Disko UI Framework porting
adam at os.inf.tu-dresden.de
Sat Nov 5 12:36:29 CET 2011
On Thu Nov 03, 2011 at 12:55:31 +0300, Ivan Loskutov wrote:
> I executed porting of Disko (http://www.diskohq.com/
> ) to use in L4Re. Archive with dependencies and patches
> 1. Differences of Disko original source code
> I changed *.cpp files name extension to *.cc. I didn’t build a library from
> *.cpp files.
> I added files enabling working with framebuffer (mmsgui/fb/mmsfbl4re.cc)
> and L4Re input
> subsystem (mmsinput/mmsinputl4rehandler.cc).
> Virtual file system(VFS) added (mmstool/vfswrapper.cc) for implementation
> of access at application resources (image, xml form files). Libwhefs
> library was used for it (http://code.google.com/p/whefs/).
> 2. Known issues
> The input system link handler has has only the functions I needed. The
> touchscreen calibration function was added to mmsinputl4rehandler.cc, but
> it’s not the best solution. The best solution would be a separate server
> for the input subsystem that routes handled events to other applications.
> To simplify the touchscreen handler coding I added a patch to input. See
> L4Re video API function bits_per_pixel returned incorrect pixel size
> because it was calculated as a sum of R, G, B and A fields sizes. I have
> used RGB32 format with A field length equal to 0 bits. That’s why the
> returned total length was 24 bits instead of 32 bits. See video.patch
There's a difference between bytes_per_pixel and bits_per_pixel. bits-pp
is indeed summing up each field but it does not reflect the size of the
pixel, which why there's bytes-pp.
Adam adam at os.inf.tu-dresden.de
More information about the l4-hackers