[00:10] Hi. I'm a developer for a GTK+ program (Geany) and we're getting bug reports that our menus that are generated at runtime are not working when using Unity. There's some Glib warnings printed on the console about libdbusmenu. Can someone help me to understand what such a menu must do to properly be supported by Unity's dbus menu? [00:13] (or even point me to the source file I should examine to figure it out) [00:13] bschaefer, did you look at your no-results branch recently? i wonder if the snafu with weird signals got resolved after home-lenses branch got merged [00:14] gord: can you help codebrainz [00:14] bschaefer, and btw hello... and good night :) [00:14] hi bschaefer [00:15] bschaefer: the Alt-F1 branch? [00:15] FWIW, i think it's the same problem as with Eclipse and Lotus (i don't know these, just based on their respective similar bug reports) [00:15] bschaefer: jason recently merged his multimonitor work in which he hacked around with the keyboard stuff [00:15] bschaefer: if you branch doesn't cleanly merge now, and you can't see how to make it, best chase jason :) [00:16] codebrainz: ted is probably the better person to chase [00:16] not online currently? [00:16] no... it seems not [00:16] * thumper wonders who else knows a lot about menus [00:17] bummer. is there a non-launchpad using location to find the source code for this? [00:17] maybe i can track down the source and figure it out [00:17] why non-launchpad using location? [00:17] it doesn't work on my computer, makes the browser freeze [00:18] really? [00:18] wow [00:18] which browser? [00:18] yep, has for a long a time [00:18] iceweasel/firefox [00:18] well that sucks [00:18] firefox works fine for me [00:18] where are you based? [00:18] western canada [00:18] it shouldn't be a problem from there [00:19] have you asked in #launchpad ? [00:19] i think it's a problem with the NVidia driver or mozilla renderer. both projects have bug reports for this issue [00:19] codebrainz: lp:dbusmenu AFAIK [00:19] maybe i can use another computer to clone the repo [00:19] codebrainz: want an http link? [00:19] sure, anything without launchpad.net in the URL seems to work fine :) [00:20] hmm... [00:20] well I know this goes to a different server [00:20] bzr branch http://bazaar.launchpad.net/+branch/dbusmenu [00:20] try that [00:20] ok, it should be fine, i think it's just the web interface causing probs, bazaar should be fine [00:20] IIRC the dbusmenu thing checks menus on startup [00:21] ouch [00:21] but there may be a way to get your app to poke things through after generation [00:21] I've not looked at the code at all though [00:21] just going from what I remember being told [00:21] OK, will see. I thought I remembered reading it hooked into the map event, maybe I can re-trigger one after the new items are added or seomthing [00:21] perhaps [00:21] good luck [00:22] ted works on texan time :) [00:22] OK, thanks a lot [00:22] which I think is UTC-7 (or 6) [00:22] thumper, alright Ill look into his merge [00:22] thumper, hopefully it will go alright and yeah that is the branch Im talking about [00:23] bschaefer: I thought it might be :) [00:23] thumper, also hopefully I can get that no-results branch merged soon! [00:23] bschaefer: we are frozen for the 5.2 release [00:23] bschaefer: but after that, sure [00:24] thumper, have to take a look at the new home-lenses mikkel release [00:24] ack [00:24] thumper, wait what does that mean? [00:24] thumper, so no new ui changes for now? [00:25] thumper, bzr: ERROR: Not a branch: "http://bazaar.launchpad.net/%2Bbranch/dbusmenu/" [00:25] bschaefer: acknowledged (ack/nak from old serial hardware) [00:25] (I don't know bzr at all) [00:25] bschaefer: no changes... [00:25] * thumper looks for codebrainz [00:26] it seems to url escape that plus symbol in the URL [00:26] codebrainz: yeah... that isn't the problem [00:26] (I just pasted your command directly is there something else to do first like clone/checkout?) [00:26] codebrainz: try this "bzr branch http://bazaar.launchpad.net/~dbusmenu-team/dbusmenu/trunk.0.6 dbusmenu" [00:27] codebrainz: no... branch is like clone [00:27] codebrainz: in fact you can write clone and it will do a branch :) [00:27] codebrainz: bzr help branch [00:27] OK, that one's working, thanks thumper [00:28] * codebrainz crosses his fingers that this will be C or Python code :) [00:29] ah, C ... phew [00:29] heh [00:33] thumper, o yeah awesome book suggestions! Currently reading Effective c++ and then More Effective C++, very good information and tips so far [00:34] cool [00:34] they are good books [00:35] yeah and thanks. Ill be asking for some more when im done :) [00:36] thumper, i think it might be libdbusmenu-gtk/parser.c:watch_submenu() [00:36] will have to break out gdb on an ubuntu machine to know for sure [00:38] * thumper nods [00:39] bschaefer: no problem, I have a long list :) [00:41] thumper, is that branch you gave me a good one to make a patch against? [00:41] codebrainz: if that is where the problem is, sure [00:41] (ie. trunk.0.6) [00:41] codebrainz: that is the current trunk [00:41] ah, ok, perfect [00:41] codebrainz: do you have a LP id? [00:41] codebrainz: with an ssh key? [00:41] i had one at one point [00:42] no idea what it was, stopped using since launchpad doesn't work :) [00:43] thumper, sweet! [00:43] codebrainz: can you do me a favour and see if it is still an issue? [00:44] codebrainz: it'd make it much easier to get the code reviewed if we followed the normal process [00:45] thumper, it is still an issue on my desktop, laptop with ATI graphics never had no problem, also it's the one with ubuntu installed, so I can do whatever [00:45] i just hate coding on my laptop [00:45] :) [00:46] have you looked at synergy? [00:46] heh, yeah, could use that [00:46] I use that to use the same keyboard and mouse with desktop and laptop [00:46] or plugin Das Keyboard into the laptop :) [00:46] quicksynergy has a nice gui [00:46] oh... like shiney black? [00:47] well, it has the letters on it, but it's still pretty shiny [00:47] hmm, is there packages with debugging symbols for dbusmenu? [00:48] or should i just compile/install from the trunk? [00:56] where is smspillaz? :P [00:56] Trevinho .. you'll do [00:56] im sure you can fix it [01:01] codebrainz: there are normally -dbg packages available [01:02] i thought so. don't see any though [01:02] and that checkout won't build :( [01:03] codebrainz: what exactly problem is geany having with unity menus? [01:04] oh, i see now [01:04] huh, I could have sworn that was working in Oneiric [01:04] hmm, it's getting a linker error, i checked the build system and everything looks correct [01:05] undefined reference to `g_variant_new_fixed_array` [01:05] Geany has been my every day coding environment since I switch to Gnome3 [01:06] nice. good thing they don't mess with the menu :) [01:06] oh I've been using Unity [01:07] mhall119, using the geany packages or from git? [01:07] I mean that in the upgrade to Gnome3 apps, GEdit stopped working with all the plugins that made me happy with it, so I went in search of something new [01:07] codebrainz: from the Ubuntu repos [01:07] and I could have sworn the globalmenu was working for it [01:07] hmm, maybe you can test using View->Editor->color schemes menu, apparently it's not working anymore [01:07] but I just checked, and it's not now [01:07] it's only that one submenu that's not working afaik [01:08] codebrainz: I see Default and Alternate [01:08] but the menus are in the window, not in the unity panel [01:08] all of them? [01:08] yeah [01:08] (ie. the whole main menu) [01:08] yup [01:08] and it still don't work? [01:09] I see "Default" and "Alternate" under View->Editor->Color schemes [01:09] should there be more? [01:10] codebrainz: http://ubuntuone.com/1JvJ21pD5J00KrTujTnZC4 [01:10] does it switch the theme when you pick the other one of the themes? [01:10] i think it's always showing, just not working [01:10] the syntax highlighting, yeah [01:11] hmm, weird. does it print those dbusmenu warnings on the console when run with geany -v ? [01:11] hmmm, GEdit's menus aren't in the menu bar anymore either [01:11] heh [01:11] codebrainz: It may be my setup that's wonky tonight [01:12] here, the menus are not where we put them (in our window) [01:13] yeah, GTK should automatically export them over DBus to the unity panel (when running Unity) [01:13] something is broken on my end, it's not doing that for any of them [01:14] grrrr, i really wish this would compile here. it's claiming undefined error on a function that exists in my current glib :( [01:14] (or should anyway, according to the docs) [01:14] who is our resident compiz expert? :p [01:15] snadge: what is your problem? [01:15] or at least someone who has looked at the source code and might understand the first thing about the focus code [01:16] short version.. when switching workspaces.. sometimes the input focus of the current window isnt set correctly.. it will receive typing input and visually appear to have focus, but for example.. the autocomplete (in location bar) wont function.. and when right clicking, the menu disappears as soon as you move the mouse over it [01:16] unless you manually click on the windows title.. to re-focus it [01:16] snadge: ok [01:16] snadge: here is the good news [01:16] snadge: I'm told that this is fixed in trunk [01:16] oh man.. i want access to the warez ;) [01:16] snadge: however, we've had some issues getting a release out the door [01:16] snadge: but we should have something workable very soon after alpha 2 [01:17] snadge: for precise at least [01:17] thats cool.. im running precise [01:17] that bug irritates the crap out of me too [01:17] do I have to delete .la files or something to build against GTK+ on ubuntu (I thought i remembered reading this)? [01:17] i wonder if that specific fix is backportable (not that i really care.. all my machines are running precise atm) [01:17] so if it isn't fixed once the release is out [01:17] I'll be banging some heads :) [01:17] snadge: I'm not sure [01:18] does the trunk at least build? is it publically available? :P [01:18] i am familiar with git and building deb packages [01:19] ive been waiting so long for that to be fixed ;) [01:19] its my #1 gripe about unity [01:19] every time it happens .. i have to think of the buddha.. and meditate and go to my happy place [01:21] snadge: well... [01:21] snadge: we are in the process of making things better there :) [01:22] snadge: very soon, we'll have publicly usable daily builds of compiz trunk [01:22] i wonder if these instructions are somewhat relevant [01:22] http://askubuntu.com/questions/28470/how-do-i-build-unity-from-source [01:22] it says updated jan 17th, and is precise specific [01:22] I've not built from source for three months now [01:22] :-| [01:22] ergh.. make install ? what drugs are these instructions on [01:23] snadge: if you have set the install prefix, it isn't so bad [01:23] * thumper looks at the instructions [01:24] it apparently installed into /opt [01:24] installed/installs [01:24] the instructions point it to $HOME/build/unity [01:25] see step 1 [01:25] codebrainz: while I've got you here, have you considered making a Unity QuickList for Geany? [01:26] ok.. well i'll have a look at rebuilding the deb packages instead [01:26] mhall119, i don't think any of the devs are using Unity, and no feature requests yet for it [01:26] shouldn't be too difficult ;) [01:26] codebrainz: where I can file a feature request? [01:26] :) [01:26] mhall119, on the source forge (ugh) site [01:26] and if i break unity horribly.. i'll just switch to unity-2d [01:26] it's linked from the Bugs page on geany.org [01:27] codebrainz: also, I've got my unity situation fixed, and I can confirm that the menu we discussed earlier is populated, but selecting "Alternate" doesn't do anything now [01:27] or reinstall it from precise [01:27] mhall119, ah, good to know it's at least breaking consistently :) [01:27] i got dbusmenu building from apt-get source, so should have it figured out in a while [01:29] thumper, yep, it is breaking where I guessed in the source :) [01:30] (ie. parser.c:watch_submenu() ) [01:31] and since it's a g_return_if_fail() assertion failing, which indicates a programming error, in a static function, I'm going to wash my hands of it :) [01:33] codebrainz: care to file a bug about it? [01:34] codebrainz: you can do that with email you know :) [01:34] meh, I can tell you where the problem is :) [01:34] (ie. exact line) [01:34] I'll file the bug if you like [01:34] if you can tell me what the problem is [01:35] parser.c:1124 it get's a menu widget from the gvalue, but it doesn't guard against it being null before calling watch_submenu() on line 1126 [01:35] it calls parse_menu_structure_helper() the line before, this function internally properly guards against the parameter being null [01:35] codebrainz: http://sourceforge.net/tracker/?func=detail&aid=3481844&group_id=153444&atid=787794 [01:36] I guess somewhere along the line it's trying to get the menushell of the item in question and we haven't added one yet (just a guess) [01:37] codebrainz: got a useful title for the bug? [01:38] codebrainz: and it is on dbusmenu right? [01:38] thumper give me a minute and I'll see if i can fix it and pastebin a patch for the bug [01:39] it'll be against the version in the repos, but the code doesn't seem to have changed in this part [01:39] codebrainz: I'll make a branch for it [01:39] codebrainz: and I can attribute you as the author [01:39] codebrainz: if you want to privmsg me an email address and name :) [01:40] meh, the stopping of bug reports will be more than enough in exchange for fixing it :) [01:40] (all my info is on the geany repository fwiw) [01:44] thumper, my fix gets rid of the warning, but it don't fix the menu not working :( [01:47] hmm, the code seems to assume menus are less than 2 levels deep [01:47] *less than 3 [01:48] codebrainz: and geany's aren't [01:50] the menu in question is 3 levels deep [01:53] nope nevermind, this part works ok [02:29] hmm, there's even a note about Geany in the comments of the code [02:34] thumper: bschaefer: hey [02:34] cyphermox, hello [02:35] thumper: bschaefer: I'll definitely need help with the ibus input issue in dash. I thought I had figured it out but it turns out it's broken elsewhere if I special case space [02:35] bschaefer: hey :) [02:35] I'm pretty sure this can be worked around in just unity, though [02:35] cyphermox, it can be if the dash could handle XIM [02:36] bschaefer: what do you mean? [02:36] xim "seems" to work fairly well, although kind of retardedly. [02:36] cyphermox, njpatel and I were having problems with unity handle normal text when ibus was being used. That is why we used that env var to force Ibus to handle both kinds of txt [02:36] ok [02:37] right, if the env var isn't used you don't get any text in ? [02:37] cyphermox, correct, only what the im engine that is active can do [02:37] arf [02:38] but now, with the env var it's getting space committed before characters because it's parsed quicker [02:38] cyphermox, for example if you are using pinyin and the env var is not set then only the chinese chars will work [02:38] right [02:38] well I don't think it is quicker. So what I think is happening is when the ibus sends the key event through to the im engine ie hangul it returns false for say space [02:39] because hangul doesn't handle that [02:39] but there is still preedit around and then it send that space key event through gtk_filer_keypress which then commits the text BEFORE the preedit is commited [02:39] bschaefer: aye [02:40] I was able to get the space to work in libhangul because I made it handle the space and commit it correclty [02:40] but the IM should still be handling space, because it's a common way to escape out of preedit [02:40] bschaefer: it still seems weird that you'd have to patch every single input method to work correctly then [02:40] cyphermox, yeah I tested the same thing with pinyin and it handles all key events even ones it doesn't do anything with [02:40] bschaefer: hangul isn't the only one. I've tested at least esperanto chu and simi something [02:41] hmm that is really interesting. At the same time im not sure if those engines were tested for handling all input [02:41] right [02:42] which is something ibus does and it is kinda hidden [02:42] otoh nothing else seems to use sync mode at all [02:42] cyphermox, yeahhh, which getting unity to not use that env var would be awesome [02:43] I started looking at some code that we are going to try to land in nux to get around it [02:43] no idea how I can help really [02:43] but when I spoke to dbarth this morning he asked that I at least ping you and thumper about it to make sure it was on your radar [02:44] cyphermox, yeah I got slightly lost on what to do next also; im hoping to get this working in nux to avoid it in unity [02:44] * thumper reads backlog [02:44] bschaefer: how does Nux impact IM exactly? I haven't found much [02:44] cyphermox, well we overload the TextEntery in nux to try and handle IM before it reaches Nux [02:45] ah, so that's where it comes from [02:45] cyphermox, so if we can not overload it and have Nux handle it completely that would be the best case [02:45] bschaefer: however I noticed there are Nux bits missing to catch focux for IMTextEntry, so that IM typing is even possible (via OnFocusIn, OnFocusOut, etc. in IMTextEntry.cpp) [02:45] cyphermox, but Nux doens't use gtk at all so it will be a little challenging getting it working as there isn't much documentation [02:45] can I leave it with you two to email jaytaoko about the issue and ideas for a fix? [02:46] cyphermox, yeah I had a patch for that since an event got removed from Nux to do that [02:46] thumper, yeah I can [02:46] bschaefer: thanks [02:46] bschaefer: I spoke to jaytaoko about that specific idea and basically reverted to what used to be done, with OnKeyNavFocusChange or something? [02:46] bschaefer: feel free to cc me so I stay in the lops :) [02:47] thumper, will do :) [02:48] cyphermox, I tested that signal out and it was only firing when you would leave the input area in the dash [02:48] but it wouldn't be called if you say close the dash [02:48] works for me [02:48] or open it [02:48] really? [02:48] I thought so [02:48] * bschaefer needs to re test that [02:48] I tested that a month or so ago [02:48] i was successfully typing stuff in ;) [02:48] cyphermox, haha, I just had some bools that would get set when Dash was about to open and close to get around that [02:49] cyphermox, a signal would be a much better thing for that though. [02:49] bschaefer: OnKeyNavFocusChange.connect([&] (nux::Area*) { HasKeyFocus() ? OnFocusIn() : OnFocusOut(); }); [02:49] basically, a rehash of what I had seen in bzr history [02:50] cyphermox, Good to know, and yeah that is what it use to look like; Ill have to test that out again and have that merged asap. [02:51] cyphermox, hmm well for the ibus bug I keep saying it isn't unitys fault because you can re create any that bug in GNOME or unity 2d [02:51] bschaefer: I'll have another small fix for handling that IBUS_ENABLE_SYNC_MODE variable if that needs to stay; because programs started from unity then inherit it [02:51] if you force the ibus to use that env var [02:52] guess so [02:52] cyphermox, yeah I played around with that to. To set the env var then unset it when the dash search bar is done getting made [02:52] don't need to [02:52] cyphermox, what was your fix? [02:52] you could just as well leave it there and remove it from context when launching apps [02:53] but either way works :) [02:53] cyphermox, yeah, but its still broken for unity :( [02:54] cyphermox, so ill see what I can get going with this in nux; which I am hoping to get it working with ibuses X11 code because that env var works with that [02:54] ok [02:55] cyphermox, that is the patch that was thought to fix the entire bug, but only if using ibus through XIM... [02:55] cyphermox, thanks for the info on the focus :) [02:55] well, I won't spend much more time on it; but don't hesitate to ping me if you want help testing [02:55] cyphermox, sweet will do. Yeah I have spent to much time on this :( [02:56] I got a few input methods installed and can switch to a full korean desktop to test actual searches [02:57] cyphermox, cool, if any progress gets made ill make sure you hear about it :) [02:58] ok, thanks! [02:59] thumper, also on an update for the key nav mode. There was significant changes to Launcher.cpp; but I think I see where I can make all the changes again to fix it [02:59] thumper, just have to recode it and test it [02:59] bschaefer: ack [03:15] thumper, was that ted? [03:16] codebrainz: um... don't think so [03:16] but may have been [03:16] I think he is just "ted" [03:16] but it is Ted Gould, so it may have been... [03:17] heh, ok. i think i might have tracked down the problem roughly, but not positive [03:17] it's god-aweful the things that need to be done to make GTK+ do something it wasn't designed to do :) [03:26] bzr: ERROR: Target directory "" already exists. [03:26] whut [03:32] snadge, I think you have to set up an SSH key with launchpad. [03:32] err.. but "bzr branch lp:nux nux" [03:32] appears to be working [03:32] even though its meant to default to that.. if you omit the last parameter [03:32] who writes this stuff .. seriously ;) [03:33] snadge, weird, I was getting that error message when I re installed ubuntu and forgot to set up my SSH key [03:33] yeah.. but from the error message [03:33] snadge, haha well it has to be someone :) [03:33] "you must do this to [03:33] write to Launchpad or access private data." [03:33] i dont need to write to launchpad or access private data.. so therefore i dont need to login [03:33] is my assumption [03:34] snadge, yeah...it could be as hard to detect as the halting problem :) [03:34] snadge, joking but yeah Im not sure [03:34] ok i see why the unity build instructions dont build deb packages [03:34] there is no debian directory [03:34] wtf :P [03:35] ok i'll just do it the way the instructions say the ;) [03:35] snadge, that is always the safest way [03:35] i just thought.. maybe.. i can build deb packages, since thats how they end up in ubuntu right ? [03:36] of course its a lot more work.. i would theoretically have to update the control file etc [03:36] but i just think it would be nicer to dpkg -i unity*.deb or whatever [03:37] if i break it..i can always switch back to unity2d [03:37] or reinstall the packaged version [03:38] of course if there was a ppa for trunk.. that would be even easier ;) [03:43] ill stop complaining im sorry.. its been one of those days [03:58] is doxygen really required to build unity? [03:59] or will it build without it.. thats 656mb of crap it wants to install [03:59] Doxygen not found. Documentation will not be built [03:59] it appears not [04:00] snadge, yeah it isn't needed. sorry got busy; you should be able to build from the dep packages but right now trunk requires building a few other things from trunk [04:00] like dee, libunity etc.. [04:01] ok well the cmake step has failed to produce a makefil [04:02] DBO, hey are you still up and around? [04:02] cmake .. -DCMAKE_BUILD_TYPE=Debug -DCOMPIZ_PLUGIN_INSTALL_TYPE=local -DGSETTINGS_LOCALINSTALL=ON -DCMAKE_INSTALL_PREFIX="$PREFIX" [04:02] that part fails [04:03] did you set up $PREFIX and how far did it get? [04:03] CMake Error at /usr/share/cmake-2.8/Modules/FindCompiz.cmake:58 (string): [04:03] string sub-command REGEX, mode REPLACE needs at least 6 arguments total to [04:03] command. [04:03] ahh comment out subdir test [04:05] dbus_g_thread_init(); [04:05] plugins/unityshell/src/unityshell.cpp [04:05] its not called g_thread_init(NULL) :/ [04:05] i'll just ignore that for now [04:08] ok my prefix wasnt set for some reason [04:09] could've sworn i set it [04:12] bschaefer, yeah [04:13] unity/plugins/unityshell/src/PointerBarrier.cpp:68:52: error: ‘XFixesCreatePointerBarrierVelocity’ was not declared in this scope [04:14] DBO, so Im working on the key nav bug. Now that the LauncherController is in place im reworking it with the new code. One thing is mouse_down_outside... signal wont ever get called since GrabPointer() isn't called [04:15] snadge, you have to install libxfixes there was a patch a couple days ago for that [04:16] bschaefer, go ahead and make it call GrabPointer() [04:16] we decided a11y apparently doesn't matter [04:16] DBO, well that causes a crash [04:16] ...no... [04:16] fix the crash? [04:17] DBO, well it is caused in like 3 places when you drag an icon, and when you fix that another one happens; but the problems is if you grab the pointer then you can say drag a window and while quitting form key nav mode [04:17] then you cant* drag a window while in key nav mode [04:18] DBO, but I got around that by using the window_focus changed in Launcher; but now ubus call that gave key_nav mode its own window isn't there. [04:18] Was that intended to be removed? [04:19] key_nav mode its own window? [04:19] DBO, ubus_server_send_message(ubus_server_get_default(), UBUS_LAUNCHER_START_KEY_NAV, NULL); [04:19] when this use to be called [04:20] we call UBUS_LAUNCHER_START_KEY_SWITCHER for everything [04:20] probably should be KEY_NAV for everything [04:20] and then a gvariant arg for whether or not the keyboard was grabbed [04:20] which when that was called I could used the window_focus_change to handle a mouse click out side of the launcher. [04:23] DBO, hmm well I can set it up again to close when clicked outside the launcher but the problem is the Launcher it self cant communicate with LauncherController to quit from key nav correctly (or Im missing a bus call) [04:24] because just called ExitKeyNavMode() in Launcher doesn't quit from the KeyNav mode correctly [04:24] calling* [04:24] none of the keynav logic should be in Launcher [04:24] including the click off behavior [04:24] LauncherController should be the one getting the window_focus_change information [04:25] and then calling KeyNavTerminate(false); itself [04:25] essentially, Launcher as a class knows nothing of this magical thing called keynav [04:25] save for how to render it [04:26] DBO, yeah that is what I wanted to call. Alright I can add the Window Manager to handle a window_focus_change to LauncherController [04:26] DBO, and that way makes a lot more sense and looks a lot better [04:26] easier to read... [04:27] yep [04:27] basically you shouldn't have to modify the work I have already done to launcher [04:29] DBO, Alright, and thanks. Hopefully re coding the key nav fixes should be pretty straight forward [04:29]