| ali1234 | hah. i know why indicator-sound worked briefly for elfy | 02:03 |
|---|---|---|
| ali1234 | it's a race condition | 02:03 |
| ali1234 | between the indicator and anything else that uses the service | 02:04 |
| ali1234 | because the gtk3 indicators use d-bus service activation, the gtk3 dbus-backends can get loaded before the indicator plugins | 02:04 |
| ali1234 | since they have the same service names, then the gtk2 version won't load | 02:04 |
| ali1234 | so it breaks | 02:04 |
| ali1234 | this is only a problem if you have the gtk2 and gtk3 versions installed at the same time | 02:04 |
| ali1234 | but for some weird reason indicator-sound-gtk2 depends on indicator-sound | 02:05 |
| micahg | ali1234: hrm, that's bad, it makes sense though as the gtk3 version is progressing without concern for the gtk2 version, I guess our only hope might be to try for the gtk3 indicators, but I'm very reluctant still | 02:11 |
| micahg | maybe we can rename the gtk2 dbus services to not share the same names | 02:11 |
| ali1234 | that will break all apps that use sound menu won't it? | 02:12 |
| ali1234 | like rhythmbox won't be able to put it's controls in there | 02:12 |
| micahg | hrm | 02:12 |
| ali1234 | i already thought of this :) | 02:12 |
| micahg | how did this all work before? | 02:12 |
| ali1234 | before d-bus service activation was not used at all | 02:12 |
| ali1234 | this is caused by indicator-ng | 02:12 |
| ali1234 | not gtk3 directly | 02:13 |
| micahg | I don't see indicator-ng in saucy | 02:13 |
| ali1234 | indicator-ng is new in saucy | 02:13 |
| ali1234 | it's not a package | 02:13 |
| micahg | oh | 02:13 |
| ali1234 | it's just "the new way" | 02:13 |
| micahg | aside from uploading the old gtk3 indicator stack, I'm not sure what to d | 02:14 |
| ali1234 | i have a feeling that having indicator-sound-gtk2 depend on indcaitor-sound is unneccessary | 02:15 |
| ali1234 | and that removing indicator-sound and rebooting would make indicator-sound-gtk2 work correctly in xfce | 02:15 |
| ali1234 | (force removing without removing dependants that is) | 02:15 |
| micahg | well, it used to have files needed for the gtk2 versions to work | 02:15 |
| ali1234 | which files? | 02:16 |
| micahg | ali1234: can you try that with equivs (or I can build you test packages) | 02:16 |
| micahg | idr, but that's why it was there originally | 02:16 |
| ali1234 | i can't test, because all my stuff is patched for gtk3/indicator-ng | 02:16 |
| micahg | I wish I would've know this 3 weeks ago :( | 02:17 |
| ali1234 | i'm not sure if that stuff i just said about d-bus activation is actually true | 02:18 |
| ali1234 | i need to look at the raring packages really | 02:18 |
| micahg | it would be nice to know for sure, we've got 4 weeks left (2 for beta 2) and I'd hate to drop the rest of the indicators | 02:19 |
| ali1234 | the rest? | 02:19 |
| micahg | well, if they're broke, I'm not sure we have any other choice but to drop or jump to patched gtk3 | 02:20 |
| ali1234 | there is only sound and appindcator that still have gtk2 versions | 02:20 |
| ali1234 | and appindicator is currently broken in gtk3 | 02:21 |
| ali1234 | i need to make several vms... | 02:21 |
| micahg | ok, please let me know when you find something | 02:23 |
| ali1234 | ok. it won't be tonight, it's 3am here | 02:24 |
| micahg | understood, whenever you get to it will be appreciated | 02:24 |
| Unit193 | micahg: Just dropping the dep on on the gtk3 one might be something elfy can try, if it's a quick package install. | 02:35 |
| micahg | Unit193: sure, I can generate something, can you tell me which packages? (appindicator and indicator-sound)? | 02:35 |
| Unit193 | That'd be my guess, I was going to quickly try something in VM, but just broke that... | 02:46 |
| ali1234 | ok i've got the vms going | 03:07 |
| ali1234 | hmm interesting | 03:09 |
| ali1234 | in raring, the service running is the gtk3 indicator service | 03:09 |
| ali1234 | but it works | 03:10 |
| ali1234 | so i guess the d-bus backends must have previously shared the same api, but now the gtk3 one has changed | 03:12 |
| ali1234 | moving aside the d-bus service does not cause the gtk2 one to get loaded instead | 03:15 |
| ali1234 | so the only way to really fix this is going to be to update the indicator to the new d-bus api | 03:16 |
| micahg | ali1234: let me know if that's fairly straightforward | 03:27 |
| Noskcaj | micahg, any progress on the pkg-xubuntu team? | 05:31 |
| micahg | Noskcaj: no, not yet | 05:31 |
| Noskcaj | :( | 05:31 |
| micahg | let me see if I can poke someone this weekend | 05:32 |
| Noskcaj | We'll probably need to package simple-scan and alacarte with the team too. | 06:35 |
| micahg | ok, sure | 06:36 |
| ali1234 | micahg: the whole dbus service has been rewritten from scratch | 13:38 |
| ali1234 | somewhere in this lot is what needs to be backported to -gtk2: http://paste.ubuntu.com/6110453/ | 13:39 |
| ali1234 | even the build system has been changed from make to cmake | 13:41 |
| brainwash_ | I disabled the language indicator by setting "show-language-selector=false" in lightdm-gtk-greeter.conf, but one small white pixel still remains. | 13:45 |
| brainwash_ | Anybody else noticed that? | 13:45 |
| jjfrv8 | brainwash_, I can reproduce that | 13:55 |
| brainwash_ | worth creating a bug report? I assume that (almost) nobody is affected by this little tiny glitch, it's just a white pixel after all :D | 13:58 |
| micahg | gah | 15:46 |
| Noskcaj | Did we do anything to xfce4-session this cycle? I get a heap of error messages when i turn on my PC from apps that it decided where open. | 20:24 |
| Noskcaj | Also, usb-modeset-data won't upgrade | 20:24 |
| ochosi | brainwash_: yeah, unfortunately it's known, we'll fix it eventually | 20:27 |
| micahg | hrm, ok, I'd like to know when that breaking change to the indicators landed if it was part of http://launchpadlibrarian.net/150184786/indicator-sound_12.10.2%2B13.10.20130829-0ubuntu1_12.10.2%2B13.10.20130913-0ubuntu1.diff.gz or something previous | 20:45 |
| micahg | if it was after feature freeze, we can push back | 20:46 |
| ochosi | i think at least the bugreport was earlier than FF | 20:55 |
| ochosi | micahg: any chance you can take a quick look at the gmusicbrowser patch? | 21:02 |
| micahg | yeah, in a bit | 21:05 |
| ochosi | nice, thanks | 21:17 |
| knome | jjfrv8, your merge proposal has been merged | 21:37 |
| ali1234 | micahg: i looked at the sound-indicator thing and i can't see any reasonable way to make it work | 21:37 |
| ali1234 | the whole dbus service looks like it has been rewritten and even the build system was changed from make to cmake | 21:38 |
| micahg | ali1234: can you tell which upload broke it? | 21:38 |
| ali1234 | how do i see a list of uploads? | 21:38 |
| micahg | https://launchpad.net/ubuntu/+source/indicator-sound/+changelog | 21:39 |
| ali1234 | the breakage is in the gtk3 indicator anyway - the gtk2 indicator has not changed at all since raring | 21:39 |
| micahg | right, I'm wondering if it was after feature freeze (Aug 29) | 21:39 |
| micahg | if so, I can most likely get it reverted and/or get help to make it work for us | 21:40 |
| micahg | otherwise, we might have to try to push our gtk3 stack in | 21:41 |
| ali1234 | at a guess this one: https://launchpad.net/ubuntu/+source/indicator-sound/12.10.2+13.10.20130716-0ubuntu1 | 21:41 |
| ali1234 | "sorry for the big changeset" | 21:41 |
| ali1234 | can i download those debs and try them? | 21:41 |
| micahg | sure | 21:41 |
| micahg | just click on the arch you want | 21:41 |
| ali1234 | ok i'll do a bisect and see what hapens | 21:42 |
| micahg | if it was back in July, it becomes our fault for not testing enough | 21:42 |
| ali1234 | well, not really | 21:42 |
| micahg | if it's after Aug 29, they'd need a freeze exception and I'd have some leverage | 21:42 |
| micahg | I mean, some communication would've been nice, but if it's before feature freeze, there's not much we can do | 21:43 |
| ali1234 | Reported by Elfy on 2013-07-22 | 21:43 |
| micahg | :( | 21:44 |
| ali1234 | that is the date the package after that one was uploaded | 21:44 |
| ali1234 | still think it's that one though | 21:44 |
| ali1234 | the fix would always have been "update the gtk2 indicator for the new api" anyway | 21:45 |
| ali1234 | but from looking at the diff, that would be almost as complex as pushing the gtk3 stack instead | 21:46 |
| micahg | right | 21:46 |
| micahg | I'm thinking to do the latter with a release note | 21:46 |
| micahg | unfortunately, I don't see myself having this ready before next sunday | 21:46 |
| micahg | Canonical seems to love breaking APIs... | 21:48 |
| micahg | LTS+1 is the time to do that, not LTS-1 | 21:48 |
| knome | are we cranky today? | 21:49 |
| micahg | maybe :) | 21:49 |
| * knome gives micahg a cookie | 21:49 | |
| micahg | ali1234: do we know if the gtk3 stack will work with the new APIs | 21:49 |
| * micahg has sugar overload | 21:49 | |
| ali1234 | new APIs? | 21:49 |
| * knome takes the cookie away from micahg and gives him a glass of water | 21:50 | |
| ali1234 | which ones? | 21:50 |
| micahg | ali1234: indicator stack | 21:50 |
| ali1234 | (it works with all of them actually) | 21:50 |
| ali1234 | i ported the indicator-ng changes - the indicator-sound change is between the indicator itself and it's dbus backend and they both live in the same package if you're on gtk3 | 21:50 |
| micahg | knome: I wasn't very happy to found a decision I made to preserve stability left us with a totally broken indicator stack due to existing changes this cycle | 21:50 |
| knome | mhm | 21:51 |
| micahg | ali1234: ok, thanks for your help with this, we'll have to do the best we can I geuss | 21:51 |
| knome | grumpyness doesn't fix the issue though, does it? :) | 21:51 |
| ali1234 | the sound thing does boil down to dbus activation - you can only have one activation script and it activates the gtk3 dbus backend which *used* to work wit the gtk2 indicator but now they are out of sync | 21:51 |
| ali1234 | there is a workaround on the bug actually | 21:51 |
| ali1234 | someone put a test into the activation script so it runs the gtk2 backend if XUBUNTU_DESKTOP is set | 21:52 |
| micahg | which bug is this? | 21:52 |
| ali1234 | https://bugs.launchpad.net/ubuntu/+source/indicator-sound/+bug/1208204/comments/27 | 21:53 |
| ubottu | Ubuntu bug 1208204 in indicator-sound (Ubuntu) "indicator-sound no longer functions with xfce4-indicator-plugin" [High,Triaged] | 21:53 |
| ali1234 | (and the original report on 22 july is marked a duplicate of that) | 21:53 |
| ali1234 | i'm just going to post a comment there summing up the situation | 21:53 |
| ochosi | at least i can say i've tested the gtk3 variant for quite a while now and it seems as stable as the gtk2 variant used to be | 21:55 |
| Unit193 | But, need to roll something from git head. | 21:56 |
| micahg | I wonder what's working, hacking up the dbus service file or pushing gtk3 stack... | 21:56 |
| micahg | s/working/worse/ | 21:56 |
| micahg | I have to run out for a bit, I'll be back | 21:57 |
| ochosi | well tbh the gtk3 stack seems a bit more sustainable | 21:57 |
| ochosi | hacking on the dbus stuff to keep gtk2 running for one more cycle... | 21:57 |
| ochosi | that sounds a bit like a waste of time | 21:57 |
| micahg | that's assuming it's just the service file | 21:58 |
| micahg | otherwise, agreed | 21:58 |
| knome | just for the sake it's not LTS, but LTS-1, +1 for the gtk3 stack... | 21:59 |
| ali1234 | just patching the service file to always run the gtk2 backend does work (tested by others) but it also breaks other desktops since you can only have one service file | 22:00 |
| ali1234 | adding a complex /bin/sh test to it might have weird side effects and is not tested | 22:00 |
| ochosi | can't you add a session-query like the sound-indicator does? | 22:00 |
| ali1234 | add it to what though? | 22:00 |
| ochosi | some wrapper script | 22:01 |
| ali1234 | that's what the workaround in comment 27 does | 22:01 |
| ochosi | ah right, hadn't looked at it so closely | 22:01 |
| ali1234 | Exec=/bin/sh -c 'if [ -n "$(pidof xfce4-session)" ]; then /usr/lib/indicator-sound-gtk2/indicator-sound-service;else /usr/lib/x86_64-linux-gnu/indicator-sound/indicator-sound-service;fi' | 22:01 |
| ali1234 | instead of just of Exec=/usr/lib/x86_64-linux-gnu/indicator-sound/indicator-sound-service | 22:02 |
| Unit193 | Wouldn't it be better ot -f the file? | 22:02 |
| ali1234 | well i don't know about that - it's just a proof of concept and i don't think anyone even really tested it | 22:02 |
| ali1234 | i agree that gtk3 would be less hassle long term | 22:02 |
| ali1234 | nobody is maintaining this gtk2 stuff any more | 22:03 |
| ali1234 | that's the whole reason for the problem | 22:03 |
| ali1234 | it's out of sync with gtk3 (again) | 22:03 |
| ali1234 | and it will only keep happening | 22:03 |
| ochosi | very true | 22:03 |
| ochosi | it has happened before | 22:03 |
| ochosi | we lost the messaging-menu that way as well | 22:03 |
| ali1234 | yeah, i know all about it :( | 22:04 |
| ali1234 | as i said, it's why i didn't use xubuntu before | 22:04 |
| ochosi | well nice to have you "on board" now | 22:05 |
| ochosi | i prepared a tiny patch for the power-indicator to work with xfce4-powerman | 22:07 |
| ali1234 | i only found out about the gtk3 progress because someone posted the irc log of the xmir meeting on reddit | 22:08 |
| ochosi | nice :) | 22:08 |
| ali1234 | i don't care about xmir but the indicator stuff got me interested again | 22:08 |
| ali1234 | i have not tested the power indicator cos i am on a desktop... don't even know how to make it show | 22:08 |
| ochosi | ah righ | 22:09 |
| ochosi | t | 22:09 |
| ochosi | well it's very basic tbh | 22:09 |
| ochosi | it shows a power-icon and a very short menu | 22:09 |
| ochosi | if the drivers support it, it shows brightness controls | 22:09 |
| ochosi | or maybe it needs more gnome-powerman for that | 22:09 |
| ali1234 | i've seen it in unity on my laptop | 22:09 |
| ochosi | and then it shows the de/charging time/percent | 22:09 |
| ochosi | and it has a "preferences" menuitem | 22:09 |
| ali1234 | but i rarely use that | 22:10 |
| ochosi | which i patched to open xfce4-powerman if it's run in a xubuntu session | 22:10 |
| ali1234 | ah yes... all the indicators have checks to run different prefs panels in different desktops | 22:10 |
| ochosi | (agreeable it does create a problem in case someone uses gnome-powerman in xubuntu) | 22:10 |
| ali1234 | there are lots in sound-indicator | 22:10 |
| ochosi | i know | 22:10 |
| ochosi | we could also fork the session menu if we wanted | 22:10 |
| ochosi | shouldn't be all-too-hard | 22:10 |
| ali1234 | why? | 22:11 |
| ali1234 | just to get the xfce session-save dialog? | 22:11 |
| ochosi | nah, i actually prefer xfce's action.menu | 22:13 |
| knome | be back later. | 22:15 |
| bluesabre | ali1234: hi, I'm trying to get the gtk3-indicators working, I've followed ochosi's instructions here: http://dpaste.com/1381815/ | 23:07 |
| ali1234 | ok | 23:08 |
| bluesabre | currently, everything seems to be working except for the sound indicator, see: http://imagebin.org/271044 | 23:08 |
| ali1234 | ther was one missing line iirc | 23:08 |
| ali1234 | oh? | 23:08 |
| ali1234 | that looks like it is working to me? | 23:08 |
| bluesabre | the sound indicator's volume slider and playback buttons are missing | 23:08 |
| ali1234 | yes... theme perhaps? | 23:09 |
| bluesabre | using greybird from git, ochosi added support | 23:09 |
| ali1234 | can you check the log... | 23:09 |
| ali1234 | ~/.cache/xfce4-indicator-plugin.log | 23:10 |
| ali1234 | look for missing icon warnings and so on | 23:10 |
| ali1234 | no sure if they would show up there or not | 23:10 |
| bluesabre | aha | 23:11 |
| bluesabre | WARNING Gtk gtk_menu_new_from_model: cannot find type 'com.canonical.unity.slider' | 23:11 |
| bluesabre | WARNING Gtk gtk_menu_new_from_model: cannot find type 'com.canonical.unity.media-player' | 23:11 |
| bluesabre | WARNING Gtk gtk_menu_new_from_model: cannot find type 'com.canonical.unity.playback-item' | 23:11 |
| ali1234 | yeah that's it | 23:11 |
| ochosi | ali1234: i vaguely remember having seen bluesabre's problem before here myself, but i'm not sure why... | 23:11 |
| ali1234 | ido provides several new menu models | 23:11 |
| ali1234 | you might have the wrong version | 23:11 |
| ali1234 | there is a gtk2 and a gtk3 version | 23:11 |
| bluesabre | ah, ok | 23:12 |
| ali1234 | or maybe you don't have it at all | 23:12 |
| ali1234 | i think andrzejr added possibility to compile without it | 23:12 |
| ali1234 | then it won't get loaded | 23:12 |
| ali1234 | so... make sure libido is installed and then rerun autogen.sh and rebuild the indicator-plugin (last thing you built) | 23:12 |
| bluesabre | ok, I'll try that now | 23:12 |
| ali1234 | there are a couple of those warnings i still get even with ido installed properly | 23:14 |
| ali1234 | WARNING Gtk gtk_menu_new_from_model: cannot find type 'com.canonical.indicator.switch' | 23:14 |
| ali1234 | but i can't find what library is supposed to supply that one and there is literally zero documentation from any of this com.canonical.indicator stuff | 23:15 |
| ochosi | yup, lack of docs really blows | 23:15 |
| ochosi | good luck bluesabre and night everyone! | 23:15 |
| bluesabre | woohoo, that did it, thanks a lot ali1234! | 23:16 |
| ochosi | hm, guess that should be included in the instructions... | 23:17 |
| bluesabre | :) | 23:17 |
| ali1234 | ido is indicator-ng stuff, so it's newer than those instructions :) | 23:17 |
| ochosi | ali1234: i just received an update here on the sound-indicator that supposedly fixes scrolling | 23:18 |
| ochosi | will try to install that update and then report back whether it works (again) | 23:18 |
| ali1234 | "just"? | 23:18 |
| ali1234 | i updated earlier and no change | 23:19 |
| ali1234 | earlier today i mean | 23:19 |
| ochosi | well i "just" ran update-manager now | 23:23 |
| ali1234 | ok i see the fix | 23:23 |
| ali1234 | but 1204036 | 23:23 |
| ali1234 | bug 1204036 | 23:23 |
| ubottu | bug 1204036 in The Sound Menu "[Regression]Mouse wheel to change volume stopped working on Saucy" [High,Fix committed] https://launchpad.net/bugs/1204036 | 23:23 |
| ali1234 | i have that version: it's still broken in xfce | 23:24 |
| ali1234 | andrzejr said this was the case | 23:24 |
| ochosi | hm, then it's still broken with the xfce-plugin | 23:24 |
| ali1234 | ok, at least i can try to fix it now | 23:25 |
| ochosi | yeah, but it sounded like he wasn't sure why | 23:25 |
| ali1234 | maybe related to why child-added events aren't passed | 23:25 |
| ali1234 | wait no, that is also broken on unity too | 23:25 |
| ali1234 | can't be that | 23:25 |
| ochosi | m, i can confirm that it's broken | 23:26 |
| ochosi | in xfce | 23:26 |
| ochosi | ok, bedtime (this time for real) | 23:26 |
| ochosi | ttyl ali1234 | 23:27 |
| ali1234 | night | 23:27 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!