=== godbyk_ is now known as godbyk-laptop [06:35] morning [06:44] night :) [06:48] :) [07:57] greetings everybody! [07:58] hello MacSlow [07:59] hey kvalo, kamstrup [08:11] good morning [08:18] kamstrup: morning. I'm sure you have missed my merge requests ;) so here's a new one: https://code.launchpad.net/~kvalo/indicator-network/libconnman-vala/+merge/47366 [08:19] kamstrup: this one was so strange, I didn't get it: http://bazaar.launchpad.net/~kvalo/indicator-network/libconnman-vala/revision/153 [08:28] kvalo: is it |just| the cheader_filename that you've changed in the .vapi? [08:29] kvalo: because in that case it could make sense to a 'sed' invocation in your Makefile.am to get the Connman-1.0.vapi to be purely autogenerated [08:36] kvalo: needs fixing [08:41] kamstrup: the regexp wouldn't be simple, because there are two different filenames [08:42] kvalo: ah, bugger [08:42] yeah, it really sucks [08:42] kvalo: although I guess you could fix that by making one super header that includes the two, no? [08:43] kamstrup: can you check if you have libreadline-dev installed? I'm guessing configure check doesn't fail if it's missing [08:43] kvalo: and generally I push for single header includes everywhere anyway. Including discrete headers is a path of pain for consumers and maintainers alike [08:44] kamstrup: so you are saying that I should have libconnman.h which will contain all classes? [08:44] ah, it will just include all the class .h files [08:44] makes a lot of sense [08:44] kvalo: a libconnman.h that just includes manager.h and service.h or whatever you need [08:44] kvalo: in fact for all my public [08:44] kvalo: (damn return key!) [08:45] kvalo: in fact for all my public libs i *force* this and raise a compile time error if you try to include one of the subheaders [08:46] kvalo: like http://bazaar.launchpad.net/~libzeitgeist-developers/libzeitgeist/trunk/view/head:/src/zeitgeist-monitor.h#L21 [08:46] kvalo: but of course this is not as important for an internal lib [08:47] I'll add it anyway [08:47] this gives absolute freedom for me to restructure the internal source layout, which is otherwise impossible [08:48] kamstrup: thanks for the tips [08:50] yeah, configure doesn't fail if readline is missing [08:50] * kvalo starts reading the autoconf bible [08:54] kvalo: meh, readline doesn't have pkg-config integration... [08:54] kvalo: you can just do: echo '#include ' | gcc -E - > /dev/null 2>&1 [08:55] kvalo: in configure.ac and check $? == "0" [08:55] kamstrup: I'm using AC_CHECK_LIB already, I was just missing the error case [08:55] kvalo: ah, ok [08:55] kamstrup: but yeah, readline is _old_. [08:55] kvalo: but it works :-) [08:56] sure does [09:03] kamstrup: just to keep merge requests simple, I will create a new branch for the automatic vapi generation [09:03] kamstrup: ok? [09:04] kvalo: suer [09:04] sure [09:32] njpatel: ! [09:32] kamstrup, ! [09:32] njpatel: I was thinking... [09:33] uh-oh [09:33] njpatel: why is the group_model per PlaceRendererInfo and not just shared globally across a PlaceInfo (like the sections_model already is)... [09:33] njpatel: considering https://wiki.ubuntu.com/Unity/Places [09:34] njpatel: sorry, the correct terminology would be RendererInfo and PlaceEntryInfo respectively in the above [09:34] kamstrup, I think it's because different renderers might have different groups, where sections would remain constant across the place entry [09:34] njpatel: at least it makes for some oddities in the place daemons [09:34] oh, wait, I se what you mean [09:35] kamstrup, I guess it could be global [09:35] njpatel: indeed, and you can still have different renderes even if you use the same label for the group, that would just be two entries in the group model with the same labeæ [09:35] right [09:36] njpatel: so I think it's a taad more consistent, and it's one less DeeSharedModel on startup (they are not entirely cheap) [09:36] ((because of the dbus interactions) [09:36] kamstrup, this isn't something we need to change, though, right? I mean I've just come out of doing this and if you make me go back to gdbus I might kill you [09:36] oh, you do want to change [09:36] * njpatel books tickets to Denmark [09:36] njpatel: i don't have to [09:37] njpatel: it's because I am sitting here adding the Control Center stuff to the apps place [09:37] kamstrup, I'd say leave it for now, we've got enough to do for A2. Maybe once the places stuff has settled down next week? [09:37] njpatel: and the current spec makes this rather ugly in my code [09:37] njpatel: ok, I'll put it in the gdoc of pending changes [09:37] thanks [09:42] kamstrup: I updated now https://code.launchpad.net/~kvalo/indicator-network/libconnman-vala/+merge/47366 [09:42] kvalo: ok, on it [09:44] kamstrup: thanks [09:52] dbarth_, a11y call now? [09:53] kvalo: approved, with a small request... [09:54] kamstrup: hey, I can't say no to you ;) [10:13] API, dbarth_: a11 weekly call now? [10:13] hey aruiz [10:13] hey rodrigo_ ! [10:13] :-) [10:25] API, hmm, about your last comment in my branch, the PanelViewAccessible has several children, not only 1 (the layout), as shown on accercizer [10:25] API, not sure what you mean? [10:27] rodrigo_, it has several because you are exposing layout content [10:27] API, right, and you'd prefer to just expose the layout as the only child? [10:27] no [10:27] please read again [10:27] my last review [10:28] your way to implement that could be also valid [10:28] but, there is any reason to do that? [10:28] I mean that the normal is the other way [10:28] expose the only child [10:28] becase after all [10:28] is the only child of the panel [10:28] your way is also valid [10:28] hmm, but what do we need the layout for? [10:28] but, imho, it should be a reason to do that in this way [10:29] and why repeat the code to expose the children of a layout? [10:29] we have that code on nuxlayoutaccessible [10:29] so it is valid for any layout used [10:29] hmm, ok [10:30] so, I just change it to return 1 child (the layout) then, right? [10:32] well, unless you find a good reason to do that in the other way ;) [10:32] yes, [10:32] you can take basewindow as reference [10:32] kamstrup: I want to create a uber test script/makefile target which would run all possible tests there are and print an "OK!" banner if there are no errors. would makefile target "allchecks" or similar be the best solution? [10:32] no, just didn't think we needed the full real hierarchy, but if that's how everything is done, I'll change it [10:32] so yeah, don't have a good argument to convince you :-D [10:33] kvalo: just stick this in your makefile for s in test_foo test_bar test_baz; do echo $s OK; done [10:34] kvalo: is everything using the gtester framework? [10:34] kamstrup: most are, but I want to also run distcheck [10:35] kvalo: check how I do it lp:libunity. You need a Makefile.decl copy pasted to your source root from there [10:35] kvalo: and then include it in your test makefiles [10:35] kamstrup: thanks, I'll check it [10:47] gord: could you file a bug saying that we need multimonitor support in the switcher and staticswitcher plugins so I don't forget :) [11:02] smspillaz, not right now, if i remember when i'm less busy though :) [11:03] :p [11:25] njpatel, a serializedstring gvariant would look like something like this no -> "(s)", "thing" [11:25] for the gsettings command line tool [11:29] ronoc, erm, no idea dude. kamstrup ? ^ [11:30] ronoc, njpatel: ehhh, probably? What does "serialized" mean here? [11:32] kamstrup, njpatel from the docs of the gsettings command line tool the set command takes key and then value, "The value is specified as a serialised GVariant " [11:33] ronoc: it probably refers to the format used by g_variant_print()/g_variant_parse() [11:33] ah yes [11:33] thx guys [11:39] ronoc: so what I see here is that a variant created with g_variant_new ("s", "Hello") prints as ('Hello',) [11:41] kamstrup, tried that already, doesn't like it [11:41] ronoc: he - then I'm at a loss [11:42] kamstrup, no worries - just trying to helpful on a bug report ... [11:43] ronoc: look at the format from a get command - the docs say it's the same [11:45] ronoc: gsettings get org.gnome.desktop.default-applications.terminal exec [11:45] ronoc: so it's a 'single quoted string' [11:46] kamstrup, the format of the value in my hash is an array of strings [11:46] so I have been trying gsettings set com.canonical.indicators.sound blacklisted-media-players @as ["rhythmbox"] [11:46] since get returns [11:47] @as [] [11:47] ronoc: ['foo', 'bar'] [11:48] hello [11:48] kamstrup: gsettings get com.canonical.Unity.Launcher favorites [11:48] kaleo: Where's the original Kaleo?! === kaleo is now known as Kaleo [11:48] :) [11:48] * Kaleo is me [11:48] Kaleo: oh, hi Kaleo, you here too? [11:48] ;-) [11:49] after a good night of rest yes [11:49] hehe [11:49] kamstrup, still not working, a key error is now the problem even though it has no problem fetching [11:49] weird [11:49] will come back to it later [11:50] ivanka, remember me? (last week xD) [11:51] kamstrup, got it [11:52] gsettings set com.canonical.indicators.sound blacklisted-media-players "['rhythmbox']" [11:52] ronoc: right, looks sane === MacSlow is now known as MacSlow|lunch [12:25] kamstrup: actually I can't generate the vapi automatically. I just remembered that I did other manual changes as well [12:26] kamstrup: I think I just leave as it is for now and revisit then the api is stable, and hopefully tools are more stable [12:26] kvalo: right, no worries [12:38] seb128: do you still have the issue with alt + mouseclick not dragging your window? [12:43] didrocks, still get it [12:43] you don't? [12:43] using alt-click-dnd doesn't move anything there [12:44] seb128: no, I never had it [12:44] seb128: can you try this (maybe it's a perf issue that some people have): ccsm -> move plugin -> opacity back to 100% [12:44] (and check that alt + button 1 is set as well) [12:45] didrocks, it's not happening in a guest session, just checked [12:45] so don't bother [12:45] seb128: ok, something weird again in the session :) [12:47] hate compiz, it's still crashing every second ccsm run [12:48] seb128: do you enable/disable pluginis? [12:48] no [12:48] plugins* even [12:48] just run ccsm [12:48] oh, weird, I didn't with that [12:48] every second run compiz crashes [12:48] ok [12:49] I can move the dialogs which are not focused [12:49] I have some random crashes though [12:49] that's weird [12:49] like if I've to dialogs on screen I can alt-click and dnd the one which is not focussed [12:49] but not the one which is focussed [12:49] seb128: maybe related to the new decorator? (just a stupid guess) [12:49] there is still the older one if you want to give it a try [12:50] didrocks, that fixes it [12:50] not that stupid then \o/ [12:51] in the guest session, which decorator is running? (should be the unity one still…) [12:52] or not [12:52] didrocks, it seems restarting the decorator just made it work once [12:52] didrocks, I will keep watching for it but don't bother with that for now [12:52] ok, that's weird :/ [12:52] it doesn't happen in my guest session so it's likely a profile thing [12:52] I will try to debug it later [12:52] seb128: ok, keep my in touch :) [12:52] but I've other tasks I want to get done before [12:52] k [12:52] session restart brb [12:52] thanks :) [13:08] Kaleo, so (finally) looking into the merge, do you think we could drop some of the depends in u-2d-d-s to recommends ? [13:08] the dependency list is quite long [13:10] ogra: hmmm, that's kind of a broad question [13:10] ogra: any specific depends you have in mind? [13:10] nautilus looks like a good candidate [13:11] metacity too (since you apparently can also use compiz) [13:11] ogra: metacity is a no go [13:11] why ? [13:11] ogra: we want to run by default on hardware without hw accelerated OpengGL [13:11] -g [13:11] well, there is custom patch in metacity like the decorator one [13:11] if i can use compiz instead there needs to be an option to uninstall metacity [13:12] ogra: and we do not QA on Compiz very much [13:12] Kaleo, by default the unity-2d package will pull in metacity [13:12] ogra: so for now it's experimental [13:12] ogra: which is normal [13:12] a [13:12] ah [13:12] we install recommends by default [13:12] right [13:12] its just about the optional removing [13:12] ogra: I see [13:13] and i can see people using it with pcmanfm or compiz [13:13] ogra: what do you set in the session file? [13:13] metacity ? [13:13] ogra: to back up a little here our package u-2d-d-s should depends/recommends most of the same things that Unity deps/recommends [13:13] didrocks: yes [13:13] probably even with xfce-session if there is nothing that hardcodes gnome-session functions [13:13] so, the session won't start if someone uninstall metacity [13:13] didrocks: good point [13:13] didrocks, no, but you can modify it manually [13:14] right, but I'm unsure about letting people shooting on their feet :) [13:14] * didrocks tries to think how to do that [13:14] well, we will get complaints if we have to hard deps [13:14] didrocks: ogra: what package should I look at that contains Unity's session? [13:14] Kaleo: gnome-session [13:14] didrocks: thanks [13:14] i'm fine going with what we have but try to think ahead [13:15] ogra: maybe gnomewm needs some tweaking === agateau_ is now known as agateau [13:15] so things that arent absolutely needed should be removable [13:15] Kaleo: you need to look at /usr/share/gnome-session/sessions/ubuntu.session [13:15] if we indeed have hardcoded bits that really depend on gnome stuff there is no way for recommends === MacSlow|lunch is now known as MacSlow [13:15] but i see for example ubuntu-wallpapers in the deps [13:16] ogra: strictely speaking, I can only see the decorator patch [13:16] and light-themes [13:16] these seem like recommends candidates [13:17] ogra: light-themes does not have to b there at all AFAIK [13:17] we often have people building custom images for unsupported arm HW ... [13:17] ogra: so, we have a long list of deps [13:17] ogra: it needs cleaning while keeping all existing functionalities [13:17] i just try to think of that guy who wants to run everything xfce and pcmanfm with custom themes [13:18] Kaleo, right, thats why i ask :) [13:18] ogra: and we should look into Unity's deps [13:18] ogra: and mimick what is done there [13:18] ugh [13:18] they are way bigger than ours :P [13:18] ogra: sure [13:18] ogra: but for example the indicator-* are recommends for them [13:18] http://packages.ubuntu.com/natty/unity [13:19] heh [13:19] yeah, because without the indicators… [13:19] and it's only recommends [13:19] the required bits are: [13:19] compiz of course [13:19] unity-common which is internal [13:20] libglib2.0-bin for the gsettings binary [13:20] the indiocators would have been the bits i wouldnt have made recommends :) [13:20] funny [13:20] python for some migration script and the unity binary [13:20] nux-tools for the detection bits [13:20] unity-asset-pool for the icons [13:20] ogra: well, if you install unity, you want to have a way to disconnect from your session for instance [13:21] or you maybe want the clock :) [13:21] right [13:21] ogra: http://ietherpad.com/2s0NezImPo [13:21] thats why i mean i wouldnt have made indicators a recommends [13:21] ogra: I started putting the plan there [13:21] ogra: oh, you mean a dep? [13:22] ogra: let's figure out the whole changes there, agree on it, and then push it [13:22] ogra: we have some people who wants to remove the appmenu [13:22] ugh [13:24] Kaleo, wallüpapers and themes should be recommends there [13:26] Kaleo, hmm we could drop wallpapers and themes and move them into the netbook/desktop seeds [13:26] (i think they are in there anyway) [13:27] apt-cache rdepends answers part of the question I suppose [13:27] ogra: they are [13:27] right, lets drop them [13:27] didrocks: http://ietherpad.com/2s0NezImPo contains the discussion in a more efficient format [13:27] ogra: can you leave them on the document please [13:27] oops [13:27] back [13:28] thanks [13:28] what about metacity ? [13:28] if its hardcoded in the session file we indeed need to keep it [13:29] how about gnome-session ? can we use other session managers without breaking ? [13:31] ogra: don't forget that this pacakge is meant to install the session in GDM [13:31] and what about the dep on unity ? that should be lowered to the bits we actually need i guess [13:31] ogra: that's its purpose [13:31] (unity-common/-asset-pool) [13:31] Kaleo, what if i want to use it with kdm ? [13:31] ogra: we don't depend on these [13:31] right, we need to clean up what's in -common [13:31] Kaleo, right [13:31] I'll have a look to ensure you have all you need [13:31] Kaleo, we depend on unity though [13:32] ogra: not in 2d-default-settings AFAIK [13:32] or am i misreading the doc ? [13:32] didrocks, so drop g-s-d and nautilus ? [13:32] unity-2d-launcher depends on unity-asset-pool [13:32] ogra: yes, you should IMHO [13:33] can you tag it DROP ? [13:33] sure [13:33] thanks [13:33] so only the pixbuf engine left [13:34] should be brought by something else IMHO [13:34] let me check [13:34] ogra: finally, you have more dep than unity has :) [13:34] well, can i install unity-2d and still have a working UI ? [13:34] with these deps [13:34] didrocks, could you look at #684275 (upload request for the ginn project)? [13:34] else we need to re-add recommends [13:35] ogra: that's the question, I don't really know for that one :) [13:35] bug #684275 [13:35] Launchpad bug 684275 in ginn "New source package: ginn" [Medium,New] https://launchpad.net/bugs/684275 [13:35] didrocks, can i do that with unity ? [13:35] bregma: excellent, will get to it that week, thanks :) [13:36] Kaleo, i agree with your comment in the doc [13:36] ogra: we don't use any theme engine [13:36] it should still be possible to just install unity-2d [13:36] without a toplevel metapackage [13:36] didrocks, thank, there are a few more multitouch upload requests pending I'll follow up on later [13:36] so things we need for functioning should be recommends [13:37] bregma: got it, do not hesitate! :) [13:37] ogra: you need rather a metapackage then, but we can have the same rationale with every apps… === klattimer1 is now known as klattimer [13:38] didrocks, we will have a unity-2d package, i need to migrate unity-2d-default-settings into something [13:38] in that case… :-) [13:39] at some point i would like to merge the other binary packages of the unity-2d source package into that too, but Kaleo wants them distinct for now for flexibility [13:39] ogra: what what [13:39] ogra: that's a very different problem [13:39] thats probably natty+1 [13:39] ogra: merging the binaries is an upstream thing [13:39] ogra: that will not happen [13:39] Kaleo, no, its a packaging thing [13:40] ogra: which is equivalent [13:40] you currently build four binary packages [13:40] while unity doesnt [13:40] ogra: we distribute separate binaries for reasons that make the binary package existence valid [13:40] with my merge we will have one meta and four binaries [13:40] ogra: Unity _cannot_ [13:40] ogra: Unity does not have multiple binaries [13:41] and it would make sense *at some point, i dont say _now_* to merge all binaries into the meta [13:41] ogra: I hint never :) [13:41] that doesnt make sense [13:41] ogra: we want to keep the flexibility for users [13:41] ogra: that means separate binaries and separate binary packages [13:41] why do you actually need these binary packages? do you expect anyone to use either of them standalone ? [13:42] ogra: yes [13:42] well, for instance, we have in the unity package the plugin and the unity panel service, but you can't use one without the other one [13:42] didrocks: that is actually going to change AFAIK [13:42] didrocks: because it's likely Unity 2d is going to depend on the unity panel service [13:42] Kaleo: if you are going to migrate to it, for sure [13:42] yeah, I'll make a separate bin package for it then [13:42] well, if the intention is to use the launcher standalone (and if it acrtually works without any of the other) then i agree indeed [13:42] didrocks: right [13:42] ogra: that is the intention [13:43] ogra: and it does work [13:43] ogra: in fact some people already use it athat way [13:43] yeah, tne i fully agree [13:43] -a [13:43] *then [13:43] Kaleo: just keep me in touch once all the seperating work done, so that we can work on the minimal package req. [13:43] :) [13:43] didrocks: sure, I will let you know when the time comes [13:43] i thought there were inter-dependencies between them [13:43] ogra: hopefully they will be resolved one day [13:43] ogra: oh, sorry [13:43] ogra: I misread you [13:43] ogra: there are no inter dependencies [13:43] ogra: or rather [13:44] right, sorry for assuming that :) [13:44] ogra: if they are they are going to be fixed soon [13:44] ogra: there is no runtime dependency that's for sure [13:44] ogra: you can run all four binaries on their own [13:44] k [13:44] ogra: there is some packaging dependenc for wrong reasons [13:45] ogra: and bug reports to fix them [13:45] +ies [13:45] so back to our cod [13:45] *doc [13:45] right [13:45] * ogra cant type today [13:45] pixbuf engine [13:45] so, I don't know for that one [13:45] what about that ? [13:45] let me bzr qlog [13:45] i cant imagine any qt binary using it [13:46] so it would only be pulled in as a meta dep [13:46] which makes it a recommend [13:48] https://bugs.launchpad.net/watuppa/+bug/645977 [13:48] ogra: we use a bit of GTK internally [13:49] ok [13:49] heh, not allowed ... but i belive you :) [13:49] so keep it as dep [13:49] ogra: no [13:49] ogra: that's not the reason [13:49] ogra: we depend on libgtk2.0-0 in unity-2d-launcher already [13:50] ogra: looking at the bug report I think that it could be made a recommends [13:50] k [13:51] ogra: so [13:51] ogra: what is the difference between [13:51] ubuntu-wallpapers and nautilus [13:51] for example [13:51]