/srv/irclogs.ubuntu.com/2010/09/08/#ayatana.txt

sanguine1hello i need some help with my python script and the indicator applet02:19
sanguine1i can't seem to get the applet to recognize that my script is running02:19
sanguine1it only shows the arrow next to the entry when i run the script directly from the command line02:19
sanguine1never mind, i had the wrong path for the .desktop file02:34
lucidfoxIt seems the libappindicator bindings in Maverick are bugged08:30
lucidfoxI had to manually edit them to make them work08:30
lucidfoxer08:30
lucidfoxVala bindings, I mean. The .vapi file08:30
kklimondalucidfox: that's quite possible, what have you changed?09:41
lucidfoxkklimonda> I'll post a patch against the Maverick version in a moment09:41
lucidfoxthe patch is against the .vapi file, though, and it's autogenerated09:42
kklimondalucidfox: it may still help09:45
lucidfoxkklimonda> bug #63299609:51
ubot5Launchpad bug 632996 in indicator-application (Ubuntu) "Vala bindings for libappindicator are incorrect (affected: 1, heat: 6)" [Undecided,New] https://launchpad.net/bugs/63299609:51
kklimondaah, yes..09:51
kklimondathis is what you get when you don't follow GObject naming scheme09:53
lucidfoxWell, "../src" was just developer oversight, evidently :)09:56
kklimondameh, whole API breaks all conventions of GObject so gobject introspection generates garbage and so does vapigen..09:58
Cimihi chaotic09:58
kklimondayou have _get_menu and _set_menu that take/return GtkMenu argument and then you have a menu property that is a string..09:59
lucidfoxwell yes, that's odd10:00
lucidfoxsame with status10:00
kklimondaand icon - you have _get_icon and _set_icon but property is icon-name10:01
kklimondaI've complained about it months ago10:01
kklimondaapparently not loud enough10:01
kklimondaand now we are probably stuck with this excuse for an API anyway10:02
kklimondalucidfox: for now the only thing to do is to ship your own .vapi with project10:06
kklimondaI10:06
kklimondaI'm going to poke tedg when he's online about it but, as I've written in my comment, fixing it is not trivial.10:07
lucidfoxWell, we all know what happens with new GUI elements10:09
lucidfoxFirst they appear in external libraries. Second they appear in GTK, the GTK developers do them right, and the external libraries are deprecated. :p10:09
lucidfoxIs there no way to provide hints for gir/vapigen to generate the correct bindings?10:10
kklimondalucidfox: not that I know of10:14
kklimondalucidfox: not for this kind of problems10:14
kklimondaif bindings weren't generated from .gir you could do basically everything in the .metadata file10:14
kklimondabut Ayatana devs believe it's wrong way of doing that because it can't be easility tested when you do changes to the library.10:15
kklimondaso they have opted for .gir10:15
kklimondathe poor API has been one of the reasons the inclusion of libindicator into Gtk+ has failed.10:16
lucidfoxThen the GTK developers should write GtkPanelIndicator or something that offers a better API for this :p10:16
kklimondathat's not really how things work.10:17
kklimondathere are no Gtk+ developers per se after all..10:17
lucidfox...no wonder it's stagnating then :(10:17
kklimondaindeed10:17
kklimondabut the situation is getting better recently.10:17
kklimondastill if anyone were to provide GtkPanelIndicator it would have to be Ayatana10:18
kklimondaand those who provide code would have to justify why do we need a duplicate of GtkStatusIcon10:19
lucidfoxPah! Just deprecate GtkStatusIcon! :p10:19
lucidfoxI think GTK is architecturally superior in its core concepts, but... Qt has Trolltech/Nokia's corporate backing, and it really shows. In just about everything :(10:19
kklimondawell, I agree that Gtk+ is part of a better platform.10:20
kklimondaand it's written in C10:21
kklimondawho in their right mind would even touch C++... ;)10:21
lucidfoxyeah :S10:21
lucidfoxI would actually like to see more application developers moving away from C to Vala, Python or C#, and see larger parts of the core platform stack rewritten in Vala10:22
kklimondalucidfox: Vala, in my opinion, while a great idea, is still a too thin wrapper over C10:23
Cimiiainfarrell: ivanka: do I need to move to London in order to speak with chaotic? :) I guess I'll rent a house :D please poke him *live*!10:24
lucidfoxWell10:24
CimiI've don esome cool updates10:24
lucidfoxif I wanted to attract new developers to a "sexy" platform to write applications for, I'd point them to Mono/Gtk#, but the very mention of Mono is enough to cause half the community to go RAWR MICROSOFT TAINT10:25
ivankacimi: I am working at home today and I think they are doing a fire alarm test in Millbank which means walking all the way down and then up again10:25
kklimondalucidfox: do we really need half a dozen of different VMs running on  our desktops? :/10:26
lucidfoxHardly half a dozen10:26
lucidfoxOne CLI to rule them all10:27
kklimondalucidfox: every time I see how much ram do gwibber use.. or even the smallest applet I'm irritated.10:27
lucidfoxWell, Python is just :S10:27
kklimondawriting in python or mono seem to disable something in programmers' minds and they stop tracking how much memory do their applications use.10:28
lucidfoxor it's just the VMs that are written that way10:28
lucidfoxPython has its legitimate uses as a scripting language, but writing serious desktop applications in it just sounds odd to me10:28
kklimondalucidfox: that too is a problem - but the fact that they hide memory managment doesn't help.10:28
lucidfoxMore importantly, to attract new developers, I think providing better toolchains and IDEs is more important than providing better languages10:29
kklimondalucidfox: meheheh.. :/10:29
lucidfoxQt has the right idea with Qt Creator10:29
Cimiivanka: oh thank you boss ;)10:29
kklimondaa good IDE built around clang with great profiling tools, support for valgrind, good gdb integration, with a nice GUI.. we can dream of it ;)10:30
lucidfoxand *not* using autotools.10:30
kklimondameh, I got used to autotools.10:31
kklimondasure, they ain't pretty10:31
lucidfoxYou did, but requesting new developers to understand the arcane syntax is just too much, I think10:31
lucidfoxnot just syntax - all the arcane gotchas that come with it10:32
ivankacimi: hehe10:32
lucidfoxnot to mention that a build system based on shell scripts is unportable by design10:32
lucidfox"@Andr?s, why would Gnome want to switch to CMake? From my experience with it, it?s a pretty rubbish system? I had a lot of trouble with it. For example, it couldn?t even do pkg-config without a 3rd party script."10:37
lucidfoxEpic facepalm10:37
davidbarthCimi: ok, wassup with the resizer?10:37
davidbarthCimi: hi, btw ;)10:37
Cimidavidbarth: what about the cody's patch to add resize grippers without needing to add them in the statusbar?10:40
davidbarthCimi: i don't think he has it, or that it is ready10:43
Cimidavidbarth: so it won't land in maverick?10:43
davidbarthCimi: plus it's past feature and UI freeze now10:44
davidbarthCimi: let's check with bratsche once he wakes up for the level of complexity, but i doubt it's that easy, so which means a risky change to propose at that stage10:45
Cimichaotic: ?11:06
=== bilalakhtar is now known as cdbs
=== cdbs is now known as bilalakhtar
=== bilalakhtar_ is now known as bilalakhtar
=== MacSlow is now known as MacSlow|lunch
=== agateau_ is now known as agateau
=== MacSlow|lunch is now known as MacSlow
=== rgreening_ is now known as ghost
=== ghost is now known as rgreening
bratscheCimi, davidbarth: I have a patch somewhere (forgot what computer it's on now), but it hasn't received much testing or review.  I seriously doubt seb128 would accept such a patch at this point, and I kind of wouldn't blame him. :)15:06
Cimiok15:06
bratscheI can look for the patch though.15:06
bratscheAnd then we can figure out what we want to do with it.15:07
bratscheI found my original patch, but this isn't the latest one: https://bugzilla.gnome.org/show_bug.cgi?id=59172115:08
ubot5Gnome bug 591721 in gtk "provide a way for all resizable windows to have a resize grip" [Enhancement,New]15:08
davidbarthbratsche: putting it in the same gtk daily build as the one that i'd like to have for argb15:08
bratscheThere was one that also had kind of a weird hack to either GtkScrollbar or GtkScrolledWindow, and I guess I never posted it on bugzilla.  I'll try to find it.15:09
Cimichaotic: ?16:47
chaoticCimi: hi, been in meetings17:01
=== oubiwann is now known as ouibiwann
=== ouibiwann is now known as oubiwann-away
=== oubiwann-away is now known as oubiwann
=== funkyHat_ is now known as funkyHat

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!