slickymaster | cy tomorrow guys | 00:08 |
---|---|---|
Unit193 | Bye. | 00:08 |
slickymaster | Unit193, ;) | 00:08 |
ochosi | brainwash: no worries, just let me know when you do get to it | 00:22 |
bluesabre | ochosi, ali1234: thoughts? https://code.launchpad.net/~smd-seandavis/lightdm-gtk-greeter/gtk_image_background | 02:41 |
bluesabre | it seems to work for me, nouveau and nvidia-proprietary | 02:42 |
bluesabre | oh crap, missed a line that I needed to fix | 02:42 |
bluesabre | but yes, should still work ;) | 02:43 |
bluesabre | fixed | 02:44 |
bluesabre | make | 03:05 |
bluesabre | woops | 03:05 |
pleia2 | make: *** No targets specified and no makefile found. Stop. | 03:08 |
bluesabre | :D | 03:08 |
pleia2 | :D | 03:08 |
bluesabre | that command took forever to execture | 03:09 |
bluesabre | or execute | 03:09 |
bluesabre | not even sure what execture is | 03:09 |
pleia2 | slow channel | 03:09 |
micahg | so, that's it, no XMir for 14.04 | 03:16 |
Unit193 | Yes indeedio, I saw this. Means I'm useless again. | 03:18 |
Unit193 | (And I quite agree with that, no unstable and very new *display server* for a LTS.) | 03:20 |
holstein | i thought it was kind of cavalier | 03:21 |
micahg | Unit193: I'm sure we can find another project for you to spearhead ;) | 04:59 |
ochosi | bluesabre: hihi, you pushed your config-file :D | 09:32 |
ochosi | ali1234: i think on top of bluesabre's branch it'll be much easier to implement the fading in the greeter | 09:40 |
slickymaster | morning all | 09:56 |
ochosi | morning slickymaster | 09:56 |
slickymaster | ochosi: and a good one to you | 09:57 |
ochosi | ta | 09:57 |
slickymaster | knome: FYI https://code.launchpad.net/~slickymaster/xubuntu-docs/xubuntu-docs/+merge/196213 | 10:08 |
knome | slickymaster, ta, will get to that later | 10:43 |
slickymaster | knome: tk | 10:43 |
knome | elfy, ping | 10:44 |
ali1234 | ochosi: the RetainPermanent stuff is there for a reason | 10:52 |
ali1234 | it's supposed to ensure the background stays there between lightdm -> desktop transition | 10:52 |
ali1234 | ideally what we want to do is draw a gtk image until lightdm is going to quit, then dump the image to the root pixmap once just before exiting | 10:53 |
ochosi | yeah, that sounds like a plan | 10:53 |
ali1234 | however this doesn't necessarily fix the crashing on nouveau | 10:54 |
ali1234 | otoh it isn't necessarily the cause either | 10:54 |
ochosi | well bluesabre mentioned that his gtkimage branch worked with nouveau | 10:55 |
ochosi | so we could hope that it really helps | 10:55 |
ali1234 | yeah but he removed the RetainPermanent stuff | 10:56 |
ochosi | oh | 10:57 |
ochosi | that has to be added back then | 10:57 |
ali1234 | yes, at least for a test | 10:57 |
bluesabre | ah | 11:04 |
bluesabre | I can add that back | 11:04 |
bluesabre | but this should at least eliminate the glitches for when sitting at the login screen | 11:05 |
ochosi | sounds good | 11:05 |
bluesabre | (also fixes the crazy memory leak) | 11:06 |
bluesabre | but, I suppose I'll be adding that back with the RetainPermanent code | 11:06 |
ochosi | yeah, i think we can merge that branch pretty soon :) | 11:06 |
ali1234 | bluesabre: it has to be not just added back, but reworked so it only gets called once at the very end, instead of at the start | 11:07 |
bluesabre | ah | 11:07 |
bluesabre | I see | 11:07 |
bluesabre | that makes sense | 11:07 |
ochosi | yeah, so the leak would indeed be fixed | 11:07 |
ali1234 | not necessarily | 11:08 |
ali1234 | well, yes, the *leak* | 11:08 |
ali1234 | but not necessarily the display corruption/crashing | 11:08 |
ali1234 | because it leaks by design | 11:08 |
bluesabre | right | 11:09 |
ali1234 | the entire point of that code is to leave a pixmap on the display after it exits | 11:09 |
ochosi | i thought the display-corruption was helped by using the gtkimage? | 11:09 |
ali1234 | that may be just papering over it | 11:09 |
bluesabre | the GtkImage draws with Gtk, so it eliminates the corruption at least when viewing it at the login screen | 11:09 |
bluesabre | it will probably come back with drawing using cairo/x | 11:10 |
ali1234 | i don't think it is the way it is drawn | 11:10 |
ochosi | and the leak will only be 1 image instead of all the images that were seen in the greeter by switching users | 11:10 |
ali1234 | i think it's the RetainPermanent followed by client disconnect | 11:10 |
ali1234 | but i have no way to test | 11:10 |
ochosi | well i can switch to nouveau if it's just about needing a tester ;) | 11:11 |
ali1234 | actually... this other computer i just found has nvidia onboard graphics | 11:14 |
ali1234 | i can probably test on that if i update it | 11:14 |
ali1234 | it's also on 12.04 though | 11:15 |
ochosi | fine by me, if it doesn't work for some reason and you need me as a nouveau tester after all, let me know | 11:21 |
ochosi | bluesabre: i think we should probably set up a shared branch | 11:21 |
ochosi | otherwise you'll always have to merge stuff back and forth | 11:22 |
bluesabre | probably | 11:22 |
bluesabre | I'll do that in a sec | 11:22 |
ochosi | cool, ty | 11:22 |
bluesabre | and that way I can leave tasks for you | 11:23 |
bluesabre | :D | 11:23 |
ochosi | :) | 11:24 |
ochosi | i added a funny list-mode to tabwin today | 11:24 |
bluesabre | https://code.launchpad.net/~lightdm-gtk-greeter-team/lightdm-gtk-greeter/gtk_image_background | 11:25 |
ochosi | http://www.zimagez.com/zimage/screenshot-11222013-122522pm.php | 11:25 |
ochosi | bluesabre: but ali1234 isn't part of the team yet so can't push, right? or is this a public branch? | 11:26 |
bluesabre | adding him now :> | 11:26 |
ochosi | nice | 11:26 |
bluesabre | btw, do you know who Adam Jiang is? | 11:26 |
ochosi | nope | 11:27 |
bluesabre | https://launchpad.net/~lightdm-gtk-greeter-team/+members#proposed | 11:27 |
ochosi | not the guy who wrote the another-lightdm-gtk-greeter | 11:27 |
ochosi | cause that guy's name is andrew | 11:27 |
bluesabre | interesting tabwin stuff | 11:28 |
ochosi | i'd like to add that as an option | 11:28 |
ochosi | it's actually quite a small code-change in the ochosi/tabwin branch | 11:28 |
ali1234 | right, computer is set up | 11:50 |
ali1234 | how do i set up lightdm/light-locker for testing? | 11:52 |
ochosi | ali1234: i usually just build with --prefix=/usr and then install it over the existing copy | 12:14 |
ochosi | if i want to revert, i just use apt-get install --reinstall | 12:14 |
ochosi | for light-locker it should suffice to use the daily-PPA, no need to build from source | 12:14 |
brainwash | ochosi: uhm, now I'm kinda lost, all this talk about the background issue | 12:21 |
ochosi | brainwash: if you wanna test a branch, this is the one: https://code.launchpad.net/~lightdm-gtk-greeter-team/lightdm-gtk-greeter/gtk_image_background | 12:21 |
bluesabre | wait a sec on that | 12:22 |
bluesabre | :) | 12:22 |
brainwash | ochosi: just tested the new xfdesktop (4.11) and it messes with the copy process of the root pixmap | 12:22 |
bluesabre | I might push a nice fix in a sec | 12:22 |
ochosi | brainwash: do you use xfdesktop from git or from the 4.12 ppa? | 12:22 |
brainwash | ppa | 12:22 |
ochosi | bluesabre: ah, yeah, i would've mentioned the caveat ;) | 12:22 |
ochosi | brainwash: mainly asking cause the interaction of greeter and xfdesktop only works with the patch for accountsservice | 12:23 |
ochosi | otherwise the user-wallpaper doesn't get set | 12:23 |
brainwash | it does get set | 12:24 |
ochosi | yeah, with the PPA version it should | 12:24 |
ochosi | so what's the problem with the root pixmap exactly? | 12:24 |
brainwash | xfdesktop crashes | 12:25 |
brainwash | because of the compositor coyping the root pixmap | 12:25 |
brainwash | tried to delay the launch of xfdesktop | 12:26 |
brainwash | but it did not work, let me try again | 12:26 |
brainwash | https://code.launchpad.net/~thad-fisch/lightdm-gtk-greeter/root-window | 12:27 |
ali1234 | nope. | 12:28 |
ali1234 | you can't delay xfdesktop | 12:28 |
ali1234 | the greeter intentionally sets resources that stay around after it exits | 12:28 |
ali1234 | the support for doing this seems very poor on nouveau | 12:29 |
brainwash | you cannot? | 12:30 |
ali1234 | you can delay it | 12:30 |
ali1234 | but you can't delay it until those pixmaps go away, because they never go away | 12:30 |
ochosi | i think the problem is xfwm4's compositor | 12:31 |
ochosi | cause it draws over the root pixmap in grey | 12:31 |
ali1234 | i have noticed that | 12:31 |
ochosi | while it'd better do nothing i guess (at least in our case, where xfdesktop is in use) | 12:31 |
brainwash | yes, if xfdesktop launches while xfwm4 copies the root background, xfdesktop will crash | 12:31 |
bluesabre | seems a bit buggier, try this ochosi, ali1234: https://code.launchpad.net/~lightdm-gtk-greeter-team/lightdm-gtk-greeter/gtk_image_background | 12:32 |
ochosi | ali1234: so basically that grey flicker is what brainwash wants to get rid of :) | 12:32 |
ochosi | bluesabre: wait, buggier in what respect? :> | 12:32 |
ali1234 | i've installed light-locker. how do i lock the screen? | 12:32 |
bluesabre | I think the grey flicker is the GtkWindow background color | 12:33 |
bluesabre | probably xfdesktop does show() before the background image is loaded I would guess | 12:33 |
ochosi | ali1234: after the light-locker process has been started, run "light-locker-command -l" | 12:33 |
ochosi | bluesabre: no, there really is a call in the compositor to draw the bg in that grey color | 12:33 |
bluesabre | oh | 12:33 |
bluesabre | thats terrible | 12:33 |
ochosi | :) | 12:34 |
bluesabre | btw, not buggier to where it doesn't work, but sometimes at first login the bg is just drawn as the bg-color | 12:34 |
ali1234 | ok that works for me, no graphics corruption | 12:34 |
ochosi | well that's what you see when logging in normally | 12:34 |
bluesabre | still an improvement though | 12:34 |
ali1234 | with the standard greeter that is | 12:34 |
ochosi | standard greeter? | 12:34 |
ali1234 | the one from the install cd | 12:34 |
bluesabre | anyway, gotta run | 12:35 |
bluesabre | feel free to add any personal fixes or garnishes to that branch :) | 12:35 |
ochosi | okey :) | 12:35 |
ochosi | seeya later bluesabre | 12:35 |
ochosi | and hf@work | 12:35 |
brainwash | bug 1232804 | 12:36 |
ubottu | bug 1232804 in xfwm4 (Ubuntu) "Improve "login greeter -> desktop" transition in Xubuntu" [Undecided,Confirmed] https://launchpad.net/bugs/1232804 | 12:36 |
brainwash | explains the grey background | 12:36 |
ali1234 | if you kill xfdesktop and run some xsetroot commands, it doesn't change the root window colour, presumably because of this | 12:37 |
ali1234 | you just see the grey | 12:38 |
ochosi | so that should probably really be fixed in xfwm4 then... | 12:39 |
ali1234 | the grey is #7f7f7f | 12:40 |
ali1234 | http://git.xfce.org/xfce/xfwm4/tree/src/compositor.c#n872 | 12:40 |
ochosi | yup | 12:41 |
brainwash | how should it be fixed in xfwm4? | 12:46 |
ochosi | not sure | 12:46 |
brainwash | the code for copying the root background is already there | 12:47 |
brainwash | it works | 12:47 |
brainwash | but xfdesktop messes up | 12:47 |
brainwash | and in 4.11 xfdesktop messes up big time | 12:47 |
brainwash | why does it draw a white background? | 12:47 |
ali1234 | so does it make any difference with any of the branches of the greeter? | 12:47 |
brainwash | now we have a grey and a white background... | 12:48 |
ochosi | ali1234: does what make any diff? | 12:48 |
brainwash | double flicker | 12:48 |
ali1234 | does anything make any difference? | 12:48 |
ochosi | :D | 12:48 |
ali1234 | because there's too many variables here | 12:48 |
ali1234 | what exactly are the requirements? | 12:48 |
ochosi | well, i think with retaining the background in the greeter, letting xfwm4 copy the root pixmap and fixing xfdesktop to cooperate could do the trick (as far as i understood) | 12:49 |
ochosi | the requirements are a flicker-free login | 12:49 |
ali1234 | and unlocking | 12:49 |
ochosi | yeah | 12:49 |
ochosi | although | 12:49 |
ochosi | unlocking is somewhat impossible | 12:49 |
ali1234 | and also it shouldn't crash or display corrupted graphics | 12:49 |
ochosi | cause unlocking includes a VT-switch | 12:49 |
ochosi | so there will be a blank screen | 12:50 |
ochosi | yeah, not crashing and not corrupting the screen would be nice :) | 12:50 |
ali1234 | nouveau just totally corrupted the display and froze the system | 12:52 |
ali1234 | can we just say "nouveau isn't supported because it's crap" | 12:52 |
brainwash | on normal screen locking? | 12:53 |
ali1234 | no just while i was trying to open firefox | 12:53 |
brainwash | it happened once for me after hibernate | 12:53 |
brainwash | scrambled screen | 12:53 |
ochosi | yeah | 12:53 |
ochosi | well getting it to work on everything else first and then trying to add some workarounds for nouveau would be nice | 12:53 |
ali1234 | yeah | 12:53 |
ochosi | but i know it's not very stable... | 12:54 |
brainwash | this whole thing is somewhat messed up, we should not have started messing around with it :D | 12:54 |
ali1234 | if we're getting corruption with nouveau, 99% it's a nouveau bug | 12:55 |
ochosi | brainwash: yeah, it's all *your* fault ;) | 12:55 |
brainwash | ochosi: do you know why xfdesktop behaves differently since 4.11? | 12:56 |
ochosi | no, i haven't looked at the changes | 12:56 |
ochosi | but i'm in touch with eric | 12:57 |
brainwash | what is the deal with the white background (overlay)? | 12:57 |
ochosi | so i can discuss this issue with him when he's back from his holiday | 12:57 |
ochosi | (~1-2weeks) | 12:57 |
ali1234 | well, tell you what... i'll concentrate on this compositor thing | 12:57 |
ali1234 | it's fine on nvidia... | 12:57 |
ochosi | sweet | 12:57 |
brainwash | so can we merge https://code.launchpad.net/~thad-fisch/lightdm-gtk-greeter/root-window ? this change alone does no harm | 12:58 |
brainwash | well, but on the other side, I'm not an expert, so it could be bad | 12:59 |
brainwash | :) | 12:59 |
ochosi | it doesnt look very harmful, but i'd prefer if ali1234 and bluesabre also took a look at it | 13:00 |
ochosi | the code is from xfwm4, right? | 13:00 |
brainwash | not directly, xfwm4 queries this X window property, so I simply set it for the lightdm background | 13:01 |
ochosi | yeah, that's what i meant | 13:02 |
ali1234 | brainwash: what's the reasoning for it? | 13:03 |
brainwash | compositor.c | 13:03 |
brainwash | http://git.xfce.org/xfce/xfwm4/tree/src/compositor.c#n806 | 13:03 |
brainwash | line 840 | 13:04 |
brainwash | it worked nicely with the small startup delay added to xfdesktop 4.10 | 13:06 |
ali1234 | and this causes the crash when xfdesktop is starting? | 13:06 |
ochosi | so the fact that both xfwm4 and xfdesktop want to paint the root window causes the crash? | 13:06 |
ali1234 | ok i see exactly what is happening | 13:06 |
ali1234 | nah | 13:06 |
brainwash | yes, https://bugs.launchpad.net/ubuntu/+source/xfwm4/+bug/1232804/comments/3 | 13:07 |
ubottu | Ubuntu bug 1232804 in xfwm4 (Ubuntu) "Improve "login greeter -> desktop" transition in Xubuntu" [Undecided,Confirmed] | 13:07 |
brainwash | comment 3 | 13:07 |
ali1234 | this might even explain the graphic corruption | 13:07 |
ali1234 | i need to look at xfdesktop source... | 13:07 |
ochosi | oh the suspense... | 13:09 |
brainwash | but these changes are only tested by me so far, or do you mean that there is a general xfdesktop issue? | 13:09 |
ali1234 | you have to think about how everything interacts | 13:09 |
ali1234 | see | 13:09 |
ali1234 | the greeter makes those RetainPermanent pixmaps | 13:09 |
ali1234 | then xfwm4 starts up and reads the pixmap address for the root image | 13:09 |
ali1234 | that's the RetainPermanent pixma the greeter stuffed in there | 13:09 |
ali1234 | then it memcpys it | 13:09 |
ali1234 | meanwhile xfdesktop starts up and also tries to set the root pixmap (presumably) | 13:10 |
ali1234 | thus, we have a race condition, an un-owned pixmap that is likely to get freed at any time, and therefore some nasty invalid pointer operations | 13:10 |
brainwash | doesn't xfdesktop draw a new window? | 13:10 |
ali1234 | the root window is the root window | 13:10 |
ali1234 | who knows what it really does to it? | 13:10 |
ali1234 | it does draw a new window but that isn't necessarily the only thing it does | 13:11 |
ali1234 | so... time to look at the code | 13:11 |
brainwash | yes, so an extra delay fixed this | 13:11 |
brainwash | but xfdesktop 4.11 is different again | 13:11 |
brainwash | and we should focus on the 4.11 stuff, right? | 13:12 |
ali1234 | yeah | 13:12 |
ali1234 | actually i would be interested to see what happens if you disable xfdesktop completely and then retry all the branches on nouveau | 13:13 |
ali1234 | i suspect xfdesktop is "cleaning" all the RetainPermanent pixmaps out | 13:14 |
ochosi | well that in itself is not a bad idea, eh? | 13:14 |
ali1234 | it is | 13:14 |
elfy | knome: a 30 minute pong or you'll have to ping me again | 13:14 |
ali1234 | it is if xfwm4 is going to take a copy of the pointer | 13:14 |
ochosi | hm, all in all it's strange no-one noticed this before | 13:15 |
ali1234 | yeah, i always think that | 13:15 |
ochosi | guess in most cases it doesn't cause huge problems | 13:15 |
ali1234 | "how did this code *ever* work?" | 13:15 |
ochosi | :) | 13:15 |
ali1234 | brainwash: so with that root properties patch, if you kill xfdesktop, do you still see your background? | 13:20 |
brainwash | ali1234: all I see is a white background | 13:21 |
ali1234 | even with the older version? | 13:21 |
ochosi | or: you deactivate/uninstall xfdesktop and try again | 13:22 |
ochosi | then xfwm4 should show the wallpaper set by lightdm | 13:22 |
ochosi | (or did i misunderstand something?) | 13:22 |
ali1234 | yes, it actually memcpy's the pixmap | 13:22 |
ali1234 | or at least it tries to | 13:23 |
brainwash | why that? | 13:23 |
ali1234 | well, i guess with a compositor you never see the true root window | 13:23 |
brainwash | right, thet's compositing | 13:23 |
ali1234 | on intel i get a black screen if light-locker is installed | 13:32 |
ochosi | ali1234: you mean in the greeter? (as in: happening as it should) | 13:36 |
ochosi | i mean: does it blank the screen on locking | 13:37 |
ali1234 | no i mean as in it boots up to a blank screen | 13:37 |
ali1234 | and no matter what i do the screen stays blank | 13:37 |
ochosi | that's more than odd | 13:37 |
ochosi | can't say i get why that would happen | 13:37 |
ochosi | unless lightdm_greeter_get_lock_hint would return TRUE in any case | 13:39 |
ochosi | ali1234: did you check the /var/log/lightdm/* logs? | 13:45 |
ali1234 | fixed it, i had to delete displays.xml | 13:51 |
brainwash | odd fix | 14:09 |
ali1234 | hmm | 14:27 |
ElderDryas | Seeing as how pleia2 likes notes here rather than -ot (iirc), http://www.dedoimedo.com/computers/xubuntu-salamander.html | 15:38 |
ElderDryas | What you guys do, slip some I-Love-Xubuntu juice into his coffee? He used to despise XFCE :) | 15:39 |
ochosi | ElderDryas: hehe, jolly-good read that | 15:44 |
slickymaster | ElderDryas, ochosi, +1 on that | 15:52 |
ali1234 | i never really thought xfce was any good until recently | 16:05 |
ali1234 | 4.10 seems to have fixed many many bugs and it's actually viable now | 16:05 |
ElderDryas | 4.8 started it, 4.10 took it most of the way...now I look forward to 4.12. But that seems a ways off now :( | 16:08 |
slickymaster | ochosi: ping | 16:36 |
ochosi | slickymaster: pong | 16:38 |
slickymaster | ochosi: looking at the code in http://git.xfce.org/apps/parole/tree/src/parole-player.c#n2441 I see that the Page Down key is for skip back 600 secs and the Page Up key is for the skip ahead 600 secs | 16:38 |
ochosi | exactly | 16:38 |
ochosi | 600secs = 10mins | 16:38 |
ochosi | (i guess that's easier to grasp for most ppl) | 16:38 |
slickymaster | ochosi: yeah, I know, but the thing is, I'm not discovering the keys for the seek medium options | 16:39 |
ochosi | it's ctrl+left/right | 16:39 |
ochosi | http://git.xfce.org/apps/parole/tree/src/parole-player.c#n2464 | 16:40 |
ochosi | GDK_CONTROL_MASK | 16:40 |
slickymaster | ochosi: never mind, it was right in front of me | 16:41 |
ochosi | no problem | 16:41 |
* slickymaster thinks that he really needs a pair of glasses | 16:41 | |
ochosi | actually it's good for you to know, because the show/hide menu shortcut is also CTRL+m | 16:41 |
slickymaster | yeah, that one I saw | 16:42 |
ochosi | cool | 16:42 |
slickymaster | i spent almost ten minutes looking at the switch block and was unable to see them, I really need glasses | 16:43 |
ochosi | hehe, that happens from time to time | 16:43 |
slickymaster | thanks, ochosi | 16:43 |
ochosi | main question being: didn't you *see* it or couldn't you *read* it ;) | 16:43 |
ochosi | (only in the second case you *really* need glasses) | 16:43 |
slickymaster | ochosi: that's the million dollar question, isn't it? | 16:44 |
ochosi | hehe, i guess | 16:44 |
slickymaster | ochosi: just one other thing, is there any particular reason for the fact that the arrow keys (up, down, left and right) are italic formatted? | 16:53 |
slickymaster | I mean in http://smdavis.us/doku/doku.php?id=usage&#keyboard_shortcuts | 16:54 |
ochosi | let me check how other xfce projects did that | 16:54 |
ochosi | hm, i guess that was jjfrv8's decision | 16:55 |
ochosi | tbh i didn't really notice that before | 16:55 |
ochosi | i don't mind either formatting, but yeah, not sure the italics are needed | 16:55 |
slickymaster | ochosi: I don't think that it's visually coherent | 16:56 |
ochosi | then remove the italic formatting, i'm totally fine with that | 16:56 |
slickymaster | ok | 16:56 |
ochosi | jjfrv8: fyi ^ | 16:56 |
slickymaster | ochosi: I think I've covered them all. Anyway it's at your disposal to check and review http://smdavis.us/doku/doku.php?id=usage&#keyboard_shortcuts | 17:10 |
slickymaster | please ping me in case I've skip anything or if you think that something needs to corrected | 17:10 |
slickymaster | I'll bbl, after dinner | 17:10 |
ochosi | slickymaster: thanks! and bon appetit | 17:11 |
ochosi | i'll be away for most of the rest of the weekend | 17:11 |
ochosi | returning on sunday evening | 17:11 |
slickymaster | ochosi: enjoy your weekend. We'll talk then | 17:12 |
ochosi | slickymaster: thanks you too! seeyq | 17:12 |
ochosi | seeya | 17:12 |
slickymaster | cy | 17:12 |
ochosi | bluesabre: i just reviewed slickymaster's addition of the playback kb-shortcuts and changed a few small details, care to review that? http://smdavis.us/doku/doku.php?id=usage&#keyboard_shortcuts | 17:15 |
ochosi | bluesabre: i think it's mostly fine and everything should be covered now, not sure about the "skip" wording (should it be "seek" instead?), i also noticed we don't have a shortcut for next/previous track | 17:15 |
ali1234 | ok, getting good progress on this greeter stuff | 17:25 |
ochosi | ali1234: really? do tell! | 17:25 |
ali1234 | i've got everything set up for testing, results are as expected | 17:26 |
ali1234 | just testing the compositor pixmap memcpy stuff | 17:26 |
ali1234 | brainwash' patch doesn't actually work, but the idea is right | 17:26 |
ali1234 | we need to make the compositor copy the root window instead of just filling it with grey | 17:27 |
ochosi | indeed | 17:27 |
brainwash | why does it work for me? | 17:27 |
ali1234 | brainwash: i dunno. if it's because "different drivers" then i will rage | 17:27 |
ali1234 | i'll get to the bottom of it anyway | 17:28 |
brainwash | and why shouldn't it work? it's just setting a property.. | 17:28 |
brainwash | =S | 17:28 |
ali1234 | i dunno. all i know is that even with the patch applied i still get a grey root tile | 17:28 |
ali1234 | i have not yet traced it to see exactly what is happening | 17:29 |
brainwash | you also recompiled xfwm4, right? | 17:29 |
ali1234 | no, did you patch that too? | 17:29 |
brainwash | amg | 17:29 |
ali1234 | or it needs none-default configuration options? | 17:29 |
brainwash | right, it's not a default compile option | 17:29 |
ali1234 | well, that seems silly | 17:29 |
brainwash | read my bug report =S | 17:30 |
ali1234 | so how do i do that? ./autogen.sh MONITOR_ROOT_PIXMAP=yes? | 17:31 |
brainwash | you need to define it | 17:33 |
ali1234 | like, manually in the source code?? | 17:34 |
ali1234 | ah i see | 17:36 |
ali1234 | this code doesn't just copy the background once | 17:36 |
ali1234 | it updates it continuously, so that you really can see the root window at all times | 17:36 |
ali1234 | we don't need that | 17:36 |
ali1234 | we just want to copy it once | 17:36 |
ali1234 | right, rebuilt it | 17:42 |
ali1234 | it works... but | 17:42 |
ali1234 | not very well | 17:42 |
ali1234 | i also got xfdesktop crashes... but without updating to the newest version | 17:44 |
brainwash | comment 3? | 17:47 |
brainwash | it worked for me if I slightly delayed the launch of xfdesktop | 17:47 |
brainwash | I cannot get it to work properly, so it's not very likely that this "improvement" will make it into the final LTS release | 17:48 |
ali1234 | that's just a workaround | 17:55 |
brainwash | yes | 18:07 |
brainwash | needs to be implemented properly or not at all | 18:07 |
ali1234 | brainwash: xfdesktop already has a patch that makes it wait for 5 seconds after xfwm4 starts | 19:03 |
brainwash | ali1234: right, does it work? | 19:04 |
ali1234 | well, obviously not | 19:05 |
ali1234 | what i do see is that xfdesktop also uses those desktop pixmap atoms atoms | 19:05 |
ali1234 | aaaaaaaaaaaaah | 19:16 |
ali1234 | https://mail.gnome.org/archives/wm-spec-list/2002-January/msg00001.html <- this thread explains a lot of things | 19:18 |
ali1234 | one of which is that if xfdesktop wants to set the root pixmap that way it is doing it, it is liable to get killed later on | 19:19 |
ali1234 | right... i think i understand the problem | 19:38 |
ali1234 | yeah... fixed | 19:47 |
ali1234 | so the thing is, if you enable MONITOR_ROOT_PIXMAP it does just that. it doesn't just copy the pixmap one time at startup | 19:47 |
ali1234 | it also enables a bunch of other code, and that other code is what causes the crash | 19:47 |
ali1234 | and it isn't really a crash, it's actually defined behaviour | 19:48 |
ali1234 | because now xfwm4 and xfdesktop both try to control the root pixmap, they fight eachother | 19:48 |
ali1234 | so the fix is to enable that first copy at startup only | 19:48 |
ali1234 | and not the other stuff | 19:48 |
ali1234 | basically the problem is they all try to manipulate the root bg pixmap, and they all use a different standard to do it | 19:52 |
ali1234 | they should all use the same standard, and then everything will be fine | 19:52 |
ali1234 | also none of them follows the standard they picked quite correctly | 19:52 |
brainwash | ok, finally we (actually you) understand what is going on :) | 19:54 |
ali1234 | the good news is this is totally fixable | 19:54 |
ali1234 | it probably also explains why the corrupted graphics on nouveau | 19:55 |
ali1234 | it all comes down to those RetainPermanent pixmaps. that's a standard, but there's no standard way to manage them once you've created one on the root | 19:55 |
ali1234 | oh the other good thing about this is that fixing it will also fix the greeter memory leak | 19:56 |
ali1234 | without having to implement gtk images etc | 19:57 |
brainwash | I will test your magical fix then :) | 19:59 |
brainwash | if it's ready | 19:59 |
ali1234 | brainwash: basically, instead of #define MONITOR_ROOT_PIXMAP, remove the #ifdefs from root_tile() only | 20:12 |
ali1234 | and that's it | 20:12 |
pleia2 | drc: thanks :) | 20:17 |
drc | pleia2: you're welcome...for what? | 20:17 |
pleia2 | drc: dedoimedo link, added to press page | 20:19 |
drc | oh...yeah...You'll have to forgive me, that was this morning and my wife says goldfish have a longer memory than I do :) | 20:20 |
pleia2 | hehe | 20:20 |
brainwash | bug 1254087 | 20:26 |
ubottu | bug 1254087 in xfce4-whiskermenu-plugin (Ubuntu) "[wishlist] Option to Use Whiskermenu as Desktop Menu" [Undecided,New] https://launchpad.net/bugs/1254087 | 20:26 |
Noskcaj | brainwash, I saw that, how hard would it be to fix? | 20:27 |
ali1234 | i pushed the root pixmap fix to my xfwm4 zoom2 branch | 20:28 |
brainwash | don't know, but why replace the xfdesktop right-click menu?! | 20:28 |
brainwash | ali1234: got any prebuild packages too? | 20:29 |
ali1234 | no | 20:29 |
ali1234 | you still need to patch the greeter to set the atom too | 20:29 |
brainwash | because many people need to test this stuff :) | 20:29 |
ali1234 | but that needs "improving" | 20:29 |
brainwash | currently everything is a bit messy, branches and fixes everywhere | 20:31 |
* ali1234 pokes #ubuntu-x for information | 20:36 | |
Noskcaj | brainwash, the bug is about right-click>Applications> | 20:40 |
brainwash | mmh, so xfdesktop would need support for the whisker panel plugin | 20:42 |
brainwash | reassign the report to xfdesktop? | 20:45 |
ali1234 | hmm... i'm going to look at what unity-greeter does wrt RetainPermanent | 20:48 |
ali1234 | we should support whatever they do too, if they use different atoms | 20:48 |
brainwash | different? xfwm4 only checks for 2 I think | 20:53 |
brainwash | oh RetainPermanent | 20:54 |
ali1234 | yeah there's at least 4 different ones | 20:57 |
ali1234 | (xfdesktop checks for 2 different ones) | 20:57 |
ali1234 | basically you make a RetainPermanent pixmap and then stuff it's X resource ID into an atom so you can find it again later | 20:57 |
ali1234 | but every desktop uses a different atom for this | 20:58 |
ali1234 | unity-greeter defines RetainPermanent but does not use it | 21:01 |
ali1234 | so now i'm really puzzled | 21:01 |
ali1234 | it looks like compiz just copies whatever is on the screen at startup | 21:13 |
ali1234 | no checking of atoms | 21:13 |
ali1234 | it does monitor atoms for later background changes though | 21:14 |
brainwash | and gdm/mutter? :P | 21:16 |
ali1234 | who cares about them? | 21:18 |
knome | bluesabre, pleia2: either/both of you around? | 21:24 |
knome | jjfrv8, or you? :) | 21:24 |
pleia2 | you're supposed to say what you want first so I can figure out whether to say yes | 21:24 |
knome | nah, not saying is part of my evil plan! | 21:25 |
* skellat is lurking | 21:25 | |
knome | skellat, i don't specifically need you at the moment but if you have anything you'd like to discuss, i'm here! | 21:26 |
knome | pleia2, do you have time to go through the blueprints today? | 21:26 |
pleia2 | maybe after work | 21:26 |
knome | what UTC is that? | 21:27 |
pleia2 | 1 | 21:27 |
knome | umph | 21:27 |
knome | might not be around at that time today | 21:27 |
knome | but we can look at it later, no hurry | 21:27 |
pleia2 | sunday would be better | 21:27 |
knome | (and i could have a quick stab at it myself as well) | 21:28 |
pleia2 | knome: lmk when you publish testing post | 21:28 |
knome | ah, i could do that soonish | 21:28 |
knome | are you lurking around? | 21:28 |
pleia2 | yeah | 21:28 |
knome | i'm probably doing it in the next 30mins | 21:29 |
knome | perfect, i'll ping you then :) | 21:29 |
=== Mapley is now known as rainbowdash` | ||
knome | heh, ok, so ~ubuntu-testcase is going to be removed ;) | 21:41 |
knome | does anybody see any reason for ~xubuntu-dev to be a member of ~xubuntu-team? | 21:47 |
knome | everybody in ~xubuntu-dev is explicitly a member of -team as well (except people who are in ~dev because of other team memberships) | 21:47 |
knome | if we drop the membership of ~xubuntu-dev, then LP would show the real amount of -team members | 21:48 |
ali1234 | here's a tonne of information about this: https://github.com/mate-desktop/mate-desktop/pull/46 | 21:52 |
knome | pleia2, prodded out http://xubuntu.org/news/help-us-test-xubuntu-14-04-lts/ | 22:05 |
knome | and tweeted | 22:05 |
ali1234 | hmm you know what? i think this code from xfwm4 MONITOR_ROOT_PIXMAP was written by someone who was very confused, and they just got the atom names wrong | 22:11 |
slickymaster | night all | 22:25 |
pleia2 | knome: thanks | 22:25 |
knome | np | 22:25 |
knome | hey slickymaster | 22:25 |
slickymaster | hi, knome | 22:26 |
knome | slickymaster, has elfy been in contact with you since yesterday yet? | 22:26 |
slickymaster | knome, no | 22:26 |
knome | slickymaster, ok, let him take the time then :) | 22:27 |
slickymaster | knome, ok | 22:27 |
slickymaster | is there any problem? | 22:27 |
knome | slickymaster, not at all! | 22:27 |
slickymaster | ok | 22:27 |
slickymaster | ochosi, saw the changes you made on to the playback table on http://smdavis.us/doku/doku.php?id=usage#keyboard_shortcuts and I must say that you were right from the beginning | 22:30 |
slickymaster | ochosi, people can understand 10 mins but probably wouldn't understand 600 secs | 22:30 |
slickymaster | ochosi, good call | 22:30 |
ali1234 | ochosi: i summarized all my knowledge about the greeter pixmap stuff on the bug report, if you don't wanna read all the scrollback: https://bugs.launchpad.net/ubuntu/+source/xfwm4/+bug/1232804/+attachment/3899352/+files/xsetroot.patch | 22:34 |
ubottu | Ubuntu bug 1232804 in xfwm4 (Ubuntu) "Improve "login greeter -> desktop" transition in Xubuntu" [Undecided,Confirmed] | 22:34 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!