pitti | Good morning | 05:41 |
---|---|---|
mlankhorst | morning! | 07:16 |
didrocks | good morning | 07:17 |
larsu | morning! | 07:19 |
ochosi | morning folks | 07:20 |
didrocks | hey larsu, ochosi | 07:20 |
ochosi | larsu: quick question, where is your gtk3.14 staging PPA? (so i can start to prep the xubuntu vivid themes) | 07:20 |
larsu | ochosi: in the desktop team ppa | 07:22 |
didrocks | pitti: hey, welcome back! I hope you had a safe flight back :) | 07:22 |
didrocks | pitti: I saw that you were cooking this week-end | 07:22 |
didrocks | and pushing new systemd :) | 07:22 |
didrocks | pitti: I wanted to know if there was any reason you didn't include bug #1400682 patch? | 07:23 |
ochosi | larsu: oh indeed. note to self: don't start asking around *before* having had coffee... | 07:23 |
ubot5 | bug 1400682 in systemd (Ubuntu) "Add xdiagnose fallback when display-manager fails to start" [Undecided,New] https://launchpad.net/bugs/1400682 | 07:23 |
larsu | ochosi: haha, no worries ;) | 07:24 |
pitti | bonjour didrocks, ça va ! | 07:55 |
pitti | didrocks: yes, flight was uneventful, and I could sleep a bit too | 07:55 |
pitti | didrocks: the main reason was, I did the merge/update mostly on the plane, and I didn't have that patch downloaded yet :/ | 07:55 |
pitti | didrocks: but there is a test regresssion and a timedatectl crash on i386 anyway, so I'll need to upload a new version today anyway | 07:55 |
pitti | didrocks: I have that bug open now :) | 07:55 |
didrocks | pitti: perfect, thanks! | 07:59 |
didrocks | pitti: nice that you could sleep, how long was it btw? | 07:59 |
didrocks | I guess something like 9-10 hours? | 07:59 |
pitti | didrocks: 11 hours | 08:00 |
pitti | 18:55 to 5:30 on the way back, 22:55 to 10:50 on the way there (1 h time zone difference) | 08:00 |
didrocks | yeah, quite long, at least, no jetlag :) | 08:01 |
ochosi | larsu: quick question, is file-roller on your list of apps to remove CSD? | 08:11 |
mlankhorst | I should not be allowed to have smartphones :P | 08:14 |
didrocks | mlankhorst: breaking one? | 08:16 |
mlankhorst | yeah I cracked the screen on the aquaris | 08:16 |
didrocks | it's still a build machine at least | 08:17 |
mlankhorst | true but I was using it to test.. if I can figure out how to get the screen permanently unlocked and bright it would be fine | 08:18 |
didrocks | mlankhorst: you can maybe get it repaired? | 08:19 |
didrocks | like I don't think it's a particular screen compared to android versions | 08:19 |
larsu | ochosi: I thought seb already patched that?! | 08:20 |
larsu | ochosi: it doesn't have csd on latest vivid, or are you talking about a new version? | 08:21 |
mlankhorst | didrocks: is it possible to unlock the screen without using the touchscreen? I already have the phone setup | 08:21 |
mlankhorst | adb shell etc works | 08:21 |
ochosi | larsu: this is vivid with the ubuntu-desktop PPA (gtk3.14) and file-roller: http://i.imgur.com/n36Mjdz.png | 08:21 |
didrocks | mlankhorst: not that I know of | 08:22 |
ochosi | larsu: i'll switch mirrors, maybe mine is lagging behind... | 08:22 |
didrocks | mlankhorst: I would really try to go to a repair shop, the model from what I know is similar to the android one | 08:22 |
didrocks | so they should have that screen | 08:22 |
larsu | ochosi: which desktop? I have a normal title bar on unity | 08:23 |
ochosi | larsu: xubuntu | 08:23 |
mlankhorst | didrocks: it's identical, but yeah I guess I can take a look | 08:23 |
ochosi | larsu: did you only add a check for unity? | 08:23 |
ochosi | larsu: my guess is that both mythbuntu and ubuntu-studio (maybe even u-mate) would appreciate "!=gnome" instead of "==unity" | 08:24 |
ochosi | and lubuntu probably | 08:24 |
larsu | ochosi: it checks for GNOME: http://bazaar.launchpad.net/~ubuntu-desktop/file-roller/ubuntu/view/head:/debian/patches/bz_unity_headerbar.patch | 08:24 |
ochosi | weird | 08:25 |
didrocks | larsu: we really need a framework to tweak that on runtime, by capability I guess | 08:25 |
didrocks | like "support-overlay-scrollbar" | 08:25 |
didrocks | "support-appmenu" | 08:25 |
didrocks | "disable-foo" | 08:25 |
ochosi | larsu: not sure you noticed, but it has a normal menu plus a headerbar, so something is fishy there | 08:25 |
didrocks | (yeah, I know, I introduced this per-session name based in 2009/2010, but I regret now :p) | 08:26 |
ochosi | larsu: better visible here: http://i.imgur.com/waneudI.png | 08:26 |
larsu | didrocks: ya... we talked about having another xsetting, like we do for global menu and buttons in dialog headers | 08:27 |
didrocks | larsu: yeah, that would make sense | 08:27 |
larsu | didrocks: consensus so far has been "oh no, not another xsetting" | 08:27 |
didrocks | as there is no phone, tablet or whatsoever | 08:27 |
didrocks | done properly, there is "touch screen", "keyboard", "screen size", "battery" | 08:27 |
larsu | ochosi: probably because gtk-shell-shows-menubar is 0 | 08:27 |
didrocks | and changing the app/framework behavior on that | 08:27 |
didrocks | larsu: did you try to compare that to media queries? | 08:28 |
larsu | didrocks: compare what? | 08:28 |
larsu | interesting idea, though... | 08:28 |
didrocks | like, media queries are really trying to do that and base on capability rather than form factor/session | 08:28 |
larsu | right | 08:28 |
didrocks | larsu: just taking web media queries as an example as part of that discussion | 08:28 |
didrocks | as an argument :) | 08:28 |
didrocks | larsu: not the most important thing right now, but this discussion just makes me thinking that :) | 08:29 |
larsu | didrocks: makes a lot of sense to me. We'd still need something like xsettings to provide the values, though | 08:30 |
=== ara is now known as Guest73317 | ||
didrocks | larsu: yeah, that's clearly one of the way to achieve this | 08:30 |
didrocks | just talking about the general concept, before we introduce an unity8-desktop for instance :) | 08:30 |
didrocks | finding a grub script typo -> *not* fun | 08:31 |
larsu | the unity8 guys have a plan | 08:31 |
larsu | I don't know what that is, though | 08:31 |
didrocks | we should follow the same one for desktop for sure (not implementation-wise, but for coherence) | 08:31 |
didrocks | anyway, sorry for side-tracking :) | 08:32 |
didrocks | waiting for grub2 to pass tests (again :p) | 08:32 |
ochosi | larsu: right, so should i report a bug about this? | 08:33 |
larsu | didrocks: haha, no worries :) | 08:34 |
larsu | ochosi: I'm not sure what the issue is? | 08:34 |
larsu | you don't have a global menu in xubuntu, do you? | 08:35 |
larsu | so having that xsetting set to 0 is correct, and you get a local menu | 08:35 |
ochosi | larsu: ehm, the issue is that file-roller has CSD..? | 08:35 |
ochosi | the menu part is fine | 08:35 |
ochosi | it was just confusing for me that there could be a mix | 08:36 |
larsu | I don't understand how that is confusing. If you tell gtk that you want a header bar and a menu inside the window, that's what you get | 08:36 |
larsu | what's in your XDG_CURRENT_DESKTOP? | 08:36 |
ochosi | larsu: it's XFCE | 08:40 |
ochosi | sry, g2g, bbl | 08:43 |
larsu | ochosi: ah, it is a bug in the patch | 08:43 |
larsu | I can reproduce | 08:43 |
larsu | ochosi: which version do you have installed? The patch I linked to is not yet in vivid | 08:46 |
larsu | http://bazaar.launchpad.net/~ubuntu-desktop/file-roller/ubuntu/revision/150 | 08:46 |
larsu | Laney: is file-roller file-roller blocked on anything? | 08:46 |
willcooke | morning | 08:58 |
didrocks | hey willcooke, feeling better? | 09:01 |
Laney | heya | 09:02 |
didrocks | morning Laney, good week-end? | 09:02 |
larsu | morning Laney, willcooke | 09:03 |
willcooke | didrocks, not really. Paracetamol helps | 09:04 |
didrocks | urgh :/ | 09:04 |
didrocks | willcooke: take it easy, soon holidays! | 09:05 |
willcooke | heh # | 09:05 |
Laney | hey didrocks et larsu | 09:05 |
willcooke | looking forward to it | 09:05 |
Laney | not bad thanks, did some xmas shopping ;-) | 09:05 |
Laney | you guys? | 09:06 |
Laney | larsu: file-roller> not sure offhand, did robert-ancell do that one maybe? | 09:06 |
didrocks | played some nice board games and finished watching breaking bad :) | 09:06 |
larsu | Laney: don't know, thought you might | 09:06 |
Laney | still haven't played that game I got in DC :( | 09:06 |
Laney | larsu: I'll take a look | 09:07 |
larsu | thanks | 09:08 |
larsu | the one we have now checks for Unity to disable csd | 09:09 |
larsu | the new one checks for !GNOME | 09:09 |
Laney | they did that upstream? | 09:09 |
larsu | no, robert_ancell did (and you committed it apparently) | 09:09 |
larsu | http://bazaar.launchpad.net/~ubuntu-desktop/file-roller/ubuntu/revision/150 | 09:10 |
Laney | i see | 09:10 |
Laney | "(blocked on gtk 3.14)" | 09:11 |
darkxst | Laney, about everything is blocked on gtk 3.14 now (for us anyway!) | 09:11 |
didrocks | pitti: I hope I'm using git-dpm well and patching the grub patch to generate other installed init system FYI | 09:12 |
Laney | darkxst: ya, lots for us too, might be that we can deal with the remaining issues post upload | 09:20 |
Laney | I need to test ubiquity though | 09:20 |
ricotz | Laney, hi :), just a small request, could you define the mir minimum-requirement in the gtk+3.0 packaging? | 09:31 |
Laney | what is it? | 09:32 |
pitti | didrocks: sure -- if it looks ok and dpkg-source is happy, it's good :) | 09:37 |
ricotz | Laney, idk, at least the trusty version is not sufficient | 09:38 |
Laney | I could make it >= what-we-have-now | 09:38 |
Laney | you probably just want to disable that for the PPA | 09:38 |
ricotz | i know and i will, still would be nice to keep track on such things | 09:40 |
ricotz | also there is a 3.14.6 ;) | 09:42 |
Laney | yep | 09:44 |
Laney | larsu: I see you got an ok for your icon patch ;-) | 09:57 |
larsu | Laney: I dumbed it down considerably with only 2 lines left | 09:58 |
Laney | ya | 09:58 |
Laney | seems this is acceptable | 09:59 |
larsu | speaking of which - virtmanager has large buttons because one of the toolbar icons is 48x48 | 09:59 |
* larsu figured that out after inspecting css for 20 minutes | 09:59 | |
Laney | what about the borders? | 09:59 |
larsu | borders have always been on those, no? | 10:01 |
larsu | I can remove them if you prefer | 10:01 |
Laney | don't think so | 10:02 |
Laney | one sec, lemme boot utopic | 10:02 |
Laney | larsu: http://people.canonical.com/~laney/weird-things/virt-manager.png trusty | 10:06 |
Laney | http://people.canonical.com/~laney/weird-things/virt-manager-new.png that's with 3.14.5 | 10:07 |
larsu | I stand corrected. Thanks. | 10:08 |
desrt | Laney: you never told me how great schroot is | 10:10 |
Laney | desrt: dude, welcome to the enlightened | 10:10 |
desrt | i jhbuild on fedora on my ubuntu linode | 10:11 |
desrt | yo dawg and all that | 10:11 |
larsu | schroot? | 10:11 |
larsu | desrt: good ... err ... morning? | 10:11 |
desrt | hi :) | 10:11 |
larsu | you're up early | 10:11 |
desrt | larsu: it's a system for maintaining images of other distros that you can chroot into | 10:12 |
desrt | it has some extremely lovely features | 10:12 |
larsu | coool | 10:12 |
desrt | like the way i'm using it now, it unpacks a fresh .tar each time | 10:12 |
desrt | so i can explode it however i want | 10:12 |
desrt | and when i logout, it deletes it | 10:12 |
desrt | it also has a lot of really great features like automatically bindmounting /sys /proc /dev and (insanely usefully) /home | 10:12 |
darkxst | larsu, jhbuild is a major hack, schroot would be better, but it helps finding ubuntu issues | 10:13 |
desrt | i've used it from time to time when Laney wants me to fix a bug on some weird arch or another but i only really "learned" it this weekend | 10:13 |
desrt | i should done that years ago | 10:13 |
desrt | *shoulda | 10:13 |
didrocks | desrt: you can even mount the overlay as tmpfs to avoid the unpacking :) | 10:13 |
desrt | didrocks: linode has no aufs support in their kernels :( | 10:13 |
didrocks | argh, so switch to btrfs and use nspawn coming with systemd 219 ;) | 10:14 |
desrt | meh | 10:14 |
desrt | containers are too heavy | 10:14 |
desrt | i also really really really like how desrt on the outside is desrt on the inside | 10:14 |
desrt | it's totally perfect | 10:14 |
larsu | darkxst: this seems to be for a different use case than jhbuild... | 10:14 |
didrocks | desrt: ahah, nicely said :) | 10:14 |
larsu | (which I'm actualy quite fond of these days) | 10:14 |
* desrt did a pile of jhbuild hacking over the weekend | 10:15 | |
darkxst | larsu, yes GNOME upstream has no sbuild/schroot stugg | 10:15 |
desrt | i added support for finding out about _all_ of the things that are needed to build the lower bits of gnome | 10:15 |
desrt | including the stuff that we couldn't previously add checks for (libxml2-python, docbook-xsl) | 10:15 |
desrt | and now i have a new mode that dumps the full list in machine readable form like pkgconfig:egl,c_include:jpeglib.h,... | 10:16 |
darkxst | desrt, like making system modules satisfy jhbuild? | 10:16 |
desrt | and another script that runs that list against a Contents.gz from apt-file and comes up with various solutions | 10:17 |
desrt | darkxst: ya... | 10:17 |
willcooke | desrt, you're up early | 10:17 |
Laney | or late ... | 10:17 |
desrt | early :) | 10:17 |
desrt | Laney: maybe you could help me with some of the issues i hit | 10:17 |
desrt | i asked in #debian but it was pretty slow over the weekend | 10:17 |
Laney | #debian-devel might be better for that kind of q | 10:18 |
desrt | first issue is that i want to find a way to effectively say: apt install a b 'c | d' e f 'g | h' | 10:18 |
desrt | where c/d are alternatives for the same functionality and g/h the same for another | 10:18 |
desrt | like jpeg vs. jpeg-turbo or something | 10:18 |
desrt | currently my script takes all of the possibile combinations of the alternatives, does apt dry-runs on them all and calculates a 'cost' based on the number of reported Inst and Remv lines (with removals costing 1000 times more than adds) | 10:19 |
desrt | then it picks the lowest cost | 10:20 |
Laney | I'd build a .deb with that in Depends: and let apt figure it out | 10:20 |
desrt | unfortunately that involves doing apt dry-runs of ~1000 operations over 192 possible combinations... which takes 5-10 minutes :( | 10:20 |
desrt | Laney: ya.... i was hoping for a better way | 10:20 |
desrt | more direct | 10:20 |
Laney | that's what sbuild does when it resolves build-deps, for example | 10:21 |
desrt | also: Depends: lines are a bit "opinionated", right? | 10:21 |
Laney | or there's a script called mk-build-deps | 10:21 |
desrt | like, the thing on the left is the one that is taken unless the others are already installed, no? | 10:21 |
Laney | ya | 10:21 |
Laney | well, kind of, if it's not installable then it'll try the other one | 10:21 |
desrt | does it do that even if the 'cost' is high? | 10:21 |
desrt | like for example say i had libfooa and libfoob mutually conflicting | 10:22 |
desrt | they have corresponding libfooa-dev and libfoob-c | 10:22 |
desrt | *-dev, sorry | 10:22 |
desrt | then package 'c' depends on libfooa-dev | libfoob-dev | 10:22 |
desrt | and my system had libfoob installed | 10:22 |
desrt | will apt uninstall libfoob in order to get me libfooa-dev just because i didn't have libfoob-dev? | 10:22 |
desrt | because in that case i'd sort of hope for the more reasonable "oh... he's already got libfoob... so probably he just wants the -dev of that"... which is what my 'minimum cost' thing was tilting for | 10:24 |
desrt | you know.. mk-build-deps is not bad...... and it would additionally give me an artifact that i could install in order to force all of the depends of my jhbuild to stay installed | 10:25 |
Laney | I think apt is smart enough to prefer not to remove packages if it can | 10:26 |
Laney | but you get into mvo territory at this point | 10:26 |
desrt | i already sent him an email :) | 10:26 |
Laney | or you could do some experimentation | 10:27 |
desrt | i'll see if i can play with mk-build-deps | 10:27 |
desrt | ya... | 10:27 |
desrt | see what it gets me | 10:27 |
Laney | 'equivs' is a tool for building packages which don't do anything other than have metadata | 10:27 |
Laney | so you can test using that | 10:27 |
desrt | great advice. thanks. | 10:27 |
Laney | or even use it to generate your fake package for dep solving | 10:27 |
desrt | i had another issue, but it's less likely you know about it (or even that it has a solution at all) | 10:27 |
desrt | i want a command that i can call that will answer the question "which package do i have to install in order to get a given uri into the system xml catalog" | 10:28 |
desrt | ie: i want to know what will make 'xmlcatalog /etc/xml/catalog http://docbook.sourceforge.net/release/xsl/current/' return a positive result | 10:29 |
desrt | (my current thinking towards a best solution here is to hardcode the package name 'docbook-xsl'... i don't plan on the xml depends of jhbuild undergoing a dramatic expansion any time soon... but a generic solution would be nice...) | 10:31 |
Laney | sorry, had a guy around checking the energy efficiency of the house... gone now... | 10:39 |
desrt | very nice of him | 10:40 |
Laney | I think that particular operation is done by packages invoking update-xmlcatalog in their mantainer scripts which are usually generated by dh_installxmlcatalog during the build | 10:40 |
Laney | soooo no I don't know of a way you could determine that :) | 10:40 |
desrt | ya.. i'm aware of that much, but it's sort of the opposite of what i want | 10:40 |
desrt | i'd need somewhere a database of all the people who ever do that | 10:40 |
desrt | in general, we need to start getting better information about this sort of stuff into apt... | 10:41 |
* desrt wants to apt install pkgconfig:gl or apt install c_include:jpeglib.h or apt install python2:libxml2 | 10:41 | |
desrt | or (xml:http://whatever) | 10:42 |
desrt | Laney: is there a better way to install-with-deps than using apt-get install -f? | 10:45 |
desrt | the proposed solution in my contrived case above is removing the package with the alternative depends | 10:46 |
Laney | maybe gdebi, but I'm not sure what resolver it uses | 10:46 |
Laney | or make a local repository using dpkg-scanpackages and put that into sources.list | 10:46 |
Laney | ISTR some talk about apt getting "apt install .deb" | 10:47 |
desrt | Laney: long overdue, that one.... | 10:48 |
Laney | DonKult @ #debian-devel might know more - he's an apt guy | 10:49 |
desrt | Laney: the idea of adding a local repository is interesting -- i don't even actually need the .deb files in that case | 10:51 |
desrt | can just ask apt "what would you do" based on the Packages file | 10:51 |
desrt | in any case, the result of the experiment is that apt does 'the wrong thing' | 10:51 |
desrt | (ie: it wants to remove libfoob) | 10:55 |
Laney | ubiquity seems ok with new gtk | 11:06 |
Laney | closing in on the upload | 11:11 |
xnox | \o/ Laney | 11:29 |
Laney | xnox: remembered to check it this time ;-) | 11:31 |
Laney | ps. nice christmas tree | 11:31 |
xnox | =)))) | 11:31 |
Laney | shared mine with you on g+ | 11:33 |
xnox | Laney: haven't been on g+ for a while. | 11:37 |
xnox | .... lennart invited me to systemd hack-fest. | 11:37 |
Laney | haha | 11:37 |
Laney | it seemed like a reasonable option from the android share menu | 11:37 |
xnox | pitti: are you going for lennarts/systemd fosdem hackfest? | 11:37 |
pitti | xnox: yes, I will | 11:38 |
xnox | didrocks: as well and marco d'Itri. | 11:39 |
xnox | pitti: i'll ponder about it. | 11:41 |
Laney | larsu: do you need to get someone to review https://code.launchpad.net/~larsu/ubuntu-themes/gtk-314 ? | 11:59 |
larsu | Laney: yes, but I've just started working on it again | 12:09 |
larsu | seb128 tried it last week and didn't notice any issues | 12:09 |
Laney | yes I've been trying it too | 12:09 |
Laney | not hugely qualified to review though | 12:09 |
larsu | that's fine, I'm not hugely qualified to write it ;) | 12:10 |
Laney | if we're going to upload gtk then this needs to go in too | 12:10 |
larsu | yep | 12:10 |
* larsu is fixing toolbar buttons right now | 12:11 | |
Laney | ok, please put up a MP when that's done :) | 12:11 |
larsu | "done". lol. | 12:12 |
Laney | I believe in you! | 12:12 |
larsu | haha - there are little issues all over the place | 12:12 |
larsu | and changing something in the theme _always_ breaks something else | 12:12 |
larsu | we need some way to properly test theming | 12:13 |
Laney | it doesn't have to be 100% to get it in vivid | 12:13 |
larsu | of course | 12:13 |
larsu | let me finish the toolbar/flat button stuff and then I'll MR | 12:13 |
Laney | did you notice that the session selector in unity-greeter got a border too? | 12:14 |
Laney | :) | 12:14 |
larsu | I did now :) | 12:15 |
Laney | put it on the pad | 12:15 |
Laney | not a priority | 12:15 |
larsu | yep | 12:15 |
larsu | thanks | 12:15 |
=== pstolowski is now known as pstolowski|lunch | ||
pitti | didrocks: FYI, I sent a fix for the timedatectl regression upstream and fixed test_profile in Debian git; I'm now trying to reproduce the NSpawn.test_boot failure (doesn't happen locally) | 12:22 |
pitti | didrocks: so uploading a new version will still take a bit, but I'll see to getting it done by tomorrow | 12:23 |
pitti | I'll be off for the afternoon | 12:23 |
desrt | pitti: take care | 12:25 |
pitti | hey desrt, how are yoU? | 12:25 |
desrt | busy :) | 12:25 |
didrocks | pitti: thanks for keeping me posted, now that I sent the grub thing, I'm about to finish some desktop developer stuff. Then, moving on on more systemd work | 12:25 |
didrocks | pitti: still want to discuss with you about fsck before pushing more :) | 12:25 |
didrocks | but let's do that tomorrow | 12:25 |
desrt | pitti: you going to the systemd hackfest in brussels? | 12:26 |
pitti | desrt: (OTP) yes, I will | 12:30 |
pitti | desrt: will you be on fosdem, too? | 12:30 |
desrt | yes | 12:30 |
desrt | hopefully will get some kdbus stuff nicely tied up there | 12:31 |
willcooke | didrocks, cheapo ebay laptop battery update.... wont charge above 60% now. | 12:41 |
didrocks | willcooke: yeah, so the cheap was really "cheap" :/ | 12:42 |
willcooke | quelle surprise | 12:43 |
didrocks | well, see the positive side: it didn't explode when you plugged it | 12:43 |
willcooke | :D | 12:43 |
didrocks | ;) | 12:43 |
larsu | yet | 12:43 |
willcooke | LOL! | 12:43 |
didrocks | roh | 12:44 |
pitti | didrocks: I responded to bug 1400682 | 12:51 |
ubot5 | bug 1400682 in systemd (Ubuntu) "Add xdiagnose fallback when display-manager fails to start" [Undecided,New] https://launchpad.net/bugs/1400682 | 12:51 |
xnox | Laney: i don't see any pictures shared with me on G+... did you use the right account instead of a defunct @canonical.com one? | 12:53 |
Laney | don't know, I typed your name in the app ^o) | 12:53 |
didrocks | pitti: agreed, doing this change and closing the bug report then | 12:55 |
Laney | xnox: https://plus.google.com/u/0/109160032876474505377/posts that one | 12:55 |
xnox | Laney: right - so I see if it I go to your page, but I didn't get a notification about it =/ | 12:56 |
xnox | Laney: yours is lovely and tall and nice toys - just enough. | 12:57 |
xnox | Laney: mine is much shorter than that and more fluffy | 12:57 |
Laney | it expanded after we got it home | 12:57 |
Laney | wanted something skinnier | 12:57 |
Laney | i like the fluffiness ^_^ | 12:57 |
didrocks | pitti: I guess on the bug report, you wanted to tell "graphical.target.d" or really "display-manager.service.d" (and so, using RequiredBy=) | 13:02 |
ochosi | larsu: ah right. the version of file-roller i have is 3.12.2-0ubuntu1 (so standard vivid) | 13:21 |
ricotz | Laney, was it intended to drop "--include-image-data" from update-icon-cache? | 13:32 |
didrocks | pitti: done, tested and uploaded in https://launchpad.net/ubuntu/+source/xdiagnose/3.7.1 | 13:32 |
* didrocks goes for some cycling | 13:33 | |
=== pstolowski|lunch is now known as pstolowski | ||
pitti | didrocks: I actually meant display-manager.service.d/, so that we can keep to Wants= | 13:37 |
pitti | didrocks: oh wow, thanks! | 13:37 |
didrocks | pitti: the issue is that we would have graphical.target still active | 13:38 |
didrocks | I'm unsure we would want that | 13:38 |
pitti | didrocks: you didn't ref/close the bug? | 13:38 |
didrocks | right now, it's overriding well as a Requires | 13:38 |
pitti | oh, invalid | 13:38 |
didrocks | pitti: yeah, it's not exactly the same approach/component, so didn't retarget it | 13:38 |
pitti | didrocks: anyway, great to see that this .d/ thing works! | 13:39 |
didrocks | yeah ;) | 13:39 |
pitti | didrocks: now, 218 isn't in vivid yet, but I'll work on that as fast as possible :) | 13:39 |
didrocks | pitti: we can just override display-manager.service.d if you prefer in the future, I'm just not feeling confortable letting graphical.target as active when the graphical server didn't really start | 13:39 |
didrocks | pitti: actually, it's working on 217 as well | 13:39 |
didrocks | 218 just introduce the edit command | 13:40 |
pitti | didrocks: ok, that seems fine | 13:40 |
pitti | didrocks: oh right | 13:40 |
didrocks | this override mechanism are how generators are working in /run/system/ actually :) | 13:40 |
didrocks | I just didn't map into my brain, but yeah, way nicer | 13:40 |
didrocks | thanks for the suggestion :) | 13:40 |
* pitti hgus didrocks | 13:42 | |
pitti | ok, I'm off for some hours to do some christmas shopping | 13:42 |
* didrocks hugs pitti back | 13:43 | |
didrocks | enjoy! (and good luck) | 13:43 |
* didrocks really off to some exercise now | 13:44 | |
ochosi | larsu: another question, if i may, is patching CSD in gnome-calculator also on your list? | 14:11 |
dgadomski | hello everyone | 14:13 |
dgadomski | Trevinho: hey, do you remember the focus storing change (fix for bug #1125442)? | 14:14 |
ubot5 | bug 1125442 in Compiz 0.9.11 "Always Visible and On Top Windows Steal Focus on Workspace Switch" [Low,In progress] https://launchpad.net/bugs/1125442 | 14:14 |
dgadomski | Trevinho: there was a regression caused by the change (bug #1393020), do you have any idea how to differentiate viewport moving caused "manually" from those caused by clicking on the launcher (like in the bug description) on the compiz level? | 14:16 |
ubot5 | bug 1393020 in compiz (Ubuntu) "[regression] "Remember Focus" does not play well with unity launcher so shouldn't be default enabled or better yet the rev should be reverted" [High,In progress] https://launchpad.net/bugs/1393020 | 14:16 |
Trevinho | dgadomski: hi | 14:16 |
Trevinho | dgadomski: mh, no I don't remember a way now... but I could investigate | 14:17 |
Trevinho | dgadomski: however, Brandon Schaefer (now offline, it will be back in about 2hrs) said me he prepared a fix for that changwe | 14:18 |
Trevinho | dgadomski: so, if you wait him to join yo could have a chat with him about that | 14:18 |
larsu | ochosi: again, works on unity :) | 14:19 |
larsu | ochosi: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/gnome-calculator/vivid/view/head:/debian/patches/git_no_headerbars_in_unity.patch | 14:19 |
larsu | ochosi: probably we should update this to !GNOME as well | 14:19 |
ochosi | yeah: if (!is_desktop ("Unity")) | 14:19 |
dgadomski | Trevinho: great, I will talk to him, thanks! | 14:19 |
ochosi | larsu: yeah, would be great in order for things to be consistent, at least across desktops | 14:20 |
Laney | ricotz: that's only in the gtk2 one, right? and I think it's going to be dropped from there too | 14:20 |
Laney | so... yes | 14:20 |
draven33 | need some help with compiz in 14.04 i activatewd dual loghin compoz and metacity an can't align the 3d windows on the cube they form outside the cube | 14:54 |
ricotz | Laney, i meant the gtk3 3.14.5 packaging | 15:14 |
Laney | where do you think it got dropped? | 15:14 |
ricotz | Laney, iirc I added it in the early gtk3 packaging for gnome3 ppa, and I thought it is needed still | 15:16 |
ricotz | ignore me if it isn't needed anymore and i will drop it as well | 15:18 |
Mirv | I wonder if any core-dev would be happy to ack compiz https://ci-train.ubuntu.com/job/ubuntu-landing-012-2-publish/46/artifact/packaging_changes_compiz_1%3A0.9.12.0+15.04.20141210.2-0ubuntu1.diff ? ...or if I'll ask for better changelog. all of that is related to the line "added support for multi-arch installations" | 15:50 |
Mirv | mostly the cleanup to remove all debian/tmp/ from *.install could have been mentioned as well | 15:51 |
bschaefer | dgadomski, hello! | 16:54 |
dgadomski | bschaefer: hey | 16:55 |
bschaefer | did you have some questions about the workspace needing reverting? | 16:55 |
dgadomski | I wanted to fix bug #1125442, but Trevinho told me that you were working on some fixes already | 16:56 |
ubot5 | bug 1125442 in Compiz 0.9.11 "Always Visible and On Top Windows Steal Focus on Workspace Switch" [Low,In progress] https://launchpad.net/bugs/1125442 | 16:56 |
bschaefer | dgadomski, well i wasn't actually work on a fix atm | 16:56 |
bschaefer | i fixed a bug your prev merge caused but that was a simple fix. The main other issue would be to not re-focus a cached window when clicking on a launcher icon | 16:57 |
bschaefer | Trevinho, do you remember how the launcher icon actually focus an icon on a differnet workspace? | 16:58 |
bschaefer | we call a workspace switch in unity or compiz? | 16:58 |
dgadomski | bschaefer, oh, sorry, I pasted wrong link, this was the bug I was thinking about: bug #1393020 | 16:58 |
ubot5 | bug 1393020 in compiz (Ubuntu) "[regression] "Remember Focus" does not play well with unity launcher so shouldn't be default enabled or better yet the rev should be reverted" [High,In progress] https://launchpad.net/bugs/1393020 | 16:58 |
bschaefer | dgadomski, right, i assigned my self that bug just because i reverted your merge | 16:59 |
bschaefer | (not really a fix though) | 16:59 |
Trevinho | bschaefer: all unity side should bein PluginAdapter::FocusWindowGroup | 16:59 |
* bschaefer looks | 16:59 | |
Trevinho | bschaefer: so, there's no direct request to switch to another vp... it just happens when you ask compiz to focus a window | 17:00 |
bschaefer | Trevinho, i see, so we can patch that bit of code in compiz i suppose? | 17:00 |
bschaefer | that should be in ... screen/window.cpp? | 17:00 |
Trevinho | bschaefer: I suppose | 17:00 |
Trevinho | bschaefer: probably on activate window | 17:01 |
bschaefer | Trevinho, well ideally we need to think of a better way to cache said window | 17:01 |
bschaefer | dgadomski, but ideally, we need to come up with a way to restore focus on that cached window | 17:06 |
bschaefer | if and only if its on that workspace and we are not focusing a new window on that workspace | 17:06 |
dgadomski | bschaefer: I was debugging this and I saw that moveViewport (where the window is cached) is called in both cases and I was not able to determine how to differentiate between "manual" viewport changing versus the launcher-initiated | 17:07 |
bschaefer | right, that was a worry of mine as well... though we might have to figure out a better way to cache the window | 17:08 |
bschaefer | dgadomski, theres a place somewhere in compiz | 17:08 |
bschaefer | that focuses a window | 17:08 |
bschaefer | thats what the launcher uses, and in there it switches the viewport | 17:08 |
bschaefer | so we can detect when the launcher is being used, or rather we can detect when a window is being focus (hopefully?) | 17:09 |
* bschaefer looks at compiz | 17:09 | |
bschaefer | Trevinho, dgadomski looks like the wall plugin causes the window to be moved | 17:21 |
bschaefer | you can test that out by disabling the plugin then attempting to click on a launcher icon... strange | 17:21 |
dgadomski | bschaefer: what do you mean by "causes the window to be moved"? | 17:23 |
bschaefer | dgadomski, err not window to be moved but workspace/viewport to be moved | 17:25 |
dgadomski | bschaefer: yeah, and if you change the wall to e.g. cube you are still affected by this issue, so this needs to be fixed below the plugin layer | 17:26 |
bschaefer | dgadomski, right, but we need to ensure the plugins work as well | 17:26 |
bschaefer | dgadomski, the issue being that the unity launcher is using the wall plugin (or rather the wall plugin is auto moving workspaces to the "wanted" focused window) | 17:27 |
* bschaefer might not be making much sense | 17:27 | |
dgadomski | bschaefer: I see your point, the unity launcher also works in the same way with cube, so probably cube also moves viewports while focusing a window outside current viewport | 17:28 |
bschaefer | dgadomski, right, so if those plugins are disabled | 17:29 |
bschaefer | all is well | 17:29 |
bschaefer | but we have the wall enabled, meaning it'll attempt to focus a window | 17:29 |
bschaefer | on its viewport | 17:29 |
bschaefer | so we need to add a case in that plugin to be handled | 17:29 |
bschaefer | dgadomski, then i think we can get the fix re-merged :) | 17:29 |
bschaefer | compiz is ... like a N headed monster where N is the number of plugins enabled... | 17:29 |
dgadomski | bschaefer: Trevinho mentioned earlier that there may be a way to recognize the reason behind the moveViewport inside compiz | 17:31 |
larsu | bschaefer: lol, more like N = 2^n_plugins | 17:31 |
bschaefer | larsu, haha... true | 17:31 |
willcooke | +RAND(10) | 17:31 |
bschaefer | Trevinho, ^ | 17:31 |
bschaefer | n^n | 17:32 |
Trevinho | dgadomski: ehm? :P | 17:32 |
bschaefer | i like to imagine the best way this could turn out... We need to have ideally on each workspace a last focused window | 17:33 |
dgadomski | "<Trevinho> dgadomski: mh, no I don't remember a way now... but I could investigate" | 17:33 |
bschaefer | when we have landed on said workspace we need a way to check "Did we just change focus to get here?" vs "We just got here, change focus to last focused window" | 17:33 |
bschaefer | those two cases are the tricky part, but what "should" happen | 17:34 |
dgadomski | Trevinho: so if you don't remember a way not it means that there probably was a *WAY* :) | 17:34 |
bschaefer | is as soon as the launcher focuses the window, the last focused window for that workspace should be updated | 17:34 |
bschaefer | which would resolve our issues | 17:34 |
Trevinho | dgadomski: oh, yeah... there is... :) | 17:34 |
bschaefer | Trevinho, haha | 17:34 |
bschaefer | dgadomski, so we need to look at making sure we update the last focused cached window per workspace | 17:35 |
bschaefer | if we can do that, then when the launcher focuses a window on any workspace, that workspace updates its last focused window and we are set :) | 17:35 |
Trevinho | bschaefer: it's very likely that each workspace plugin (like wall or cube) implements that switch | 17:35 |
bschaefer | Trevinho, i greped | 17:35 |
bschaefer | Trevinho, but it doesn't matter, we can get around this under the hood | 17:35 |
bschaefer | Trevinho, all we have to check for is FOCUS OF NEW WINDOW | 17:36 |
bschaefer | when a new window is focused update the last focused window on the workspace its on | 17:36 |
Trevinho | bschaefer: I see.. | 17:36 |
bschaefer | so when we move to that workspace it will always get focus | 17:36 |
bschaefer | that sound reasonable? | 17:36 |
bschaefer | as...other wise im just imagining a mess of code to handle all strange edge cases | 17:36 |
Trevinho | bschaefer: yes, that seems correct | 17:39 |
bschaefer | so instead of updating the last focused window each time we move workspace | 17:39 |
bschaefer | the details could be hammered out | 17:40 |
dgadomski | bschaefer: are you thinking about caching focused window e.g. in CompWindow::activate? | 17:42 |
bschaefer | dgadomski, not 100% sure where but that might be a good place | 17:42 |
bschaefer | dgadomski, we just need a way to cache a window per workspace | 17:43 |
bschaefer | and when a new window is focused to update that cache array | 17:43 |
=== rickspencer3_ is now known as rickspencer3 | ||
* didrocks waves good evening and good night | 18:05 | |
=== alan_g is now known as alan_g|EOD | ||
dgadomski | bschaefer: caching in CompWindow::activate does not work as expected, I will continue debugging tomorrow morning | 18:40 |
bschaefer | dgadomski, sounds good, i think the screen will have to know about it | 18:41 |
bschaefer | vs the window | 18:41 |
bschaefer | good luck! | 18:41 |
dgadomski | bschaefer: thanks! take care | 18:42 |
bschaefer | you to! | 18:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!