Mir Qt dependencies update.

[Robert Carr]

Whoops! Forgot a section.

Mir trunk is sufficient now (death to input-conglomeration).

Cheers!

On 05/13/2013 11:12 PM, Robert Carr wrote:
> Hi! Hope everyone is well.
>
> Spent some time yesterday rounding up the Mir Qt dependencies 
> (primarily for Unity), and investigating Gerry's segfault with 
> Qt-inprocess on android.
>
> First! The issue I believe was straightforward, platform-api-mirserver 
> functioned on Mesa (where EGLNativeWindowType==mir::shell::surface) by 
> passing the surface through directly. Android requires an abstraction, 
> i.e. the mir::graphics::InternalClient that landed in Kevin's branch 
> to get at the native window.
>
> So I made an updated platform API branch:
>
> lp:~robertcarr/platform-api/mir
>
> Containing both mirserver and mirclient versions.
>
> Unfortunately there are two issues:
>
> 1. mir::graphics::InternalClient requires a Surface to get the 
> EGLNativeDislayType, whereas Qt wants it while initializing the 
> QScreen (before a Window[Surface] has been created). So we will need 
> to iterate on this in the morning I think.
> 2. platform-api is in the process of a big refactoring, working with 
> ricmm, hoping to land at the beginning of next week, until then will 
> use this branch.
>
> It is now built with
> -DAPPLICATION_UI_IMPLEMENTATION=mirserver (or mirclient)
>
> Qtubuntu mirclient (working) is available at 
> lp:~robertcarr/qtubuntu/mirclient
> and Qtubuntu mirserver at lp:~robertcarr/qtubuntu/mirserver, currently 
> it will not link until the native display type is fixed.
>
> The EGL header hack is still required, will sync up with Chris on this 
> tomorrow.
>
> Didn't start thinking about integration testing seriously yet.
>
> Hoping: Tomorrow, get qtubuntu inprocess working on android, early 
> next week get it all landed (at least the unity/mir dependencies, 
> while working out the unity/mir kinks concurrently).
>
> Thanks!
> Robert
>
> P.S.
> What is the best Unity list.
>