[00:24] night === robert_ancell_ is now known as robert_ancell === Amaranth_ is now known as Amaranth [02:13] c/ === Kermiac_ is now known as Kermiac === asac_ is now known as asac === Amaranth_ is now known as Amaranth [07:38] Good morning [07:38] good morning o/ [07:47] kenvandine: FYI, setting WIs in https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-empathy-indicator back to TODO, and instead moving the entire spec [07:52] good morning [09:09] good morning everyone [09:11] hey chrisccoulson [09:12] hey pitti, how are you? [09:20] chrisccoulson: Je suis bon, merci! [09:29] hi chrisccoulson [09:29] bonjour mon ami [09:29] hi rickspencer3 - how are you? [09:29] chrisccoulson, pretty well [09:30] chilling with pitti and didrocks [09:30] watching them work out of the corner of my eye while I surf facebook and play tetris [09:30] it's important to have good managers [09:30] :-) [09:30] lol :) [09:30] * pitti rubs the whip cracking blisters on his back [09:32] are you all having lots of fun at the sprint? [09:33] (except for pitti of course, with the blisters on his back) ;) [09:33] chrisccoulson, so far so good, yes [09:33] awesome:) [09:33] * rickspencer3 is writing a blog post about it atm [11:31] hey [11:31] hey seb128 [11:46] asac: do you have time for another nm related question? If I have more than one connection, which one is set as default route? Can I control this somehow? Or disable this multi-homed environment at all? [11:46] asac: well, actually more like 3 questions ... === robbiew_ is now known as robbiew [11:58] TeTeT: ethernet > wireless > 3G ... you can try to manually add routes in the IPv4Settings tab of connections and use metrics for the default route to achive that [11:59] i am not sure if the metrics trick still work, but it did at some point in the past [11:59] let me know [11:59] * hyperair wonders what do metrics do for routes [12:02] asac: can you do that programmically via d-bus? [12:03] TeTeT: you can do everything through dbus [12:03] that is possible ,) [12:04] but it might be complex [12:04] asac: is there any documentation for the API? [12:07] TeTeT: i dont think so. but i might be wrong. you can introspect dbus APIs for NM using d-feet [12:07] asac: d-feet? [12:10] asac: how to use that? I see nm-applet there, but no object paths or alike [12:11] asac: ah, have to go to system bus it seems [12:11] yes [12:13] TeTeT: in the source there is ./docs/NetworkManager\ DBUS\ API.txt [12:13] check that [12:13] * asac thinks that should be packaged [12:13] i am not sure how up-to-date or accurate that is maintained [12:14] but before doing this, first try if the metrics trick helps [12:14] asac: awesome, thanks [12:15] TeTeT: the normal way is to disconnect from the ethernet connection (there is a button now in applet) if you dont want it to have the main traffic [12:17] asac: ok, I wait for the customers use case to see if they really need to toggle this routing manually [12:18] kk === MacSlow is now known as MacSlow|lunch [12:31] asac: thanks for your help :) [12:32] np [12:57] hello everyone === MacSlow|lunch is now known as MacSlow [13:15] any idea why NetworkManager says my eth0 device is not managed? and how can I set it to manage it? [13:19] rodrigo_: does it appear in /etc/network/interfaces ? [13:22] pitti, I added it by hand [13:22] seems the installation didn't detect it [13:30] pitti: no, after rebooting it still says it's not managed [13:44] chrisccoulson1: hey [13:44] hey pitti [13:45] chrisccoulson1: just as a heads-up for boot speed [13:45] it looks like we can change the plan to move the 10s target to UNE [13:45] chrisccoulson1: which seems easier to do, since we have fewer applets, fewer processes (like no nautilus desktop rendering), and more code under our control (UNE launcher) [13:46] so, the 10s target will not apply to the standard desktop? [13:46] chrisccoulson1: so the main targets are now gconfd and panel speed up [13:46] chrisccoulson1: I don't think we will be able to meet it TBH [13:46] UNE still uses gnome-session and g-s-d doesn't it? [13:46] panel speedup will benefit both desktops, but there's still nautilus/themes/fonts/etc. [13:46] chrisccoulson1: it does [13:47] yeah, i can still look at some other desktop components to help speed it up [13:47] chrisccoulson1: http://piware.de/tmp/unity-full.png [13:47] but at least the gnome-session, g-s-d and gconf work i've done will be beneficial for UNE [13:47] chrisccoulson1: so, I have some thoughts about those and would like to run them by you if you have a minute? [13:48] chrisccoulson1: so, you have changes in the pipe for g-s and g-s-d to start in parallel with gconf, right? [13:48] chrisccoulson1: for gconfd my idea was as follows: [13:48] pitti - can do (although, i might need to hide the chat window occasionally) [13:48] - pull gvariant into glib (once Ryan confirms that it's scheduled to land in the near future) [13:48] chrisccoulson1: heh, boss mode? :-) [13:49] - rip out /var/lib/gconf/defaults/%gconf-tree.xml [13:49] - replace it with a gvariant serialization file [13:49] if gvariant is out of the question for some reason, we could also investigate how fast sqlite is [13:49] but parsing a 2.2 MB XML file which has mostly noise is just crack [13:49] do you think that's too crackful? [13:50] pitti - yeah, it's quite extreme. i wonder how much benefit we'd get by ripping out all the comments [13:50] chrisccoulson1: there are no comments? [13:50] that seems a lot of work with gvariant though. some of the work i've already done works around the long gconf delays [13:51] ah, so you think it will then just interleave with the other delays, and not be an issue any more? [13:51] i did some work to parse the gconf tree in parallel with other things, so that it doesn't delay the whole session from starting [13:51] note that we still need to change the startup to launch all programs at the same time, instead of the current serialization [13:51] (right now, most programs start 4 seconds into the session [13:52] pitti - i'd be a bit nervous about removing the current serialization, as we might introduce weird race conditions and other issues [13:53] for example, there are things that g-s-d must do before anything else loads, else you can create extra X round-trips etc [13:53] i'd prefer to try and optimise the current serialization rather than remove it [13:53] well, that's why we need to do it soon [13:53] eg, it might be possible to start the panel and WM at the same time [13:53] seb128 played around with that already; did you notice race conditions with that? [13:54] (reading backlog) [13:54] i haven't played around with starting everything together yet, but i just wanted to point out that it might cause issues for some people [13:55] seb128: wrt. introducing race conditions when starting everything in parallel instad of serializing [13:57] hey tedg [14:01] chrisccoulson1: back (sorry, spoke with Seb); he said that it didn't cause any obvious regressions, and it starts everything aruond 1.2 seconds in [14:01] pitti - as an example, i think at-spi-regsitryd needs to set up some things in the environment before anything else loads (although i can't verify that for sure whilst i'm at work), which is why it is started in the initialization phase [14:02] yeah, right [14:02] if it doesn't do that before other components load, then it might break assistive technology for some people [14:02] chrisccoulson1: for seahorse-daemon, do you think it's possible to fork, export the env, and then do the real initialization? [14:02] this seems to be the other major blocking step [14:03] pitti - i think it already does that, but i need to investigate that still [14:03] it would need to set the environment before forking, and then doing the real initialization though [14:04] as once gnome-session has moved from initialization, it doesn't allow you to change the environment (the dbus method will return an error if you try to do it) [14:05] Morning kenvandine [14:06] hey tedg [14:07] chrisccoulson1: right, that's what I was proposing [14:08] chrisccoulson1: is that SSH_AGENT_PID/GPG_AGENT_INFO etc.? [14:08] it seems to me that you could create those sockets and determine their pids/values without much inintialization [14:09] chrisccoulson1: of course could couldn't actually talk to them in the first 0.5 seconds after session startup, but who cares.. [14:09] pitti - i've got a feeling those 2 variables are set by ssh-agent and seahorse-agent, which load from a Xsession.d script [14:10] hey stormy_ [14:10] seahorse-agent really should be a proper session agent, but it's not on the default install anyway [14:10] hey kenvandine [14:12] chrisccoulson1: hm, I think that gap is actually just gconf [14:13] the delay is trivial on http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-hacked.png (which has the "start all in parallel" stuff) [14:13] pitti - the 0.5s gap after gnome-session starts is mostly gconf [14:13] but i've got a fix for that already [14:14] pitti - that trace is with my g-s-d, gconfd and gnome-session patches [14:14] chrisccoulson1: right [14:14] that's the -hacked charts [14:15] chrisccoulson1: 1006-hacked is your changes plus seb's "one phase" change [14:15] pitti - that chart is also with the packages from here: http://people.ubuntu.com/~chrisccoulson/desktop-startup-speed/ [14:15] chrisccoulson1: right [14:15] great job with those! [14:16] chrisccoulson1, do you know if g-s-d needs to start first? [14:17] seb128 - yeah, g-s-d needs to start first really, to set up the xrandr stuff, and the theme as well === korn_ is now known as c_korn [14:21] chrisccoulson1, is xrandr doing anything if there is no config change? [14:21] the g-s-d code I mean [14:22] on a stock install it should do nothing and let xorg the way it is no? [14:22] seb128 - good question. i'll have a look later and see what it actually does if there is resolution change needed [14:24] chrisccoulson1: I think we could tell by looking whether we have an xrandr configuration file? [14:24] I'm wondering what seahorse-daemon is actually doing [14:24] seahorse-agent is the one doing gpg agent [14:25] chrisccoulson1: i. e. don't run it unless we have a .config/monitors.xml? [14:25] pitti - yeah, that might be possible [14:25] chrisccoulson1, seb128: so, we need to coordinate who is doing what now [14:25] AFAICS we have [14:25] I'm doing the one phase change [14:25] 1) speed up gnome-panel (sit down for a week) -- probably robert_ancell, need to confirm [14:26] 2) start everything at once -- seb128 [14:26] 3) speed up gconfd loading -- I could look into that [14:26] 4) land the gconf dependency reduction changes -- chriscoulson [14:26] asac: any chance that nm 8.1 will land in lucid? a customer is interested in the additional bluetooth phone capabilities. A PPA with 8.1 might be sufficient [14:26] 5) investigate seahorse [14:26] 5) not launch xrandr if we don't have a conf file [14:27] * pitti defers to being 6) [14:27] 7)... [14:27] seb128, chrisccoulson1: ^ does that look complete and sane to you? [14:28] 8) profit? [14:28] yes [14:28] I think sabdfl wants to claim 8) :) [14:28] sounds good to me [14:28] me too [14:28] all these changes will speed up gnome and UNE [14:29] just that it might be enough for UNE, while desktop still suffers from more applets, nautilus desktop rendering, compiz, etc. [14:29] while on UNE we have mutter, no nautilus desktop, fewer panels, etc. [14:29] chrisccoulson1: are you interested in 5) ? [14:29] pitti - yeah, i'll look at 5 too [14:30] chrisccoulson1, pitti: you rock [14:34] * pitti hugs chrisccoulson1 [14:34] I'll update the work items now to get an updated plan [14:35] chrisccoulson1: your changes were for g-s-d and gnome-session, right? [14:38] sorry, i had to disappear to get coffee there [14:38] i have a lot of people hovering around my desk ;) [14:39] pitti - my changes were for gconf, gnome-session and g-s-d [14:39] chrisccoulson1: thanks [14:41] chrisccoulson1, seb128: I updated the WIs on https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-startup-speed -- does that look ok to you? [14:43] pitti - looks ok. the only comment about the g-s-d WI though, is that the change is to load as much as possible without needing gconf [14:43] currently, that's only xrandr, as it's the only plugin that doesn't store any settings in there [14:44] chrisccoulson1: ah, I see; hm, that means once we eliminate xrandr we won't have a benefit from that any more? [14:44] or does it do other stuff as well? [14:45] we will still probably load the xrandr plugin, but i'll see if we can reduce the number of X calls for setting up the display configuration, if there is no stored configuration [14:46] i'll look in to that though anyway [14:48] does xrdb work in a dynamic way? [14:49] would be nice to delay that one too [14:49] seb128 - i was just asking if we even need xrdb? [14:50] we do for openoffice [14:50] the xrdb code is off by default [14:50] the xrdb call we have is from the xsettings option [14:50] and asac had to add fixes there for openoffice previous cycle [14:50] which indicates it's really useful [14:52] ah, ok.i didn't realise the xrdb plugin was off by default actually [14:52] i was just reading http://mces.blogspot.com/2008/10/improving-login-time-part-1-gnome.html [14:52] i think you sent me that link didn't you? [14:55] cassidy, what would be useful to debug a "can't reconnect to icq" error? [14:56] chrisccoulson1, yes [14:56] seb128, haze logs [14:56] see http://live.gnome.org/Empathy/Debugging [14:57] 'CI [14:57] 'ic [14:57] can't type! [14:57] 'ci [14:57] cassidy, haze is not in empathy debug dialog [14:57] yeah it doesn't implement the debug iface [14:57] so you have to do it old style [14:58] bah [14:58] if I restart empathy that will work [14:58] I already got the issue [14:58] it fails to reconnect but connect fine after a restart [14:58] thanks anyway [15:16] kenvandine, hi, did you see my last dent? [15:23] kenvandine, sorry I forgot to review your and upload your xchat indicator [15:23] I will do that today [15:23] thx [15:23] no worries [15:24] rickspencer3, woot! [15:28] kenvandine, so soon I need to start pulling feeds in [15:35] kenvandine, http://theravingrick.blogspot.com/2010/01/tonights-photobomb-feature.html [15:35] hey desrt [15:35] there's my microblogging UI [15:35] I suppose the title is maybe a bit silly [15:38] rickspencer3, cool, can you make that a separate widget we can include in gwibber? [15:38] seb128: hi =) [15:39] kenvandine, yeah, just grab the .ui and .py files [15:39] in other words, it's already a separate widgets [15:40] great [15:41] kenvandine, does gwibber already use .ui files? [15:41] now it does :) [15:41] it's a simple dialog, so I could just code the layout [15:41] the accounts UI does [15:41] ok [15:41] seb128: does lucid have the nice new tooltips (from RH I guess)? [15:41] kwwii, no [15:42] mclasen, ^ did the fedora tooltip changes went to GNOME? [15:42] I didn't notice any change in 2.29 I think [15:43] vuntz: do you think that if /apps/panel/general/toplevel_id_list is mandatory, it's possible that the fact there is no applet showing up at the first launch is related to a writing issue? (if I unlock it, run gnome-panel, and then relock it, I have my applets in subsequents launch) [15:44] seb128: http://ubuntuforums.org/showthread.php?t=1310730 seems to imply they did make it into gtk [15:45] kwwii, so maybe the theme or something needs to use that [15:46] seb128: hrm, ok..I'll look into it some more, thanks for the help [15:50] kwwii, https://bugzilla.gnome.org/show_bug.cgi?id=599618 [15:50] https://bugzilla.gnome.org/show_bug.cgi?id=599617 [15:50] Gnome bug 599618 in gtk "better tooltip positioning" [Normal,New] [15:50] https://bugzilla.gnome.org/show_bug.cgi?id=599617 [15:50] Gnome bug 599617 in gtk "a modern tooltip look" [Normal,New] [15:53] kwwii, so not upstream in git [15:53] upstream in bugzilla [15:53] seb128: hrm, can we apply that patch? [15:53] which one? [15:53] round corner? [15:53] the positionning one? [15:53] mainly the positioning one but the rounding would be nice too [15:53] I would like to check with mclasen if he plans to land that upstream first though [15:53] ;) [15:53] ok, sounds good...thanks for the help [15:54] np [15:54] can you open a bug on launchpad for those? [15:54] so it doesn't slip from our list of things to do [15:56] davmor2, not sure why you opened that evolution bug [15:56] could you try asking slangasek rather? [15:56] I think I did advice you to do that before you opened the bug [15:57] it's not evolution's fault if ubiquity removes it on installs [15:57] which it doesn't there on the install I did today btw [15:57] seb128: slangasek fixed it, it happened because 3 or 4 locales still used the evo-doc-locale file which removed evo [15:58] could you close the bug then? [15:58] thanks [15:58] seb128: sorry I thought slangasek had when he fixed it [16:06] pitti: would something like this work well for you? http://pastebin.ubuntu.com/356117/ [16:06] didrocks: hrm, didn't know that, but it might not be a big surprise [16:06] the code might consider nothing is writable if one key is not [16:07] vuntz: should be that so, I'll try to find a workaround [16:08] tseliot: yes, looks fine; however, it might be that we need transitional packages and versioned conflicts/replaces [16:08] tseliot: we should test the upgrade [16:08] pitti: right, which is why I suggested that we use transitional packages yesterday [16:09] tseliot: on a normal system where jockey is already installed, it should work fine [16:09] tseliot: you'd just need them on a system where envy is installed, but jockey isn't [16:09] pitti: a corner case but still worth testing === \vish is now known as vish [16:16] didrocks: or you can fix the bug :-) [16:19] hello vuntz [16:19] how are you? [16:25] seb128: I'm sad. Snow is melting here :/ [16:25] ;-) [16:25] seb128: I see you're having fun with login performance [16:26] indeed [16:26] it's rather down to gconf and gnome-panel being slow now [16:26] do you think gnome-panel could do async applets loading? [16:27] seb128: it's supposed to be async already [16:27] hum [16:27] note the "supposed" :-) [16:27] what robert_ancell observed seems to indicate it's buggy [16:27] ;-) [16:27] nice to see you clean gnome-panel btw [16:29] vuntz - you like the snow? [16:29] I like the snow too [16:29] when it doesn't make me miss my train [16:29] like today! [16:30] heh [16:30] i don't mind the snow, but i don't like taking 3 hours to get to work ;) [16:30] it took me 1.5hours this morning [16:31] chrisccoulson1: snow is amazing! [16:31] chrisccoulson1: it just makes the world a better place! [16:31] vuntz - did you build a snowman? :) [16:32] snow is great, when it's in the mountains where it belongs ;) [16:32] vuntz, you spend your days playing in the snow? ;-) [16:32] chrisccoulson1: I didn't think about it, actually. Maybe next time... [16:32] rickspencer3: you know, before you said that, I used to respect you. :-) [16:32] yeah, snow does make everything look quite nice. especially where i live - it's normally fairly dull and grey, but the snow brightens everything up :) [16:32] the snow is almost gone in ireland [16:32] :( [16:33] vuntz, you don't ski or snowboard? [16:33] I visit the snow the every weekend, it does not need to visit me [16:33] heh [16:33] rickspencer3: ski, but only a bit. I never tried snowboard [16:33] I just like the atmosphere when there's snow [16:34] people being stucked in cars for hours? ;-) [16:34] vuntz, come to Seattle in the winter, I will show you snow (where it belongs in the mountains ;) ) [16:34] you watch them from you chair since you don't need to go to work? ;-) [16:34] seb128: well, of course, the fact that I don't need to take a car/bus/train to work helps ;-) [16:34] and we can ski on it, as intended [16:34] hehe [16:35] vuntz - driving a car to work sucks in the snow ;) [16:35] i've spent too much time this week in my car [16:36] chrisccoulson1: you could also choose to not go to work. I mean, if you really hate being in the car... ;-) [16:37] vuntz - it did cross my mind, but i'm more professional than that ;) [16:37] of course, at some point, you'd have to stop eating too, since you wouldn't be able to buy food [16:37] oh rickspencer3 photobomb is awesome any chance it will replace fspot hehe [16:37] some of my colleagues stayed at home though [16:37] fagan, I think it already has [16:37] * rickspencer3 checks seeds [16:37] hmmm not yet, probably tomorrow [16:38] rickspencer3: hah it is growing quickly though (I stress the word quickly hehe) [16:38] fagan, goocanvas is quite awesome [16:38] and plus segphault has lots of useful code for me to snipe ;) [16:39] * fagan needs to get said useful code [16:40] fagan, $bzr branch lp:grabbersnap [16:40] I did the micro-blogging with the Gwibber API of course [16:40] rickspencer3: thanks [17:24] right, home time :) [17:59] hate this libmozjs thing [18:00] gjs is broken in lucid and I can't figure why [18:00] gjs is broken in lucid and I can't figure why! [18:00] ups lag [18:00] ldd lists libmozjs.so twice [18:00] once correctly and one not found too [18:17] didrocks, process_workspace_switch_grab in src/core/keybindings.c [18:24] njpatel: thanks [18:33] hey didrocks === rodrigo__ is now known as rodrigo_ [19:32] Hi. I have a machine running Karmic and I am trying to disable Compiz. [19:33] I go to Appearance, last tab, select None for desktop effects; Metacity launches [19:33] and next reboot I'm back to "Normal" and Compiz. [19:33] Is the choice of WM still dictated by a gconf key? === asac_ is now known as asac [19:51] asac: hey, just wanted some more info about removing OOo from UNE [19:52] is it being removed or replaced or neither? [19:53] djsiegel: removed and then we see how well we can integrate gdocs as a first step [19:53] asac: oh, ok [19:53] thats armel UNE btw [19:53] cant speak for the rest [19:54] djsiegel: for email we do something similar: check out: https://blueprints.edge.launchpad.net/ubuntu/+spec/mobile-lucid-arm-webservice-for-email [19:54] thats a bit better understood [19:54] e.g nicely integrated [19:54] would be happy to get your feedback [20:37] anyone, re compiz/metacity?\ [20:48] crb: what paroblem? [20:49] problem === cyphermo1 is now known as cyphermox === robbiew is now known as robbiew_ === cyphermox_ is now known as cyphermox === fta__ is now known as fta_ [23:55] * crimsun is pleased that PA is shaping up nicely for 10.04 [23:56] also, I'll write up the SRU for the p_udev leak shortly [23:57] \o/