wiki.ubuntu.com/MainInclusionReportBeep [13:18] ahh, ok [13:18] doko: there are a couple of others, you should join #ubuntu-arm :) [13:19] devio-udeb and oldsys-preseed [13:19] cjwatson: I thought we wanted to have this in #ubuntu-devel ... at least I did write this in my email [13:19] I don't think there are MIRs for devio-udeb and oldsys-preseed yet [13:20] I don't think we need a separate mailing list; IRC channel meh *shrug* [13:21] Mind you, over time, the IRC channel probably ends up being less dev and more support, and so people will want to retreat here, but for now it's mostly dev stuff. [13:21] superm1: well, we wouldn't remove images that are still part of an official release from cdimages.u.c, so mythbuntu/releases is staying :) [13:21] cjwatson, pitti, please promote https://bugs.edge.launchpad.net/ubuntu/+source/beep/+bug/300442 [13:21] Ubuntu bug 300442 in beep "Main Inclusion for Beep" [High,In progress] [13:22] superm1: but I've nuked mythbuntu/daily at least, yes - thanks for the nudge [13:22] doko: done, as I say will need the other two before it's worth giving back d-i though [13:22] ok [13:39] cjwatson: see #300676 and #300678 [13:51] kirkland: Sorry, I think I missed it... Why were we putting it into /var ? Something about configuration data? [13:51] soren: here's the key .... [13:52] soren: $HOME/.ecryptfs/wrapped-passphrase contains your mount passphrase, symmetrically encrypted with your login passphrase [13:52] soren: it is unencrypted in userspace by pam on login [13:52] soren: and then the mount passphrase is used to mount all of $HOME [13:53] soren: okay .... now .... [13:53] Oh, ecryptfs is not the data directory? [13:53] Sorry, I mean .ecryptfs. [13:53] soren: right the mount is this: [13:53] Ok, sorry :) Go on. [13:53] soren: mount -t ecryptfs -o lots_of_foo $HOME/.Private $HOME [13:54] Oooh.. [13:54] soren: persia: okay, this is where the unionfs or bind mount might work .... [13:54] Ok. [13:54] soren: because i need .ecryptfs to (again) be available in $HOME, for password changing, unmounting, etc. [13:55] Got it. [13:55] I thought .ecryptfs was the data directory.. [13:55] soren: but, i need to *absolutely* make sure that anything edited within there doesn't get written to the disk through the kernel ecryptfs crypto [13:55] here's a thought: [13:55] soren: no, just config [13:55] kirkland, I think a bind mount would be better. aufs does layering in a way that's causing me enough confusion I'm not getting you a mount line. [13:55] the data dir is /home/$USER/.Private [13:55] Before doing the ecryptfs mount, how about bind mounting the ecryptfs directory to somewhere else. [13:55] ? [13:56] you'd end up with a *lot* of mounts with multiple users logged in, is the only thing [13:56] soren: please suggest a mountpoint you find suitable? [13:56] cjwatson: Good point. [13:56] one per user is reasonable, more than that I start to find uncomfortable [13:56] kirkland: I'm not sure. [13:57] I'm getting more and more excited about the idea of a setuid binary in ecryptfs-tools to manage /var/lib/ecryptfs/$USER [13:58] Only problem there is namespace: limits to one encrypted mount per user. [13:58] kirkland: Let me just say: My primary reservation against /var/lib/ecryptfs/$USER was that I thought you were dumping lots and lots of data there... Since it's just configuration, I'm less opposed (but still a bit) to it :) [13:58] make it /var/lib/ecryptfs/$USER/$mountname, then? [13:59] Do ecryptfs's have uuid's? [13:59] * soren takes a quick break [13:59] Mithrandir, That sounds sensible. [13:59] soren: purely configuration [13:59] soren: purely $USER configuration, and i though /var/lib was less evil than /etc [14:01] hi again folks! [14:01] I've just resolved my avahi issue! [14:01] The latest ubuntu kernel seems to have problems with the ATL1E driver [14:02] after applying this patch it all works great: http://kerneltrap.org/mailarchive/linux-netdev/2008/11/11/4060524 [14:02] persia: a setuid binary to create that dir -- i like that [14:02] When can I report this to kernel maintainer? [14:02] where [14:02] * persia credits cjwatson for the suggestion : what you see here is an IRC echo [14:02] persia: i don't think it necessarily limits the mounts, as we could teach ecryptfs to deal with the data in the user's /var/lib/ecryptfs/$USER dir directly [14:03] kirkland, As pointed out above, /var/lib/ecryptfs/$USER/$mountname makes it moot. [14:03] persia: thx, i hadn't bounce back to the other channel [14:04] speakman: try #ubuntu-kernel [14:15] soren: cjwatson: persia: there is one other option, but it involves kernel changes .... [14:16] ecryptfs supports a mount option, ecryptfs_passthrough, which allows for an unencrypted file in the lower filesystem [14:16] you can read and write the file in the ecryptfs mount, and it remains unencrypted [14:16] That sounds nice and clean to me. No fussing about with stuff. [14:16] however, that feature doesn't yet work recursively on a directory [14:16] oh :( [14:16] ie, new files created in a passthrough dir are encrypted [14:16] that could be changed [14:17] and the kernel maintainer agreed that it would be doable, and a good feature [14:17] but i'm not sure how long that would take [14:18] kirkland, Now, how would this work for bind-mounting /home out of a chroot (e.g. a default schroot setup)? [14:18] Would I have to enter my password 8 times for an sbuild run like I do for pam_mount? [14:18] persia: no [14:18] persia: see keyctl [14:18] persia: keyctl show [14:19] persia: the key will be loaded into your kernel keyring [14:19] persia: if i'm understanding your question correctly [14:19] Cool. [14:19] persia: fwiw, i can debuild in an ecryptfs home no problem [14:19] which uses something similar, i think [14:19] No. Not at all. [14:20] sbuild, then [14:20] debuild uses the host system binaries. Very dangerous way to test-build if there's any chance you have stray development libraries around and an agressive ./configure. [14:20] Yeah, by default sbuild uses schroot. [14:21] I want to have a pam_time which would work somewhat like pam_stack and skip modules if not too long had passed, so I can do stuff like "the screensaver should be able to unlock using my finger if it has been locked for less than 15 minutes". [14:24] Mithrandir, That's a nifty idea. Interesting applications with pam_blue as well. === LucidFox_ is now known as LucidFox [14:27] persia: yup. [14:28] UTSL! [14:28] (note: pam modules are icky black magic) [14:37] err [14:37] wtf [14:38] why does the default apt sources list include the banshee and bzr PPAs ? [14:38] o_O [14:39] it doesn't ... [14:39] if I added random PPAs, I wouldn't have added banshee :) [14:39] cjwatson: mine has [14:39] I reinstalled today [14:39] with what installation image? [14:39] jaunty daily? [14:39]