=== yaboo007 is now known as yaboo [05:33] how can I force the screen resolution in 11.10 desktop [05:34] seems it tries to do 1600x1200 when my monitor only does 1280x1024 res [06:07] Good Morning. [09:11] Good morning [09:29] dobey: not sure whether someone already told you, but the distro uploading question was resolved with jasoncwarner_/rickspencer3; so it's fine to continue what we have always done, i. e. you can upload [09:30] pitti, dobey of course with the caveat that the distro acceptance tests were run and passed! [09:30] :) [09:30] sure; I checked the policy, and it doesn't say "canonical platform team", just "distro" [09:30] as distro == ubuntu developers, you as an uploader are responsible for the distro side [09:31] the AC still applies, of course === ejat- is now known as ejat [09:34] dobey: and since that was resolved, if you still want a U1 packageset, the DMB is open to it === zyga_ is now known as zyga [09:41] /c/c [09:49] Kaleo: (and Saviq): I updated the list of tasks in https://wiki.ubuntu.com/UbuntuTV/MergePlan, and i'm starting to go through it in order. Pleae confirm you're ok with it when you have time. [10:14] mvo, hi, I have a work item to "[mpt] Invite people to port release-upgrader to aptdaemon" [10:14] mvo, would you have a few minutes to tell me what I need to know to write something informative? :-) [10:20] Hey [10:21] New rhythmbox has file conflicts with -plugins on upgrades [10:21] tentative de remplacement de « /usr/share/rhythmbox/plugins/artdisplay/rhythmbox-missing-artwork.svg », qui appartient aussi au paquet rhythmbox 2.90.1~20111126.89c872b0-0ubuntu3 [10:27] and speaking of broken upgrades... [10:28] kenvandine, seb128: I have a work item to "Specify how unity and other critical packages shouldn't be removed by an update". The idea is that we'd have a list of packages where, if the an update wants to remove it, Update Manager would know the update is broken. Does that seem like a good idea to you? [10:29] lool, hey, seems a buggy replaces in ricotz's update, do you want to add the replaces? I can try to have a look later, at a the platform rally and in a meeting atm though [10:39] mpt, don't we do that with what ubunt-desktop depends on already? [10:41] mvo, does update-manager do anything special if an update would remove ubuntu-desktop? [10:42] seb128: The bzr branch seems out of date; missing push? [10:42] (I mean at lp:~ubuntu-desktop/rhythmbox/ubuntu) [10:42] lool, done, sorry about that [10:42] ups [10:45] lool, ok, fixed, please check I didn't screw, I was missing a revision when I sponsored the update yesterday [10:46] so pulled, fixed the conflict and push, should be ok [10:50] seb128: ok [10:50] lool, thanks [11:02] uploaded [12:49] what's up with the build farm? It'll take nine more hours for my packages to build? :'( [12:59] BigWhale: backlog appears to be disappearing slowly now [13:14] kenvandine: how safe is it to update to precise [13:14] ? [13:15] seiflotfy: no known major breakages today; you might run into a rhythmbox file conflict if your mirror is slightly behind [13:16] but another run of dist-upgrade will sort that out [13:16] and update-manager just ignores these anyway [13:16] pitti so on an x201 it should be ok? [13:16] seiflotfy: yes, I've used precise on this since day 1 [13:16] u on an x201 too ? [13:16] seiflotfy, should be safe [13:17] how is the battery issue? [13:17] is it covered now [13:17] my last ubuntu caused me a whole new battery [13:17] it was draining like crazy [13:18] seiflotfy: yes, I have an x201 (Arrandale) [13:18] seiflotfy: "issue"? [13:18] seiflotfy: I've run natty, oneriric, and precise on this box without an issue [13:18] well somehow power consumption wiht the last kernel was terrible [13:18] in the train I got 9 W (11.5 hours) [13:19] ok i just got me a new battery today [13:19] 9 cells [13:19] and [13:19] 6 hours and 11 minute [13:19] not bad [13:19] but i need more [13:20] :P [13:20] seiflotfy: well, my 11 hours is with wifi killswitch and 50% brightness [13:20] with wifi I only get 8 h [13:22] do have anything in you rc.local settings? [13:22] no, stock ubuntu install [13:23] seriously ? [13:23] ok wow [13:23] let me install precise then [13:23] seiflotfy1: you might want to check powertop what's draining your bat [13:23] seiflotfy1: wasn't much different in oneiric/natty, though [13:23] 10W with wifi and reasonably active, 8W with killswitch and idle [13:23] and precise? [13:23] same [13:23] hmmmmmm [13:23] I think precise is a tad better [13:24] I didn't keep exact records [13:24] and of course it's hard to reproduce workloads exactly over several hours [13:24] seiflotfy1: powertop is pretty good to debug rogue processes or devices [13:24] http://www.omgubuntu.co.uk/2011/11/linux-power-regression-overheating-problem-on-thinkpad-fixed/ === jasox_ is now known as jasox [14:12] pitti: thanks [14:49] mpt: sorry, I didn't manage to respond to the aptdaemon question yet, I will try to get it later today [14:50] mvo, I wasn't sure whether you saw my other question too: mvo, does update-manager do anything special if an update would remove ubuntu-desktop? [14:54] pitti: http://paste.ubuntu.com/800630/ [14:56] mvo, those questions might be quicker to answer on Mumble [14:56] desrt: pushed [14:57] mpt: I have a real-life meeting in 5min :/ - I have not seen this question, but the anser is that if that happens it will fail hard and refuse to upgradde [14:57] pitti: thanks [14:58] mvo, how can I test how update-manager presents that failure? [14:58] (Does it put up an error message?) [15:03] mpt: hm, i don't think there is a really easy way, but its the generic: http://paste.ubuntu.com/800640/ [15:03] message [15:05] thanks mvo, I'll tweak that text [15:08] thanks! [15:53] cyphermox, hello [15:54] cyphermox, are there known issues with nm with probing wired interfaces which arent connected? [16:02] nm probing wired interfaces which are not connected? [16:03] ricotz: not that I know of. If you're mentioning some errors in dmesg about speed/duplex, I investigated that yesterday [16:03] I'm not *certain* it's not NM, but reasonably sure since NM listens for netlink messages to know when there is carrier [16:04] chrisccoulson: git://gitorious.org/performance-scripts/mainline.git [16:05] hello all! quick scream for help: the precise fresh installation in my laptop is stuck in "removing conflicting operating system files..." any idea what should I do? [16:05] chrisccoulson: http://people.gnome.org/~federico/news-2006-03.html#timeline-tools [16:12] cyphermox, i mean nm seems to try to connect while is there no cable plugged in and doenst give up trying [16:13] ricotz: I'll need much more information than that -- please file a bug (ubuntu-bug network-manager) and include syslog, and the output of 'ethtool ' [16:14] ricotz: I'm thinking that might be something device-specific, maybe some carrier-detection pieces got broken [16:15] cyphermox, i will try to get some better information, thanks [16:15] thanks! [16:15] btw the driver is skge [16:15] yeah.. one funky one or something :) [16:15] ethtool will probably be the key === m_conley_away is now known as m_conley [17:35] good night everyone [18:18] pitti: oops. there's https://bugzilla.gnome.org/show_bug.cgi?id=638117 that specifically mentions not saving the state in gnome-bluetooth [18:18] Gnome bug 638117 in general "Remember off/on state between sessions" [Normal,Unconfirmed] [18:19] guess there's a way to fix this in bluez though; the init script or upstart job could save the state on shutdown, kind of like alsa does [19:17] anybody around? i'm looking at bug #631638. it sounds like upstream is unlikely to pick this up in time for precise, so it seems like it would make sense for us to pick up the patch... [19:17] Launchpad bug 631638 in hundredpapercuts "palimpsest thinks that USB2 is 480MB/s instead of 480Mb/s" [Undecided,In progress] https://launchpad.net/bugs/631638 [19:49] broder: yes we should definitely take that patch for precise [19:49] jbicha: ok, cool. i'll put something together [19:51] or rather, i'll have one of the people at this bug jam with me put something together :) [19:52] \o/ [19:59] steve holt [20:14] where the heck is the shotwell irc channel? [20:15] or yorba or whatever [20:17] shotwell developers see those bugs by themself in launchpad ubuntu [20:18] i guess no need to upstream [20:18] and i dont know any channel [20:21] bil21al: i don't want to upstream a bug, i want to bug developers with questions [20:21] and i hate mailing lists [20:22] hmmm ask seb128 he may be know that oe brain murray in ubuntu bugs channel somebody there must knew about that [20:53] kenvandine: are you around, or at the pub? [20:53] dobey, i am around, but still working a bit heads down [20:53] * kenvandine should be at the pub :) [20:54] kenvandine: heh. i was just wondering what the plan was to get gwibber to stop using python-webkit for precise :) [20:54] dobey, stop hanging out with seb128 :) [20:54] he asks me everyday :) [20:54] lol [20:54] no plan right now... we need to rewrite the accounts UI [20:55] and i have no time for it right now [20:55] kenvandine: well, i'm poking about to find room for pyqt; and you, shotwell, and u1 are the only things that depend on gtk2 webkit still [20:55] the accounts ui is all python right? [20:55] dobey, shotwell doesn't anymore:) [20:55] yes [20:55] don'g say pygi [20:55] don't [20:55] kenvandine: well, trunk doesn't. but shotwell in precise still does. there isn't a relase of gtk3 shotwell yet afaict [20:56] yes there is [20:56] was uploaded today :) [20:56] oh [20:56] well hooray then [20:56] yup [20:56] :-D [20:56] so gwibber ;) [20:57] dobey, shhhhh [20:57] it would be hard to use pygi, i think [20:58] the accounts dialog does some python import muckery to inject python files from other places [20:58] i am sure it would introduce problems [20:58] well... s/sure/afraid/ [20:58] ugh, lots of stuff uses pygtk, not just accounts :-/ [20:58] nope [20:59] not in gwibber [20:59] not that we still use [20:59] there is some files that can be removed [20:59] oh [20:59] there is junk in trunk, that shouldn't be there [20:59] yes [20:59] that confuses grep :) [20:59] gwibber/microblog/plugins/*/gtk/*/ [20:59] and gwibber/accounts.py [20:59] should be all we have that we care about [21:00] and there is gwibber-service-{sina,sohu} which are in separate projects [21:00] ok [21:00] but provide accounts files [21:00] what projects? same name? [21:02] well i guess they don't use webkit, otherwise they'd show up in rdepends [21:03] kenvandine: oh, new shotwell is in depwait for new valac [21:04] kenvandine!! [21:04] weird [21:05] since vala-0.16 was uploaded 8 days ago [21:05] kenvandine, there was this guy who's name was JamesT and he wanted to run Gwibber on Net BSD! [21:05] wow [21:05] (he wasn't really a jamesT ... but ... ) [21:05] dobey, it is probably in universe [21:06] and shotwell is in main [21:06] kenvandine: ah, it is in universe [21:08] BigWhale, i made a huge improvement to gwibber-service's power consumption and CPU usage today [21:08] awesome! [21:08] people will be happy [21:08] dobey, i finally was able to get rid of threading.Thread in gwibber :) [21:08] heh [21:09] multiprocessing too? [21:09] python multiprocessing.Pool was causing constant polling looking for jobs to do even when idle [21:09] so eating up power [21:09] i bet [21:09] dobey, no.... i wish [21:09] i can't wait to get the new twisted in [21:09] but i couldn't close the pool when it was run in the thread [21:09] hopefully my reactor will get approved/landed soon [21:10] so getting that shit untangled i was able to close the pool down after finishing operations [21:10] kenvandine, threading in python is horrible :/ [21:10] yeah [21:10] which fixed the power problem... and a nice side effect is refreshes are now quite a bit faster [21:10] and use very little CPU [21:10] I have to start hacking on gwibber again ... [21:11] i went from 100% cpu usage for ~20s to spiking at 14% over the course of ~10s [21:11] and most of that 10s is under 10% [21:11] it is in ppa:unity-team/ppa if anyone wants to help test it :) [21:11] I am mostly finished with Kazam, just few minor things left ... [21:11] along with all the new unity crack and gwibber ported to the new APIs :) [21:13] does compiz no longer wake up 60 times a second with the new unity? :) [21:14] no idea [21:15] but i know half of the top 12 power hog bugs in precise have been fixed this week [21:17] the big one in u1 is fixed upstream [21:17] gwibber-service was causing 7% of the wake ups on a clean install [21:18] bug 906916 [21:18] Launchpad bug 906916 in gwibber "gwibber-service has a child that polls at 10Hz which is causing a lot of wakeup events" [High,Triaged] https://launchpad.net/bugs/906916 [21:18] does it run even if not using gwibber? [21:18] or only after you add accounts? [21:19] i think even if you have no accounts [21:19] yeah, that's bad :) [21:19] i had tried fixing the multiprocessing stuff in the past, but it kept causing more problems... i hit the sweet spot today :) [21:20] mixing threading and multiprocessing sucked [21:20] yeah [21:21] as soon as i looked at that code when the couchdb/keyring issues came up, i was thoroughly appalled [21:21] yeah... your patch didn't work though :) [21:21] it did work, just didn't fix *all* the cases :-/ [21:21] well, it fixed some problems and caused others [21:21] it was just a mess [21:22] s/caused/exposed/ :) [21:22] indeed [21:22] but things are much better now it souns like [21:22] yup! [21:22] sounds even; this keyboard is picky sometimes [21:22] :) [21:23] i need quite a bit of testing though [21:23] that code scares me [21:23] heh [21:24] you guys need to try the crack in the ppa :) [21:24] i need to upgrade my workstation to precise [21:24] then i can just ignore the fact that gtk 3.2 is apparently broken somewhere [21:25] who is responsible for getting software into universe? [21:25] since gtk in precise 'just works' where i'm hitting that brokenness [21:25] BigWhale: anyone with permissions to do so [21:26] BigWhale: motu and core devs all have the power [21:26] or powah, if you prefer [21:26] the force, please ... [21:26] BigWhale, i could [21:26] :> [21:27] kenvandine, ahh someone I know! ;) [21:27] although i am kind of drowning in work atm [21:27] kenvandine: popey keeps posting pictures of beer on g+ :( [21:27] you could create a package and i'll review it :) [21:27] kenvandine, well first I'll need few pointers on what still needs to be done [21:27] dobey, there has been plenty of beer :) [21:28] although i spent last night and tonight working late :( [21:28] kenvandine, well there's no real rush, except that I'd like to get it into precise :) [21:28] pile of new releases and crap to deal with [21:28] i'm sure there has been [21:28] i love it when there is an abi break that requires a bunch of rebuilds [21:29] are you guys occupying that bar on the street? :)) [21:29] and i love api breaks even more where i have to do last might porting [21:29] BigWhale, most of the time :) [21:29] kenvandine: i am totally about to break some API :) [21:29] * kenvandine smacks dobey [21:29] didrocks had me update from a PPA which broke gwibber... forced me to port the lens :) [21:30] kenvandine: do you not want a fast gtk3 version of u1ms? :) [21:30] dobey: the compiz 60 wakeups a second can be worked around by drm.vblankoffdelay=1 if you're on intel [21:30] kernel command line parameter ya can add to grub [21:30] dobey, well yes :) [21:30] dobey, sometimes it's worth it :) [21:30] the libdee api break was totally worth it [21:31] huge performance fixes that speeds up unity and gwibber :) [21:34] so dropping libwebkitgtk-1.0-0 off the CD will get us about 7 MB back [21:35] woot [21:35] dobey: a unity update wouldn't fix that though, its a kernel problem where vblank interrupts aren't turned off for 5000ms after they are used so you get your refresh rate number of wakeups per second pretty much always. that might be fixed in the 3.3 kernel, but only on intel where vblank interrupts are reliable, or you can add the kernel command line option to work around it without a newer kernel [21:38] Sarvatt: well, a backport of that fix to 3.2 kernel would be fine as well. doesn't have to be a unity or compiz update if the bug is elsewhere. kernel update is fine :) [21:39] maybe post 12.04 release when it actually goes in :) [21:42] in the meantime GRUB_CMDLINE_LINUX_DEFAULT="quiet splash drm.vblankoffdelay=1" in /etc/default/grub would help if you want to save power, its a seriously small amount though [21:42] like less than 2%, cking on the kernel team did some exact measurements [21:42] Sarvatt: i don't think gaining an extra hour is a "seriously small amount" [21:43] no reducing the 60 wakeups a second from compiz wont give you an extra hour by any stretch of the imagination [21:43] let me dig up ckings power measurement posts on the lists [21:44] Sarvatt: the battery icon's estimated time remaining increased by an hour on my laptop, simply by switching off of unity to fallback gnome, and turing off ubuntuone-syncdaemon (which has 10 wakeups/sec) [21:44] http://mid.gmane.org/4F05A844.6000707@canonical.com [21:44] dobey: we should totally switch to gnome fallback for precise then! :) [21:45] syncdaemon is keeping the hdd alive afaik, its very different and cant just compare the wakeups [21:45] no it isn't [21:45] it only hits the hdd at most once every 2 minutes, if not actively doing anything [21:46] its touching a log file which is keeping the hdd spun up I mean [21:46] i could be wrong, still digging [21:46] 30 times per hour, yes [21:46] not 10 times per second [21:46] all of his results were here http://zinc.canonical.com/~cking/power-benchmarking/ [21:46] he filed that gwibber bug [21:47] yes, he filed a bug or two against u1 as well [21:53] this laptop sucks for power anyway :-/ === m_conley is now known as m_conley_away [22:41] Can somebody look at this merge proposal and check it for mistakes: https://code.launchpad.net/~stebalien/gnome-disk-utility/fix-units/+merge/88283 [22:56] man gsettings is annoying :( [23:17] kenvandine: so, i just changed the color for my twitter account, in gtk3 ;) [23:18] anyway, i need to write this e-mail, and call it a day for sure [23:18] dobey, what do you mean in gtk3? [23:18] kenvandine: pygi [23:18] in a branch [23:18] oh... do you mean actually loading the accounts panel for the twitter account? [23:18] woot! [23:18] yes :) [23:19] dobey, you are my hero! [23:19] it still needs some fixes and testing, but i got pretty far already [23:19] and there doesn't seem to be a gnomekeyring gir [23:19] push that before it gets lost :) [23:19] ugh! [23:19] but maybe we can use txsecrets from ubuntu-sso-client [23:20] which just talks to keyring directly over dbus :) [23:20] so works with any keyring that implements fd.o secrets api [23:46] later, need to not hack for a while :) [23:57] Sleepy time for me too