[04:54] Good morning [05:03] pitti: moin [05:04] hey desrt, how are you? [05:18] pitti: bad. [05:19] the inotify manpage lists lots of limitations, caveats, inherent races and downright bugs that will ruin your day [05:19] and it's incomplete :( [05:19] * desrt has been finding a few other issues.... [05:20] * desrt is finding that inotify has an unfortunate interaction with O_TMPFILE [05:22] desrt: hm, I have a similar feeling about wetware :/ [05:24] inotify is nice enough to tell you about 'write close' after you close() a writable file descriptor on a file [05:24] which is great... because open() writable of a new file gets reported as a create [05:24] so you can know not to look at the file as soon as you see a CREATED event [05:25] since the process still have to write() write() write() close() it before there is interesting content there for you [05:25] and this even deals nicely with the usual 'safe' paradigm of creating files by writing to a temp file, calling fsync() and then renaming it into place [05:26] since you get a MOVE_TO event, which is distinct from the CREATE event [05:26] so you know that you don't need to wake for CLOSE_WRITE [05:26] when using O_TMPFILE, though, you get a CREATE event when you link the final result into place [05:27] and since the original file was created anonymously, it doesn't have a dentry pointing back at the directory [05:27] so you get no CLOSE_WRITE when it closes [05:30] same problem applies to link() in general as well as mkdir(), symlink(), bind(), ... [05:32] at least with mkdir(), bind(), mknod(), etc. you can inspect the file to see that it is special... link() is trouble, though [06:57] good morning [07:09] hey didrocks [07:10] evening darkxst [07:11] yes! a nice one for once, finally stopped raining ;) [07:13] bonjour didrocks, ça va ? [07:13] hey darkxst [07:14] didrocks: happy systemd day! [07:16] hey pitti, how is it systemd day? [07:16] darkxst: today and tomorrow is a virtual systemd porting sprint [07:16] http://pad.ubuntu.com/systemd-porting-sprint [07:16] darkxst: see https://lists.ubuntu.com/archives/ubuntu-devel/2015-January/038635.html [07:19] pitti, oh I see, I have a rather large backlog of emails atm, so must have missed that [07:19] although I think most of gnome is in pretty good shape, except maybe gdm, but that is fixed in debian already I believe [07:20] (and in no way critical, it mostly works as is) [07:20] pitti: happy systemd day as well! :) [07:20] didrocks: I've got ubiquity and oem-config working \o/ [07:21] pitti: oh nice! was it hard or it was just complex to test? [07:21] didrocks: writing the units was relatively simple; it's just uber-hard to test [07:21] pitti: yeah, I guessed so :) it reminds me when i had to fix casper and test it live during christmas 2009 :p [07:22] didrocks: i. e. break=casper-bottom, chroot, manually call dhclient, use netcat to transfer your .debs, etc. [07:22] I learned more than I wanted on casper-bottom scripts that I always wanted to :p [07:22] didrocks: I noticed that the machine-id-commit job fails on the live system [07:22] waow netcat! [07:22] fun :) [07:22] oh? [07:22] didrocks: yeah, scp doesn't work on the crappy initramfs term [07:22] didrocks: it's apparently not fatal, but a bit of a wart [07:22] pitti: ok, did you open a bug on machine-id-commit? [07:22] no, not yet [07:23] brb [07:23] please do and assign me, I'll have a look after this systemd days [07:41] hey didrocks pitti desktopers [07:41] morning seb128 :) [07:45] bonjour seb128, ça va ? [07:46] seb128: nous avons essayé ton schnaps à week-end avec nos voisins -- c'est delicieux ! [07:46] seb128: j'aime les marons, avec beaucoup des herbes [07:46] "herbs" [07:47] lut pitti, ça va bien ! [07:47] super ;-) [07:47] pitti, "herbes" ;-) [07:48] but you say "d'herbes" [07:48] seb128: ah right, the old trap [08:12] didrocks: bug 1411140 [08:12] bug 1411140 in systemd (Ubuntu) "systemd-machine-id-commit.service fails on live system" [Low,Triaged] https://launchpad.net/bugs/1411140 [08:12] thanks! :) [08:45] morning all [08:45] hey willcooke! [08:49] hey willcooke [08:49] morning willcooke! [08:49] hi seb128! [09:00] OT: Is Digital Ocean any good? [09:00] Are* [09:04] hullo [09:04] what up Laney [09:04] There are some men at the end of the road installing a new phone cabinet [09:05] I think our village might be about to enter the 21st century [09:05] hey lane [09:05] hey Laney [09:05] grrrr [09:05] there is a crumbs under my tab key [09:05] -s [09:05] :D [09:05] willcooke: it's a sponsor of LAS, that's all I can say about them :) [09:05] morning Laney! [09:07] village life eh [09:07] being in Alconbury over Christmas was like the dark ages internet wise ... [09:07] ha! I had to get an ISDN2 installed when we lived in Alconbury [09:07] the WRT54G that I gave them 10 years ago is still working fine, which is amazing [09:07] hey seb128 & didrocks [09:08] Solid router that [09:08] morning Laney! [09:09] yo larsu [09:09] how's it going? [09:09] great! How about you? [09:12] pretty good [09:12] climbing's been in the news this morning ;) [09:14] Laney, only time climbing gets in the news here is when bush walkers fall off a cliff [09:17] http://www.climbing.com/news/the-dawn-wall-goes-free/ [09:20] nice unity SRU to trusty ;-) [09:26] Trevinho, ^ nice work! :-) [09:27] +1 thanks Trevinho === greyback__ is now known as greyback [10:40] willcooke: thanks (and to seb too :)) [10:42] * Laney gets some mysterious "jayatana" package installed on his system ... :) [10:43] Laney: it's nice, it gets the appmenu on java apps [10:43] ya [10:43] Laney: I've been using it for some time, it works well [10:44] need to find something javaish to use [10:45] Laney: well... jdownloader or... any simpler app example like the standard jEdit [10:46] Laney: never heard about it :p [10:46] Laney: I guess it would be nice to make possible to trusty users to get it without a ppa also, but... It's not possible to add new packages at this point, right? [10:47] didrocks: the changelog denies you :P [10:48] lies! [10:48] :) [10:59] mdeslaur: do you want to handle https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/1366790 as a security update? [11:00] Launchpad bug 1366790 in gtk+3.0 (Ubuntu Trusty) "Fix for CVE-2014-1949 (GTK 3.10.x)" [Undecided,Confirmed] [11:00] marga: thx for the debdiff [11:00] Laney, yw [11:01] Laney, I wasn't aware until recently that it affected gnome-screensaver as well. [11:23] is it just me, or are there more compiz/unity crashes recently? [11:24] hm, I now have a firefox window in the background that I can interact with. But I can't move it to the foreground and it doesn't appear in alt-tab [11:27] attente_, hi, do you mind if I land lp:~mitya57/unity-gtk-module/lp1383382, or do you have other pending unity-gtk-module changes? [11:35] mitya57: is this still an issue for you (the transparent stacks part)? http://bazaar.launchpad.net/~ubuntu-art-pkg/ubuntu-themes/trunk/revision/343 [11:35] I can't reproduce it... [11:36] also, I got a messaging menu entry for new sms and new phone call after connecting to an irc server in polari. yay telepathy! [11:36] larsu, oh, no, that is not needed with new Gtk+ [11:36] noon [11:36] That property existed only in 3.10 AFAIK [11:36] mitya57: which property? [11:36] -GtkWindow-decoration-button-layout [11:37] oh, you were asking about another part [11:37] mitya57: ah okay I can remove that as well. I was talking about making the stack in header bars transparent [11:37] right. Weirdly in the same commit :P [11:37] let me try [11:37] thanks [11:37] widgets are transparent by default [11:38] and I can't find a place that overrides the background for stacks [11:39] larsu, I can't reproduce that issue with polari anymore [11:40] mitya57: ok, I'll remove the rule for now. Thanks for testing! [11:43] larsu, btw, what do you think about lp:~l3on/ubuntu-themes/gnome-shell-fixes ? [11:44] mitya57: hm, lots of unrelated commits [11:44] some of which are already merged (like the gtk-icon-source stuff) [11:45] we definitly need the shadows [11:45] I haven't added them yet because I want to keep these windows look as broken as possible [11:45] so that more people complain to Trevinho ;) [11:46] Some parts of that branch are quite important for GNOME Shell/Flashback, but I can't say anything about other parts [11:46] there's also some changes for gedit? [11:46] would be nice if those were separate merge requests [11:46] also, the branch doesn't cleanly merge into trunk anymore [11:47] For gedit, nautilus, gnome-terminal and unity-greeter... [11:48] Probably I need to cherry-pick the stuff I need (i.e. gnome-panel.css) and create a separate MP for that [11:50] larsu: eheh [11:51] Trevinho: just messing with you ;) [11:51] (still want that patch, though) [11:56] larsu: i know... once I'll make my hands dirty with compiz soon === MacSlow is now known as MacSlow|lunch [12:26] Laney: yeah, I'll handle it as a security update [12:26] great [12:26] Laney: thanks [12:28] mdeslaur: we have lp:~ubuntu-desktop/gtk/ubuntugtk3trusty btw [12:28] if you can push there after it's uploaded [12:29] Laney: sure [12:29] rock on [12:29] [12:29] hehe [12:34] mdeslaur, Laney, I'm lacking backlog but I guess you are speaking about bug #1366790 ? [12:34] bug 1366790 in gtk+3.0 (Ubuntu Trusty) "Fix for CVE-2014-1949 (GTK 3.10.x)" [Undecided,Confirmed] https://launchpad.net/bugs/1366790 [12:34] seb128: yes [12:34] great [12:34] * seb128 cross for his list === alan_g is now known as alan_g|lunch === MacSlow|lunch is now known as MacSlow === alf is now known as alf_ === alan_g|lunch is now known as alan_g [13:31] ochosi, Laney, seb128: I'd very much appreciate some testing on this branch: https://code.launchpad.net/~larsu/ubuntu-themes/stop-using-unico/+merge/246566 [13:31] * larsu tried lots of apps, but it's quite a change [13:31] (even though not that much css changed actually) [13:32] hm: Launchpad encountered an internal error during the following operation: generating the diff for a merge proposal. It was logged with id OOPS-3fcd77bfd8f5bd561dee1cdd48f40c79. Sorry for the inconvenience. [13:32] larsu: did you try some swing apps like eclipse which are using a gtk theme? [13:32] didrocks: no, good point, thanks [13:33] larsu: refresh, I can see the diff [13:33] me too [13:33] 1742 lines (+74/-1478) 8 files modified [13:33] got the email anyway [13:33] weird [13:33] nice! [13:33] didrocks: most of that is removing a file that wasn't used (but had lots of references to unico) [13:33] larsu, sure can try, but I'm probably not the best to notice minor stylish changes since I don't usually pay much attention to details [13:34] larsu: shut up and take the credit dude! ;) [13:34] didrocks: HAHA ok :) [13:34] :p [13:35] larsu: btw, any reason you still use so many pixmaps, e.g. for buttons? [13:35] seb128: you're the only one I know that uses radiance. Just look out for something that's really off (probably enough if you wait for the merge) [13:35] larsu, ok [13:36] ochosi: (a) history (b) performace (c) I don't want to be the one recreating them pixel perfectly with css gradients [13:36] larsu: really? pixmaps are faster than css? [13:36] the other two are perfectly understandable [13:36] i just didn't know about (b) [13:36] ochosi: yes, quite a lot [13:37] that sort of sucks tbh [13:37] but still, good to know [13:37] ochosi: blitting an image is much faster than asking cairo to draw a bunch of gradients and borders and shadows... [13:38] anyway, i'll give it a shot! [13:38] thanks! [13:40] anything to look out for specifically or just a general checkup to see whether anything is off [13:40] ? [13:40] ochosi: focus rings [13:41] you only see them when you keyboard-navigate [13:41] bonan matenon, ubuntanoj [13:41] yeah, just wanted to say the new open-file dialog really sucks big time in that respect [13:41] hm? [13:42] oh, just in general - not related to the theme? [13:42] adwaita does a slightly better job there [13:42] it has that dotted outline inside some widgets [13:42] mostly listviews [13:42] it has them in all widgets [13:42] it's much better [13:42] still, not very visible.. [13:43] yeah, it has them everywhere, but with buttons i see no issue in Ambiance, that's quite obvious [13:43] with listviews, there's just no focus outline whatsoever [13:43] yeah :/ [13:43] but as i said, even in adwaita it's too subtle to be really clearly visible [13:44] it's a bit inconsistent that menus are dark and popovers are bright [13:44] (sorry, i know that's not what you asked me to look at...) [13:45] I agree. I gave them a bright background because popovers are more like dialogs than they are menus [13:45] the other focus outlines i could spot look good! [13:45] also, dark background on them is fugly [13:45] also, we do have bright menus as well, because design [13:46] mitya57: sure, lgtm [13:47] larsu: oh, you do? humm... :] [13:48] right, so from my brief look around i'd say the only problem i see are listviews [13:48] but those had the problem before, no? [13:49] need to check [13:50] indeed [13:50] ok [13:50] it's a problem, but not for this MR [13:50] thanks for testing! [13:51] i wonder why shotwell has such a messed up "status corner"... [13:52] there, the focus outline looks a bit meh [13:52] http://i.imgur.com/CVoe4M5.png [13:53] (but again, not a regression) [13:53] anyway, summing it up, good work there! [13:53] ochosi: this is actually not the focus outline, but the fact that the labels are selected by default when you tab to then [13:53] *them [13:53] I wonder why, that's really strange [13:53] right [13:53] yeah, that corner is theme-wise also broken [13:53] i mean it has that sunken-in style to it that doesn't really exist anymore in gtk3 [13:54] it even looks really bad with adwaita [14:06] seb128_, is it possible to poke phone settings from the command line? [14:06] mlankhorst has a broken screen and so cant get in to the settings app to disable the screen time out [14:06] willcooke, what do you want to do exactly? === seb128_ is now known as seb128 [14:08] set it to infite :p [14:11] willcooke: ok I think I have a hack for -2x [14:11] willcooke, mlankhorst, is that the "lock on idle" you want to change? === alan_g is now known as alan_g|afk [14:12] yeah [14:12] and screen turning off [14:14] mlankhorst, @ hack - cool! [14:14] you can use powerd-cli somehow to keep the screen on [14:14] I'm not 100% sure if it completely works since I get some corruption (maybe resize borked?), but probably good enough.. [14:18] willcooke: if you want to test https://mblankhorst.nl/etc/Xmir [14:18] thx mlankhorst [14:20] add -2x to the command line [14:20] ah! [14:20] I was about to ask :D [14:20] * mlankhorst is telepathic [14:20] lol [14:24] looks pretty good, a little bit of corruption to start, but thats fine [14:24] just need to get metacity to use a reasonable font size [14:25] oh right.. [14:29] willcooke: yeah looks like I need to do some hacks to fix initial screen size :P [14:29] that's why you're getting corruption [14:29] s'no biggy [14:29] or I should just clear everything instead.. [14:32] I think I will do that, it will some issues anyway.. [14:34] qengho, seb128 - can you guys point me at the main LP page for Chromium? There seem to be quite a few places. I'm looking for a specific commit [14:35] willcooke: You probably want https://code.launchpad.net/~chromium-team/chromium-browser/vivid-working [14:35] thx qengho [14:36] willcooke: That's packaging. [14:58] willcooke, what in chromium? https://launchpad.net/ubuntu/+source/chromium-browser has the versions/bugslist [14:58] seb128, sorted, thanks [14:59] yw! [15:08] larsu, I've still got this huge fonts issue in Xmir: https://lh4.googleusercontent.com/-0H4Gcl0ogpw/VLfXncNCsvI/AAAAAAAAJDM/bSsvO-dGj7k/w860-h879-no/1.png [15:08] any clues? [15:09] I'm using metacity as a WM [15:09] is it likely to be a metacity thing? [15:09] err, good question [15:10] the settings keys are set to something sensible I presume? [15:10] yeah, based on our quick conversation the other day I set them to 10 or 8 or something [15:11] how about org.gnome.desktop.wm.preferences titlebar-font [15:11] this is about big fonts in title bars, right? [15:11] * larsu has problems parsing that screenshot [15:25] larsu, sorry, otp - will check [15:31] Laney: , seb128: http://www.ubuntu.com/usn/usn-2475-1/ [15:31] ty! [15:31] yw [15:32] bregma: i would like a shortcut or command to force move all windows to display1 [15:33] bregma: or forexample force remove display [15:33] cause on display-port screen disconnect i often have windows "stuck" on the external monitor [15:33] on 14.04 LTS [15:33] * xnox ponders if turning mirror mode on and off would do the trick. [15:33] xnox, the compiz "put" plugin will do that [15:34] bregma: aha! that's what I need. [15:34] assuming it hasn't bitrotted, we don't normally enable it so it doesn't get a lot of testing [15:34] bregma: any way i can help with fixing the monitor disconnect bug though? [15:35] bregma: on reconnect it realises that it had windows on dead screen -> moves them to the first screen & initialises the external monitor afresh. [15:35] xnox, is the problem that X does not realize the monitor is gone? [15:36] bregma: correct - but i cannot tell you exactly. [15:36] mdeslaur, thanks [15:36] Laney: so ... our pure-css spinner doesn't work anymore in 3.14 (see #gtk for details if you're interested) [15:36] Compiz is relying on xrandr for screen size change events to adjust the window placement, no xrandr events to response [15:36] e.g. i see the windows in the alt-tab on the app, switching to one window of the app shwos animation switching to it -> by making it fly off the main screen into nothing. [15:36] Laney: do you know if we have a spinner icon somewhere? [15:37] Laney: could also use the one from adwaita - or the cool one from the phone [15:37] * larsu wonders if that looks good on desktop [15:38] bregma: i have openned ccsm but there is no put plugin =( [15:39] larsu: what's the phone one? [15:40] xnox, that's odd -- I see it under Window Management [15:40] bregma: hm, i see in the move plugin that I can do Alt-F7 and move the window to the other screen with keyboard, that would do for me. [15:40] grid, move, place, resize, scale, snapping [15:40] huh, well if move works, use it [15:41] Laney: /usr/lib/x86_64-linux-gnu/qt5/qml/Ubuntu/Components/Themes/Ambiance/artwork/spinner@30.png [15:42] oh this one [15:42] adwaita's in orange looks neat as well [16:06] willcooke: I think the current version should handle resize better, I'm clearing the entire buffer before copying the old contents. :P [16:06] mlankhorst, sweet [16:26] willcooke: can you redownload? I should have fixed the transparant corruption at the start [16:26] mlankhorst, thx [16:27] but I can't read the contents, if you want I should be able to add rotation though [16:27] mlankhorst, rotation would be super useful [16:28] lowriter on the screen's too tiny to read now :P [16:28] ha [16:31] working on doubled makes it easier to add rotation, same codepaths are affected. It's just a drawing call I need to fix and messing around with width/height [16:58] tons of if cases later.. [16:59] yeah, input doesn't rotate but otherwise it works :P [17:00] I'll try to hook up the full thing later [17:00] eod! [17:08] thanks a lot mlankhorst [17:54] * willcooke -> EOD [17:54] ta ta === alan_g|afk is now known as alan_g === tedg` is now known as tedg === meetingology` is now known as meetingology === mfisch is now known as Guest97735 === Ursinha_ is now known as Ursinha === Saviq_ is now known as Saviq === kgunn is now known as Guest4292 === broder_ is now known as broder === balloons is now known as Guest8667 === maxb is now known as Guest47523 === Pici is now known as Guest32967 === charles_ is now known as Guest52246 === larsu is now known as Guest6673 === TheMuso` is now known as TheMuso === dkessel_ is now known as dkessel === Zdra is now known as xclaesse === robru_ is now known as robru === DalekSec_ is now known as DalekSec