[00:42] <sidi> I'm having a pretty insane problem with UnityCore/GLibSource... I made my own GSource class that inherits Source, based on Timeout but with a custom dispatch function... started having systematic SIGSEGV in compiz when it attempts to dlopen libunityshell.so... I went back to GLibSource, and created an exact replica of the Timeout class (exact same code logic, just renamed it to Banana). It still SIGSEGVs if I use that class in my code, whilst using th
[00:42] <sidi> e original Timeout class causes no SIGSEGV. So my question is: How does libunityshell.so import symbols from libunitycore.so exactly? Why is it missing my Banana dummy GSource?
[00:44] <sidi> (bazaar.launchpad.net/~sidi/+junk/unity-banana/changes + apply http://sidi.mupuf.org/files/dummy-banana.diff if anyone wants to replicate)
[00:46] <sidi> (+ add  http://sidi.mupuf.org/files/ZeitgeistManager.{cpp,h} to unity-shared/)
[10:09] <dednick> greyback: is there available for daniels app state changes?
[10:09] <dednick> *silo
[10:09] <dednick> greyback_:  is there a silo available for daniels app state changes?
[10:10] <greyback_> dednick: nope, it's just been reviewed
[10:33] <dednick> greyback_: hm. why have we put nearly every single public method that Application/Session/MirSurfaceItem into an interface?
[10:34] <greyback_> dednick: for testing I believe
[10:34] <greyback_> and eventual migration into unity-api probably
[10:35] <dednick> ok. testing makes sense I guess.
[10:37]  * greyback_ hates C++ or that
[10:38] <dednick> yeah. it's pretty nasty. there's a lot of things that shouldn't be in the interface in there. :/
[10:39] <greyback_> dednick: feel free to say so in the MR comments. I didn't look closely at that I'll admit
[10:40] <dednick> well, i mean shoudn't be in there for third part to use. internal testing it's kinda needed if you want to mock.
[11:41] <pete-woods> hi guys, I can't seem to install the unity8-autopilot package any more?
[11:41] <pete-woods> should I be doing something different now?
[11:42] <pete-woods> ah, never mind
[11:43] <pete-woods> seems to be working now
[13:48] <tedg> mzanetti, One of the support folks sent me a question about using the scope:// URL for accessing a specific scope.
[13:48] <tedg> mzanetti, It doesn't seem to work for me. It brings up the dash, but the dash doesn't switch.
[13:48] <tedg> mzanetti, Do you know what the format of that URL should be?
[13:49] <mzanetti> tedg, yes, one sec
[13:50] <mzanetti> tedg, http://bazaar.launchpad.net/~mzanetti/machines-vs-machines/trunk/view/head:/app/ui/info/Feedback.qml#L28
[13:50] <mzanetti> tedg, that opens the store...
[13:51] <tedg> Huh, I can't get the same format to work with YouTube.
[13:53] <tedg> pstolowski, By chance do you have any ideas here? ^
[13:53] <tedg> It would seem odd the store is somehow special here.
[13:53] <tedg> Also, that URL seems odd in that it only has the package name, not the scope name in it.
[13:55] <pstolowski> tedg, it's not package name, it's *scope id* as understood by scopes machinery
[13:55] <tedg> pstolowski, How do I get from an AppID to a scope ID?
[13:56] <pstolowski> tedg, pete-woods may know if you can get that from the manifest? ^
[13:58] <pstolowski> tedg, but since scope ids are not going to change, why not just put youtube scope id there?
[13:58] <tedg> pstolowski, I don't know what it is?
[13:58] <tedg> pstolowski, That's what I was hoping to figure out :-)
[13:58] <pstolowski> tedg, ah, i understand the problem
[14:00] <pete-woods> tedg: com.ubuntu.scopes.youtube_youtube is the youtube scope ID
[14:01] <tedg> pete-woods, Cool, so it's just the short AppID?
[14:01] <pete-woods> tedg: yep
[14:01] <pete-woods> package ID + hook name
[14:01] <tedg> Cool, that's easy.
[14:01] <tedg> That works, thanks pete-woods, pstolowski and mzanetti!
[14:01] <pete-woods> :)
[14:02] <pstolowski> tedg, also, if you manually go to ~/.cache/unity-scoopes/<scopedir>/ you will find <scopeid>.ini file there, that's you scope id
[14:03] <tedg> Ah, okay. I'm old school and looked in the click dirs :-)
[14:04] <tedg> Where "old school" is "used a technology that's been around a bit over a year or two" :-)
[14:04] <tedg> Canonical-old-school
[14:04] <pstolowski> :D