=== m_conley is now known as m_conley_away [07:22] morning [07:59] kvalo: morning! [08:01] smspillaz: morning. how was australia day yesterdat? [08:01] yesterday even [08:09] good morning [08:11] kvalo: I didn't really get up to much suprisingly, and I missed the fireworks -_- [08:11] didrocks: morning! [08:11] hey smspillaz! how are you? [08:11] didrocks: I have fixed the "wrong window positions bug", now working on the gconf bug [08:11] didrocks: feeling awesome as usual [08:11] didrocks: morning [08:11] hey everybody! [08:11] smspillaz: awesome! [08:12] smspillaz: oh, missing fireworks sucks [08:12] MacSlow: guten morgen [08:12] smspillaz: I don't see that one in your email, am I stilll awake? [08:12] even still* [08:12] hey kvalo, MacSlow! [08:12] kvalo: yeah, I messed up the time (thought it was 9pm instead of 8pm) [08:12] kvalo, hyvää huomenta [08:12] :) [08:12] didrocks: I think I commented on your bug report with the patch [08:13] salut didrocks [08:13] MacSlow: oh wow! [08:13] smspillaz: excellent then, will get to it as well :) [08:13] guton morgen hyvaa huomenta bonjour bonjourno etc etc [08:13] kvalo, google translate... *cough* :) [08:13] *guten [08:13] kvalo, I would have no idea how to correctly pronouce that though :) [08:13] MacSlow: next challenge, pronounce it correctly ;) [08:14] MacSlow: about funny words, does "saippukauppias" sound funny to you? [08:15] didrocks: I was thinking though for bug 688816, we might need a new plugin to handle that case (since it needs to load before the place plugin) [08:15] Launchpad bug 688816 in unity (Ubuntu) "Don't create windows over the launcher" [Medium,Confirmed] https://launchpad.net/bugs/688816 [08:15] kvalo, it certainly looks funny [08:15] smspillaz: I can handle that one later, not for alpha2 though [08:15] smspillaz: as you fix the wrong window position at start, it's fine :) [08:16] didrocks: ah ok, it would be like a ~100 line plugin [08:16] MacSlow: I think I need to list you some funny words and we can go through them over beer [08:16] smspillaz: yeah, I'll do it, I think I got what is needed [08:16] I could do that in like 10 minutes, although I'm wondering if that's the case then how much sense would it make a compiz-plugins-unity package ? [08:16] smspillaz: before the place plugin in any case? to avoid jumping again? [08:17] didrocks: yeah, so basically a plugin before place which hooks CompWindow::place on any window that has a special property set [08:17] smspillaz: right, but at least, it will be a nice exercice for me :) [08:17] didrocks: I'll forward the email I sent to Jason [08:17] smspillaz: excellent, thanks! [08:18] I'm just ensure about the "another plugin just for that", maybe, we should find another way, or having a plugin that handle all those special cases [08:18] argh, I keep on forgetting to back up evolution before reinstalling [08:19] didrocks: I would have opted to do it in the unity plugin, but that isn't possible since the unity plugin needs to load after place [08:19] didrocks: and you can make this plugin smart actually and not make it do anything unless the window has this property set [08:19] smspillaz: yeah, because you need to override the place action in that case :) [08:20] didrocks: indeed, well loading before it will make place actually place the window and then we just shift it to the right if it overlaps the launcher area [08:20] smspillaz: ok, can you just forward me the email? I'll get that done next week [08:20] didrocks: sure, although if I have time to I might do it so that we can get it in by the time we do the upload [08:21] smspillaz: no, I will do the upload today and no more new feature then (freeze) for alpha2 [08:21] fair enough [08:21] smspillaz: so, I just need the gconf crash bug fixed :) [08:21] didrocks: yeah, good point [08:21] didrocks: give me some other stuff to fix in case I have time on my hands [08:21] smspillaz: also, if you don't let the other try to do it, you won't get anyone helping you in the future :) [08:22] didrocks: this is true :p [08:22] smspillaz: sure, finishing triaging my emails first, still 200+ to go :) [08:22] (and besides, someone is doing work *for* me, what a perfect opportunity :P!) [08:23] :) [08:23] didrocks: so after gconf, I think I will fix 661049 and maybe I will check if our workaround to do 685968 is still working [08:24] also I should look into why the decorations of unfocused windows are going away with u-w-d [08:24] smspillaz: I've done 661049 [08:25] smspillaz: look at the status :) [08:25] oh? so double clicking works? [08:25] \o/ [08:25] yeah ;) [08:25] nice [08:25] still need merging, but will be done today [08:25] awesome [08:25] and that's using the PanelTitlebarGrabAreaView is it ? [08:25] 685968 is currently fixing by Jason [08:25] smspillaz: right [08:26] didrocks: there's a feature in compiz we can turn on to fix 685968 [08:26] smspillaz: just the double click signal in nux is broken, so it's a workaround [08:26] didrocks: yeah I noticed [08:26] that's why you didn't want to do it before :) [08:26] well it gets broken by my method of using PanelTitlebarGrabAreaView [08:26] no, even without the MouseDown [08:26] didrocks: ah weird [08:27] yeah, I was first thinking of a conflict, but it isn't [08:27] yeah since I do nasty things in that code like *cough* force ungrab the pointer [08:27] right, but I tried to remove all the grab thing [08:28] but it's really an issue nux and jay didn't have the time [08:28] hence the workaround to emulate double-click myself right now [08:30] cool [08:31] didrocks: ah ok, seems though I don't have this email I sent to jaon because I forgot to backup evolution *cough*. I'll ask Jason to foward it [08:32] smspillaz: ok, no worry :) [08:32] smspillaz: backup are good! [08:32] the server backs up for me :) [08:32] smspillaz: and even with insane mailbox, evolution is working fine [08:32] (like 1h to backup my 4 Gio of emails) [08:33] GiB in english :) [08:33] how do you even process that much email in a day :/ [08:33] well, it's just "picking the right ones" that enables to see what bugs are important [08:33] didrocks: the gconf crash is bug 691561 right ? [08:33] Launchpad bug 691561 in Unity "compiz crash on login to Unity desktop" [Critical,Triaged] https://launchpad.net/bugs/691561 [08:34] smspillaz: btw, I don't see a comment in the wrong window position [08:34] smspillaz: right [08:34] maybe launchpad failed on me again, I'll email it [08:34] smspillaz: you attach something to bug #707853 [08:34] Launchpad bug 707853 in compiz (Ubuntu Natty) "Window geometry isn't restored" [Medium,In progress] https://launchpad.net/bugs/707853 [08:34] which is the "geometry restore" [08:34] but not the "wrong position" [08:34] ah right [08:34] I think they are the same bug [08:35] (I only tested with minimal plugins though, I might be wrong) [08:35] smspillaz: ok, I'll tell you [08:36] smspillaz: btw, just so that you know, Jason and I found why the launcher became so slow for me after your launcher change [08:36] and it seems to be the same issue than alt + tab beeing slow :) [08:37] in fact, you go through all window xid and redecorates them [08:37] this seems to be insanely expensiveeeeeeee there :) [08:43] ahh right? [08:43] that's interesting [08:44] file a bug I'll look into that [08:44] (and yes, redecorating windows is insanely expensive, good fine) [08:44] (good find) [08:44] smspillaz: urgh, ABI break? [08:45] smspillaz: what the "fix window jumping" is doing exactly? [08:45] smspillaz: “implement improved borders & shadows (phase 1): DONE” – improved in which way? :-) I’m still unable to try Unity out myself due to nux crashing on fglrx. I do hope that means reducing the borders to zero pixels. That’s what OSX does and it looks great. :-) [08:46] ion: exactly [08:46] Awesome [08:46] ion: there is now a small area outside of windows to resize in [08:46] didrocks: there was one recently, yes [08:47] smspillaz: I don't get that bug btw, how to trigger it? [08:47] didrocks: also, fix window jumping is when you grab on chromium and then grab it again and it jumps a bit [08:47] smspillaz: remember than any ABI break is ubuntu not being installable for 6 hours [08:47] didrocks: heh [08:47] smspillaz: can that slip by just after alpha2? [08:47] smspillaz: like, I took the two other patchss? [08:47] didrocks: it really annoyed me, it would be good to include [08:48] trying to reproduce [08:48] smspillaz: A fully transparent area that becomes opaque on hover? [08:48] ion: no, it is always transparent (but your mouse cursor changes) [08:48] Alright, sounds great. [08:48] smspillaz: I can't reproduce it on chromium [08:48] smspillaz: just that you ensure ABI break is expensive [08:49] didrocks: well, I'm not really sure if the ABI was really broken by that update, I just bumped the core ABI just to be safe [08:49] urgh :/ [08:49] though if we can get away with not bumping it, then that's probably a good thing [08:49] I did add some new globals though [08:49] I think I'll give you some classroom about ABI breakage one day :) [08:49] good plan :p [08:49] did you change an existing struct? [08:50] don't think so [08:50] and nothing in existing methods? [08:50] negative [08:50] ok, let's try without the breakage [08:50] sure [08:51] urgh, can't seem to reproduce this gconf bug [08:51] let me try it from your packages I guess [08:52] hi njpatel [08:52] hey smspillaz [08:53] I love it when my keyboard layout gets switched to Afghanistan [08:53] smspillaz: our package, without the patch [08:53] smspillaz: and ensure you are using the gconf backend [08:54] right [08:55] smspillaz: Every time keyboard-configuration gets updated it seems to revert the debconf values to defaults which happen to be af-something. I haven’t got around to filing a bug (if one doesn’t already exist). [08:55] ion: yeah, I think a bug has been reported already [11:28] smspillaz: ok, so for your compiz backtrace [11:28] didrocks: so is there a way that I can get a backtrace from the .crash? [11:28] smspillaz: rm all *compiz* in /var/crash [11:28] smspillaz: because I think it's an old one [11:28] ok [11:28] so just ensure we get a fresh one first :) [11:29] you recompiled the packaged, right? [11:29] package* [11:29] oh, right [11:29] it didn't crash [11:29] yeah I recompiled the package without the workaround [11:29] weird [11:29] I'll try rming my config [11:29] ok, but it crashed you just told me? [11:29] yeah, I'm almost sure you are using the ini backend :) [11:31] doesn't say I am [11:31] and weird, even after I rm that stuff unity doesn't show up [11:31] are you in the right session? [11:31] env | grep GDMSESSION [11:31] and env | grep COMP [11:33] nothing in both of them, weird [11:33] oh, well, then again, sshing in doesn't help that [11:34] * smspillaz opens a terminal [11:34] smspillaz: right, those variables are handled by gnome-session, you need a process son of gnome-session [11:34] or spawned by it, at least [11:35] yeah [11:35] didrocks: what's the name of the package to right click in nautilus and open a terminal there? [11:36] smspillaz: I think it's nautilus-action or something like that… I'm not using it TBH :) [11:37] smspillaz: first, look if the gnome-session process has the right env variable in [11:37] if it's ok, you can just export the same in your ssh window I guess [11:37] nautilus-open-terminal - nautilus plugin for opening terminals in arbitrary local paths [11:38] thanks spikeb :) [11:38] that makes my life 435432523456372537483 times easier [11:39] smspillaz: I tend, when compiz is started, to Ctrl + alt + T to open one at start :) [11:39] there's that too [11:40] didrocks: GDMSESSION=gnome [11:40] didrocks: COMPIZ_CONFIG_PROFILE=ubuntu [11:40] is that right? [11:40] excellent :) [11:40] and compiz is working here .... [11:41] even with that patch uncommented [11:41] weird… is it at session start? [11:41] yes [11:41] (also, maybe the unity plugin isn't working right now because of the abi break, that might explain it) [11:41] you remember that it was utterly failing for everyone, we didn't invented that and you had it as well :) [11:42] yeah, I know [11:42] but ISTR that it crashed [11:42] we skipped lunch over it remember? [11:42] (also, that day was _awful_ since the damn airconditioning system was destroying my lips) [11:42] you had that only one day ? [11:43] lucky you [11:45] it was every day [11:45] but wednesday was the worst [11:46] heh, you should have come to the arm room ....air was so dry that we could produce blue sparks with the static electricity in th air (which is indeed helpful in a room full of bare non grounded boards) [11:46] well, Friday was horrible for me as well :) [11:46] haha ogra, did you guys end up frying anything? [11:47] persia made fun out of running around and randomly touching people [11:47] i think Gruemaster fried something [11:47] ouch [11:47] luckily a cheap and easily replaceable board [11:52] smspillaz: btw, did you try to log twice as told on the bug report? [11:52] what do you mean ? [11:52] *shrugh* we already discussed that extensively and it's on the bug report :/ [11:52] like, the crash only happens when you log twice with the same profile [11:53] not on profile change (on the gconf copy) [11:53] log or log in [11:53] log in [11:53] didrocks: right :) [11:54] didrocks: so interestingy enough I haven't hit the crash yet, but uncommenting that patch makes unity not load it seems [11:54] are we enabling dbus by default? [11:54] smspillaz: dbus isn't enabled by default [11:55] smspillaz: unity --reset to reset to the distribution default [11:55] didrocks: ok [11:55] didrocks: that was 000_workaround_gconfbackend_hang.patch right ? [11:55] smspillaz: right [11:55] (also had to disable move_checks_to_compiz.patch since it depended on it [11:55] hum, that's why it doesn't crash [11:56] we don't load the ccp plugin :) [11:56] adapt move_checks_to_compiz.patch rather to still load the ccp plugin without any arg [11:56] (I still think that this patch should be upstream btw, there is no need to "do nothing and screw the user" if you just compiz --replace) [11:57] ideally ccp should be integrated into core [11:57] agreed [11:58] smspillaz: so that's why you don't have any crash I think [11:58] and we should remove command line loading [11:58] didrocks: but I get window borders [11:58] smspillaz: readd it in adapating it [11:58] and compositing [11:58] weird [11:58] you shouldn't :) [11:58] so it is clearly loading that plugin [11:59] I think though people wanted it to be kept a plugin because we might experiment with other settings systems [11:59] or for debugging purposes, etc [11:59] well, first, let's get one right and see after… :) [12:00] for debugging purpose you still compiz [12:00] the patch doesn't prevent that [12:00] didrocks: yeah, although it gives favor to a specific plugin. hardcoded [12:01] didrocks: there is a COMPIZ_DEFAULT_PLUGINS a build-time for a reason you know :) [12:01] smspillaz: the plugin that loads all the other [12:01] oh, but that doesn't work if ccp aint loaded, duh [12:01] right === oubiwann is now known as oubiwann_ [12:01] so really it's just illogical [12:01] didrocks: I think that compiz should read the default values in the metadata then [12:02] maybe COMPIZ_DEFAULT_PLUGINS is just setting the value, not the default value === kklimonda1 is now known as kklimonda [12:02] it's setting the value, but those values need to be read by ccp [12:02] if that doesn't work I need to fix it :) [12:03] didrocks: the defaults still work if ccp isn't loaded [12:03] no it doesn't [12:03] I remember explicitely to try it [12:03] didrocks: for example, alt-button1 is a default value for the move plugin, and that works if you just compiz --replace move [12:03] so what's happening is that we're not reading the default value [12:03] and that is indeed abug [12:03] right, but not if you compiz --replace [12:03] you were speaking about default like COMPIZ_DEFAULT_PLUGINS [12:03] didrocks: it should read the default values for the core plugin though [12:03] didrocks: exactly [12:04] well, I'll let you try to reproduce the crash, on unity right now :) [12:05] didrocks: yeah [12:06] didrocks: what's that magical bzr command you did to fix patches not applying ? [12:06] didrocks: if I comment out 000_workaround_gconfblahblah 060_move_checks_to_compiz doesnt apply [12:06] ISTR you had a magic thing to fix that [12:06] smspillaz: so, bzr bd-do [12:06] quilt push -f 060_move_checks_to_compiz [12:07] do you have a reject? [12:07] yes [12:08] ok, so fix it [12:08] editing the file (not the rejected one, the original) [12:08] to readd the ccp plugin [12:08] then quilt refresh [12:08] exit 0 [12:08] bzr bd… [12:09] if you want to generate the debug symbols, you need to install a package, one sec [12:09] smspillaz: install pkg-create-dbgsym, it will generate the -dbgsym [12:16] kompiling === bregma__ is now known as bregma === gabon_ is now known as gabon === oubiwann_ is now known as oubiwann [12:41] didrocks: so weirdly enough, if I don't have any of those patches applied, unity doesn't get loaded :/ [12:41] so 000_fix_gconf_backend_init_hang is what makes unity load for me [12:41] otherwise it doesn't load [12:42] do you have 060_move_checks_to_compiz still? [12:42] again, without 060_move_checks_to_compiz ccp doesn't load, and so no settings at all [12:42] didrocks: I dropped 060_move_checks_to_compiz and implemented it myself in the source [12:43] doing the same with 000_fix_gconfbackend_init_hang [12:43] smspillaz: how implemented in the source? [12:43] didrocks: exactly the same way as the patch [12:43] you are building the package isn't it? [12:43] yes [12:43] so I just changed the source is build-area [12:43] are you allowed to do that? [12:43] and then? [12:43] bzr bd [12:43] no [12:44] ok, that explains a bit :) [12:44] why didn't you follow what I told you? [12:44] :/ [12:44] It didn't work [12:44] just got more reject errors [12:44] ask rather, it's working, I've done that for years… [12:44] well, reject is because other patch is depending on this one I guess [12:44] just update then one by one [12:44] indeed [12:44] aand how do I do that exactly ;-) [12:44] and think the pain I have at every update :) [12:44] read above? [12:45] quilt push -f [12:45] well the pain is created by this bizzare packaging system :) [12:45] no, the pain is made by upstream not accepting sensible patches :) [12:45] though, I'm glad you're there to help us :) [12:45] so [12:45] bzr bd-do [12:45] quilt push -fa [12:45] -f or -fa ? [12:45] -fa is easier, it will try to push all the patches it can [12:46] then, you have a reject, isn't it? [12:46] right, you missed that earlier :) [12:46] * smspillaz was doing -f [12:46] smspillaz: no, I told you quilt push -f [12:46] which is the same in that case [12:46] just easier to -fa [12:46] right [12:47] ok, so now fix that in the source [12:47] (the rejection) [12:47] as you have the .rej, it should be easy [12:47] then, quilt refresh to refresh the patch [12:47] then quilt -fa to try to push more patch [12:47] what do you mean "fix it in the source" [12:47] since you said going into build-area is wrong [12:48] smspillaz: look at my next sentence [12:48] 13:47:34 didrocks | then, quilt refresh to refresh the patch [12:48] it will bring back your change to the patch in debian/patches/ [12:48] ah right :) [12:48] only the debian/ directory is kept on exit 0 [12:49] so, just keep quilt push -fa ; fix ; quilt refresh ; until there is no more patch to refresh [12:49] then exit 0 [12:49] and bzr bd [12:51] so what exactly does "fix" entail though :) [12:51] just "quilt refresh" ? [12:51] do you know about patch? :) [12:52] if it was only quilt refresh, it would be automated :) [12:52] what, just patch -p0 060_move_checks_to_compiz ? [12:52] no ;) [12:52] whenn, quilt is doing that [12:52] well* [12:52] (-p1 in fact, but that's not the matter) === MacSlow is now known as MacSlow|lunch [12:52] so, when you have a rejection, it means it can't apply a part of the patch [12:53] because of the context is different [12:53] ok? [12:53] right [12:53] so, you need to reapply that part of the patch (it's called a junk or something like that manually IIRC) manually [12:53] for that, if you have a reject on foo.cpp [12:53] to what exactly [12:53] you have foo.cpp.rej [12:54] since build-area is something that you shouldnttouch[tm] [12:54] so, you need to look at the changes and reintroduce them to foo.cpp [12:54] but the foo.cpp in where? [12:54] *shrugh* [12:54] in the build-area [12:54] then, quilt refresh [12:54] ............................................................ [12:54] 13:48:22 didrocks | it will bring back your change to the patch in debian/patches/ [12:54] ahh right [12:55] that's possibly the most confusing thing I've used in a looooonnnnggg time [12:55] quit refresh is basically a diff -Nrup > debian/patches/… [12:55] so just trust me and follow the guide: [12:56] bzr bd-do; repeat: quilt push -fa ; ; quilt refresh; until all patches applies again; exit 0 [12:56] and bzr bd [12:57] ok :) [12:57] that's one of the 3 patch systems we are using if you want more info, look there: https://wiki.ubuntu.com/PackagingGuide/PatchSystems [12:57] * smspillaz wishes things could be a simple as arch linux [12:58] well, I admit that using bzr make it a little more difficult in that case until we can use source 3 for everything :) [12:58] didrocks: quilt-push -fa says no patches in series - I need to go into /debian right ? [12:59] oh right, sorry, you need to export QUILT_PATCHES=debian/patches [12:59] :) [12:59] well, I have that for such a long in my bashrc :) [12:59] huzzah it worked! [12:59] * smspillaz hugs didrocks [12:59] good luck now! ;-) [13:00] and meanwhile… nux is still building :) [13:00] lol [13:02]