Coffice milestone 2

[Sebastian Sauer]

> I am curious what are your longer-term plans - using Qt 4
> or waiting for Qt 5 Calligra port where natively-looking
> QML components for android would be finalized.

Its Qt4 atm. For sure going Qt5 would be the long
term plan but for that to happen Calligra needs to
be ported to Qt5. I think using COffice (means not
having to deal with the kdelibs=>kdeframeworks
switch atm but only with Qt4=>Qt5) it could be done
in just some hours. But its just not my focus yet
cause Android, BB10 and Sailfish work fine with
Qt4 atm.

Native Android integration isn't even on my list. I
think the value for the work needed is to small and
I better focus on a pure QML version that works out
of the box on as many platforms as possible and to
get that QML version nice/functional/beautiful
enough so users not complain :)

Where Qt5 makes lots of sense is performance. I do
take that (and a later switch to QML2) into account
already with the try to make rendering threaded so
integration into scenegraph is easy and by proper
splitting functionality vs QML1 (eg Document vs
DocumentItem). That's also the reason why I went
with rendering on images/textures.

Now beside Qt5 QML2 performance (which alone is
reason enough to target Qt5 asap) and things like
iOS support there is also Ubuntu Phone which is
using Qt5. For timeframes: I hope this can be done
within this year but taken my limited time into
account (sparetime project) and the todo-list it
may not possible for me to work on that topic at
all anytime soon / this year without somebody
else either doing the work or throwing some $ at
me to get it done; Focus, focus, ...

For Calligra Active: Its still on my list but
to proper get it working with QML2 scenegraph
and BB10 Cascades it would need a larger
refactoring to use an own render-thread and deal
with images/textures plus abstract everything
QML1 away to optionally replace with
Cascades/QML2/GL/textures.
That's quit a lot of work and first step would
be to investigate what's needed to move all of
the UI to an own thread what is exactly what I
do atm and the reason why I decided to
a) do from scratch and
b) limit focus on a viewer rather then editor for now.

Using textures also greatly simplifies performance
and pinch-to-zoom. Two of my top goals.