[00:00] hmm g_static_private_get has no info in apidoc when it was added :) [00:02] ok g_private_get was added in 2000 [00:02] so thats probably in [00:03] same for static private get [00:03] so yeah. the g_main_context_pop_thread_default [00:03] funcs can just be inlined [00:04] g_poll is confined too [00:04] feels doable [00:04] i think the main_context stuff you cannot inline into libsoup, but thats the thing we need to add to main.c in glib [00:04] the rest can be done where they are used from what i can see [00:05] ccheney: ^^ [00:05] so ... with that, do you feel comfortable to get libsoup done by static inline copying the funcs from the paste to where they are used? [00:06] ok [00:06] i think so [00:07] looking at gdk/gtk now [00:07] yeah [00:07] webkit should really be backporter friendly in general i would hope ;) [00:07] as best as i can tell webkit only uses the following two from gdk/gtk [00:07] gdk_x11_window_get_drawable_impl [00:07] ccheney: do the same for epiphany [00:07] gtk_widget_get_window [00:07] e.g. glib and gtk check [00:07] only two? [00:08] good [00:08] yea it seems to be [00:08] but maybe check all libs that come from the webkit [00:08] i thought that also comes with gtk wrappers ... would be odd if they dont have anything [00:08] it seems there is only one lib but two libexec to check [00:09] looks like they went away between hardy and karmic [00:10] just has a common, dbg, library, and dev package now [00:13] epiphany itself brings in gobject-introspection stuff along with cairo, and other things i am not sure if those are really needed for new versions though [00:13] ccheney: i am not sure why you think webkit uses only two gtk_ symbols [00:13] i see loads of them [00:13] if i check with nm -D [00:13] in karmic [00:14] i only saw two that didn't exist in hardy [00:14] http://paste.ubuntu.com/352041/ [00:14] unless i screwed it up somehow :( [00:14] ok. those are probably inline stuff too [00:14] then i misunderstood [00:15] as i said, chromium uess glib and gtk from trunk [00:15] err from hardy :-P [00:15] * asac should wait for context-switches to finish [00:20] lot more for epiphany [00:21] 7 gdk, 34 gtk [00:22] http://pastebin.ubuntu.com/352045/ [00:23] hmm [00:23] its supposed to already have some of those symbols in 2.12 [00:23] like gtk_status_icon_is_embedded [00:25] my diff is messed up somehow [00:25] thats expected ephy i could have placed a bet on ephy folks jumping on latest symbols instantly ;) [00:26] gtk_toolbar_new -> that feels old ;) [00:26] most likely since 2.0 [00:26] as there is no "Since field" in api doc [00:26] ugh wasn't sorted right [00:27] i guess gtk_status_ is new [00:27] 4 and 13 not as bad [00:27] also added in 2.10 already [00:27] good [00:27] please paste [00:27] ccheney: oh. can we put that on the wiki page somehow? [00:27] all those results we find here ;) [00:28] just thinking would be good [00:28] or a bug [00:28] ;) [00:28] http://pastebin.ubuntu.com/352048/ [00:29] ok the gdk_ stuff is not an issue in intrepid jaunty at least ;) [00:30] great :) [00:31] the entry icon stuff is 2.16 ... so intrepid too [00:31] an issue [00:32] ccheney: ok. so maybe to finish today make a wiki page about the epiphany backporting effort and add the symbols that need backporting. we can then track implementatoin of those etc. there and also document in advance if they are needed in intrepid etc. [00:32] ok [00:32] of course feel free to finish libsoup ;) [00:32] :) [00:32] for hardy [00:32] yea [00:32] * ccheney is being called to dinner, bbl === bjf is now known as bjf-afk === eeejay is now known as eeejay_away [01:14] empathy so rocks [01:15] I wonder if I can get by without vonage? === eeejay_away is now known as eeejay [01:50] rickspencer3: i just use a cheap voip provider with a cisco spa-3102 box (~ $70), you could do something like that and have sip to you empathy and to home phones [01:50] not sure if you can do that with vonage or not [01:51] with the spa-3102 i have a landline to send faxes and alarm system and can router local calls out of it for free and just use the voip for long distance [01:51] s/router/route/ [01:51] all using the same phones, just setup the dial plan to automatically send the calls where they belong [02:00] The spa3102 is a great box, I use one of those with my VoIP and a cordless handset, with naked DSL. [03:03] http://192.168.1.102/cumine/displayimage.php?album=52&pos=23 [03:03] oops -- disregard -- apologies [04:12] hey bratsche. sup [04:13] i have a question about the rgba/decorations code that just landed === asac_ is now known as asac [06:23] morning everyone [07:19] Good morning [07:21] ccheney: I'm not sure that I fully understand your question, but I don't see why it shouldn't work; the second library that you build will have a bumped dependency to the first (since you have to increase the shlibs version for both), but that shouldn't matter === cdE|Woozy_ is now known as cdE|Woozy === eeejay is now known as eeejay_away [09:00] hello there [09:00] bonjour seb128 [09:01] hey pitti [09:01] spent another hour fighting to get karmic somewhat working for my cousin [09:01] some days I realize that ubuntu is not "there" yet [09:01] toshiba laptop = fail [09:01] the pixma combi = fail too [09:01] I get that told every other week by a friend of mine [09:02] needs to get debs on the web to get that working [09:02] after fighting resume still doesn't work in karmic [09:02] works on a lucid usb stick though [09:02] but I can't upgrade him to lucid now [09:02] and fan is always blowing full speed [09:02] which is very noisy and annoying [09:02] lucid> at least there's improvement in sight [09:02] which I've no clue how to fix [09:05] * seb128 happy to have dell intel hardware [09:12] Hey chrisccoulson [09:13] hey seb128, how are you? [09:13] good thanks [09:13] you? [09:13] you managed to travel back from work yesterday? [09:14] I'm off for about an hour for some errands [09:14] yeah, we finished work early yesterday [09:14] i came in this morning, but most of my colleagues stayed at home! === cypher__ is now known as czajkowski [10:39] slomo, hi [10:39] seb128: hi [10:39] slomo, you might want to backport http://git.gnome.org/browse/gtk+/commit/?id=acb6f4ff33bb9eebd8c2b8ba4404d282426e1eeb [10:39] I just did for lucid [10:40] without it synaptic seahorse gtg crash [10:40] well it's only in experimental for debian [10:40] so you can probably wait next week update too [10:40] just letting you know [10:41] thanks, i already noticed these crashes and wanted to look for a solution later today :) [10:41] saved me some time [10:41] cool [10:41] I wish debian allowed source uploads [10:41] I would have uploaded directly to debian [10:42] it takes ages to upload gtk binaries from there [10:43] np :) i'll build and upload it now [10:44] thanks [10:45] * hyperair grumbles about stupid gnome screensaver not noticing my mouse movements while fading while i'm watching a video on mplayer. so annoying. [10:51] pitti: hey, you had some ideas for the evdev udev rules, now that it loads the driver for devices like the video bus etc [10:52] pitti: something about narrowing down the list of devices to use it [10:55] seb128 - i've put some packages for you to test in http://people.ubuntu.com/~chrisccoulson/desktop-startup-speed/ when you get the chance [10:56] chrisccoulson, thanks [10:56] although, i'm not sure how much difference they'll make. they didn't seem to make that much difference on my machine, but it's hard to tell because i get a lot of variation anyway [10:57] i've got a feeling that the gconf change makes gnome-session block for a while again, but i'm not too sure. it would be interesting to see your bootchart with those changes [11:00] I will do that in a few [11:00] pitti, is sound working on your dell in lucid? [11:01] seb128: yes, very well; I'm listening to RB right now [11:02] pitti, what profile and output do you use? [11:03] seb128: I use the headphone jack on the side to connect to my speakers [11:03] "profile"? [11:03] pitti, in the gnome capplet [11:03] I hate this thing [11:04] like I don't know if I should use "analog stereo duplex" [11:04] or "analog sterea output" [11:04] tjaalton: I'm actually afraid we need the keymap for all devices which have buttons or keys (even the lid), otherwise X.org and libxklavier has an inconsistent idea of what the keyboard layout is [11:04] or what is the difference [11:04] tjaalton: and we certainly do need evdev itself for lid switch, power button, etc.; otherwise g-p-m and friends wouldn't ever get the X events for "power button pressed", etc. [11:05] seb128: it says "analog stereo duplex" here; I don't think that I ever touched it [11:05] pitti: ok, my lid switch only has ID_INPUT, so evdev is unloaded since it says it can't handle that [11:05] pitti, thanks [11:05] pitti: same thing with "hda digital pcbeep" [11:05] tjaalton: hm, that sounds like a bug [11:06] tjaalton: hm, indeed, here too [11:06] pitti: but we'll not load evdev for ID_INPUT_JOYSTICK, since it causes only trouble [11:06] tjaalton: my lid switch has x11_driver=evdev and ID_INPUT=1, no keys [11:06] tjaalton: ah, because it doesn't have any KEY capability [11:06] just a SW==1 [11:06] pitti: mine too, but evdev is unloaded [11:06] so that should be correct [11:07] tjaalton: right, same here apparently [11:11] is launchpad working for anybody? [11:12] sloooowly [11:12] i can't access it at all :( [11:12] i just get an oops [11:13] ah - "The topic for #launchpad is: LP DB security updates, possible intermittent outages 11:00-11:30 UTC" [11:17] pitti: ok, we'll only skip ID_INPUT_JOYSTICK, since it's what we did with hal too. hdaps devices are listed as joysticks as well, so that should probably be fixed in input_id? [11:17] tjaalton: what is a "hdaps"? [11:17] tjaalton: would you mind filing an udev bug report about it if it's misdetected? [11:17] pitti: accelerometers on laptops [11:18] detects a falling laptop :) [11:18] aah [11:18] or, in this case, returns the mouse cursor to the center of the screen when you tap your laptop [11:18] :) [11:18] it should have a SW_BRACEFORIMPACT, SW_CRASH, and SW_BROKEN then :) [11:19] chrisccoulson, ok, looking to your changes now, do you think I should do incremental updates or all together? [11:19] pitti: hehe [11:20] seb128 - you'd need to do them all together for any benefit [11:21] chrisccoulson, ok, doing that then [11:30] mpt: you in england today? [11:30] and471, yes [11:30] Happy snowy new year :-) [11:30] mpt: having fun in the snow :-p [11:30] mpt: hehe === MacSlow is now known as MacSlow|lunch [11:34] mpt: we have a snow day at school today :-) [11:46] chrisccoulson, your changes don't seem to make a real difference :-( [11:46] I'm uploading the chart soon though [11:47] seb128 - yeah, it would be good to see the chart. i've got a suspicion that the gconf change blocks gconf-sanity-check from exitting, which delays gnome-session again [11:48] but i couldn't really tell on my machine [11:49] i need to invest in some real hardware really. virtualbox is no good for doing this type of thing :-/ [11:49] well virtualbox should work ok to tell you what block on what [11:50] doing yet another boot to see the difference [11:51] it's hard to measure small changes since desktop login is between 10.7s and 11.3s without changing anything [11:52] yeah, that's pretty hard. i seem to get several seconds of variation in virtualbox [11:52] but i suppose it depends on what else i'm doing on my desktop [11:53] http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-2.png [11:53] is today's stock [11:55] copying the updated one now [11:56] or not, one minute, that one boot was weird [11:59] * asac finally sponsored gnome-bluetooth merge [11:59] asac, thanks ;-) [11:59] baptistemm: ^^ [12:00] i also jumped in and fixed fontconfig build bustage last night ... [12:00] * asac wants hugs for that ;) [12:00] hmmm? [12:00] baptistemm: wasnt that your merge ;)? [12:01] long ago [12:01] chrisccoulson, http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-6change.png [12:01] I did a request long time ago for karmic-update [12:01] yes. that went to lucid now [12:01] chrisccoulson, http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-6change.png is yours [12:01] ah, so it is a 2.28.x version, right ? [12:01] or with your updates rather [12:01] i didnt see fixes for -updates [12:01] baptistemm: yes. .3 [12:01] we can instantly go to .6 or whatever is latest if you want [12:02] * seb128 hugs asac the rockstar [12:02] thanks for fixing the fontconfig thing too [12:02] and uploading gnome-bluetooth [12:02] hehe ... now feeling better [12:02] fontconfig 2.8.0? [12:02] no [12:02] not yet [12:03] that merge was too work intensive to get it up as an emergency fix [12:04] seb128 - hmmm, so, it doesn't make a great deal of difference. it definately shows that gnome-session starts g-s-d much more quickly now (almost immediately), which is nice. but g-s-d still has to wait for gconfd to finish parsing the database [12:04] right [12:05] and seahorse seems to take the gconf start hit [12:05] it takes almost 1 second now [12:05] and it correspond to the gconf start [12:05] ah yes, that might be what's doing it [12:06] let me turn off seahorse just to see [12:06] i'm slightly surprised that gconfd seems to sit idle for a little while though [12:08] i think g-s-d is still blocking on gconf there though, as you can see that xrdb is spawned later on [12:09] well it seems blocked on seahorse rather [12:10] chrisccoulson, http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-7.png [12:10] without seahorse [12:10] it wins almost 0.5s on the compiz start [12:10] interesting. that's quite a lot faster. is it consistently like that? [12:11] compared to stock lucid [12:11] I did only once without seahorse [12:11] let me try again [12:13] what is seahorse-daemon doing again? [12:13] I though that was not useful by default [12:14] chrisccoulson, http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-8.png [12:14] seems consistent [12:14] good work [12:14] so we just have to deal with the seahorse-daemon case now [12:14] and we win almost 0.5 seconds [12:14] excellent. i'll take a look at seahorse-daemon next then, and see what we can optimise there [12:15] well that thing should just not be starting by default I think [12:15] I though we had a gconf key which made that conditional [12:15] but it has the gconf init cost in any case with that [12:15] which we want to avoid [12:16] hum [12:16] yeah, adding gconf keys for conditional starting will just add the delay back to gnome-session again [12:16] Gnome bug #587539 [12:16] Gnome bug 587539 in Daemon "seahorse-daemon.desktop should not contain an autostart condition by default" [Minor,Resolved: fixed] http://bugzilla.gnome.org/show_bug.cgi?id=587539 [12:17] I don't care about the nautilus thing, it's a power user case [12:18] I'm wondering if the daemon could be started on demand rather [12:18] it doesn't seem something which needs to be running from the start [12:19] yeah, i'll have a look at that. i need to familiarise myself with what it's needed for first :) [12:26] chrisccoulson, with your changes and seahorse not blocking things + no compiz we have 8.5 seconds [12:27] that's desktop login [12:28] thats pretty good. it would be interesting to see the stock boot without seahorse as well i think, just to see how much seahorse influences it. when i look at the different bootcharts, i think my changes are only worth ~200ms at the moment [12:28] I would say 300ms but right... [12:29] we are still down to make gnome-panel and nautilus quick [12:30] right, lunch time [12:30] bbl [12:30] hum lunch! [12:30] chrisccoulson, enjoy === MacSlow|lunch is now known as MacSlow [12:30] pitti, ^ backlog might be interesting to read for you too [12:32] combined with session changes to not wait we are under 8 seconds [12:33] and nautilus start at 1.1 second [12:39] chrisccoulson, pitti, http://people.canonical.com/~seb128/bootchart/seb128-dellmini-lucid-20100106-hacked.png [12:39] just for the record [12:39] it's chrisccoulson's changes + no compiz + no seahorse [12:39] + session hack to not delay starts [12:40] I don't think we can tweak much over those changes [12:50] chrisccoulson, dropping seahorse without your changes is a no win [13:44] pitti, ping. Sent you mail re: latest likewise-open packages. I believe the packages are ready for upload. Just need to know what I should do next. [13:46] * coffeedude heads for another coffee....brb [13:56] seb128: read the backlog now, interesting! [13:56] pitti, cool [13:58] . o O { still three pulseaudio processes.. } [13:58] 5 actually [13:58] coffeedude: hey Jerry [13:58] hey pitti :) Morning [13:58] coffeedude: nice, thanks! if that fixes the upgrade case, nothing; I'll just sponsor it to lucid [13:59] pitti, perfect! [14:00] pitti, I got some weird bootchart by dropping some extra things [14:00] pitti, if I need to do anything else regarding the upload, let me know. Otherwise, i'll just prepare to start testing alpha2 and fix anything that comes up. [14:00] the indicator message applet and the background image [14:00] it didn't win much time [14:00] but cpu use was not bounded for a few seconds [14:01] not sure what is limiting if that's not io or cpu [14:01] ie there was some seconds not fully busy [14:01] I would expect things to keep busy while there is work to do [14:05] unless something is waiting for some signal or dbus activation, indeed [14:06] coffeedude: perfect, thanks! [14:06] maybe idle loop lead to lazy use [14:06] rather than full cpu use === robbiew_ is now known as robbiew [14:34] bratsche, hey [14:35] Hey seb128 [14:35] bratsche, thanks for your gnome-desktop and nautilus fixes, will you send those to bugzilla too? [14:36] Yeah I will. I need to test them with an unpatched gtk+ first just to make sure that they don't cause regressions without the rgba patch (I don't think they can, but just want to test before posting to bugzilla) [14:37] I'll do that now. [14:42] Yeah, it seems fine as I expected. I'll post the patches now. [14:47] bratsche, thanks [14:47] I will sponsor the changes [14:47] but I just have people rining at the bell there [14:48] I'm away for half an, hour [14:48] brb [14:48] Okay [14:48] Thanks! === bjf-afk is now known as bjf [15:11] Riddell: what can I do for you? [15:12] ArneGoetje: freetype broke but I think asac fixed it [15:12] ? [15:12] fontconfig broke [15:12] Riddell: I'm not aware of that [15:13] Riddell: i did an emergeny update yes. [15:13] defoma broke it [15:13] asac: see his upload 2.6.0-1ubuntu13 taken from debian https://edge.launchpad.net/ubuntu/+source/fontconfig [15:14] huh? i uploaded it ;) [15:14] doh [15:14] ArneGoetje: ^^ [15:14] guess that was the idea [15:14] ;) [15:14] yes that's what my tab completion was ment to do [15:14] seb128: does Gnome use debhelper 7 for packaging yet? [15:15] 2.8.0 merge should get done at some point [15:15] at best with massive cleanup and droppage of most of our conf.d tweaks [15:15] like we have a bunch of dupe entries ... some before and some after user conf [15:17] asac: ok, so, 2.8 still needs to be done... btw, defoma is depreciated on debian. all font packages are being fixed atm. [15:17] yes [15:17] defoma wasnt my idea. ask TheMuso why he requested a sync ;) [15:18] anyway. thats probably the reason why we now dont fail if defoma isnt there [15:18] asac: maybe he didn't know it is depreciated now? [15:18] yes. but deprecated is != removed from debian [15:18] so i dont see why not to sync it [15:19] fontconfig should go up to 2.8.0... maybe requesting a sync and then adding step by step what we want based on bugs would be the right idea to do a thorough cleanup ;) [15:19] asac: well, we hopefully won't need it anymore in the near future [15:19] yes. and fix i uploaded basically does that: do not fail if defoma isnt there ;) [15:20] asac: ok [15:20] which was also a debian NMU emergency before ;) ... so seems while deprecating defome debian broke it (which feels odd if they really wanted to deprecate, why do something like that?) [15:21] asac: we do need ot keep the lcdfilter patches in fontconfig [15:21] Riddell, no, we use cdbs [15:21] Riddell, I don't really see a reason to use dh7 [15:22] ArneGoetje: i dont know. my personal opinion is that unless someone gives real background information we shouldnt maintain that as we cannot even verify that it works or not [15:22] asac: I asked keybuk about it [15:22] did he give background? [15:22] or just "needed"? [15:22] asac: yes, he gave background [15:22] which is? [15:23] for proprietary apps? [15:23] asac: it's needed for the terminal applications [15:23] which terminal application shows the regression in case we drop it? [15:23] seb128: it's nicer in some ways, but I'm not convinced we need yet another build system. now I have to work out how to port our .pot generation to it :( [15:24] asac: probably gnome-terminal, maybe others too [15:24] asac: keybuk said we want to keep the patches [15:24] i doubt gnome-terminal [15:25] keeping patches without being able to verfiy that they work doesnt make sense [15:25] we need better background [15:25] like a bug with a step by step how to reproduce the bug fixed by this [15:26] asac: well, just remove /etc/fonts/conf.d/11-lcd-filter-lcddefault.conf and 53-monospace-lcd-filter.conf and see the difference in gnome-terminal [15:27] do you see any difference? [15:27] asac: maybe ask keybuk about the deatils? [15:28] thats what you should do. you can also maintain those changes infinitely ;) ... i am just saying, that i wouldnt do that based on the input i got from him and you so far ;) [15:29] lcddefault isnt legacy btw [15:29] anyway. have to do something ... then call [15:31] asac: well, whatever... I personally dont see any difference on my screen, but it may be required only for some cases... I'll ask keybuk again [15:32] asac: when do you have some time to hand over the po2xpi stuff for langpack-o-matic to me? [15:32] yes. a bug and testcase would be good. maybe it doesnt even work anymore [15:32] ArneGoetje: what does hand over mean? i definitly should move the .xpi directories in the langpack o matic home [15:32] anything else you want? [15:33] asac: well, shall I maintain it or not? [15:34] ArneGoetje: yes. definitly. We have a bzr branch. i would suggest to maintain it through merge requests until you feel comfortable with the code base [15:34] asac: ok... do you have a link for me? [15:34] we can move the install to a location where you can access it, but i would prefer to see changes first getting signed off thorugh merge request [15:34] e.g. you request merge request. i sign off ... you roll out to rookery [15:35] yes [15:35] https://code.edge.launchpad.net/~mozillateam/rosetta/po2xpi [15:35] i think not everything is committed [15:36] so better wait before you jump on it [15:36] asac: sure, we can do that... we need a fix, the zh package split doen't work for Lucid, while it does for Karmic. I suspect the version number parsing 10.04 vs. 9.10 doesn't work like it should [15:36] but changes are not that considerable that arent committed [15:36] hmm. could be ... [15:37] asac: I needed to ifix it in lp-o-matic too... see my latest commit [15:47] asac: new mozilla xpi tarballs uploaded to rookery, btw: /home/arne/public_html/mozilla_xpis_3.0.16.tar.bz2 and /home/arne/public_html/mozilla_xpis_3.5.6.tar.bz2 [15:48] asac: compared to 3.0.17 and 3.5.7 the translations haven't changed btw [16:15] asac: /home/asac/mozilla-rosetta/po2xpi/data/ doesn't have a 10.04 subdirectory and therefor no merge-hints.txt file. That's why the firefox/xulrunner translations in lucid are missing for zh-han{s|t}. [16:20] asac: could you please add the 10.04 directory, so that I can build new -base langpacks? Thanks. [17:01] pitti, ccheney, etc... does anyone know if anyone was able to help that teacher from France who wanted to log directly into a guest session? [17:02] he had emailed ubuntu-devel-discuss [17:07] rickspencer3: I replied to a question like that a while ago, but I don't know when or to whom any more [17:14] just out of curiosity what is the way to do that? [17:14] rickspencer3: ? [17:15] tgpraveen12, they want their students to use a bootable usb stick that boots into a guest session (I think) [17:15] oh [17:15] they don't boot from the USB stick, only save their docs [17:16] rickspencer3: is there a way to set my ubuntu install to start in guest mode always? [17:16] so just want to be able to log in as guest from GDM [17:16] tgpraveen12, that's what I'm asking [17:16] oh, that would be an interesting feature for users-admin too === andreasn__ is now known as andreasn [17:57]