[hatari-devel] FYI: new --desktop option

[Eero Tamminen]

Hi,

On perjantai 18 maaliskuu 2011, Thomas Huth wrote:
> Ok, looking at your screenshots, this seems like an little vs. big
> endian problem to me. Could it be that Mac OS X uses different
> endianess modes in windowed and fullscreen mode?
> Assuming that one mode is AARRGGBB and the other is BBGGRRAA, and
> Hatari tries to write 0x00ffffff for white, you will get yellow in the
> other mode instead.
> 
> Could you maybe try to add a call to VIDEL_updateColors() somewhere at
> the end of HostScreen_setWindowSize() to see whether this changes
> anything?

As the colors being set are saved by hostscreen.c, I was thinking that
it could be enough to remap them at that place.  See the attached patch.

Hm.  I replaced HostScreen_updatePalette() with palette remapping, but now
I started to wonder should that sill be called too?


(The patch also tries to speed up videl updates a bit by removing one
intermediate hostscreen function call per pixel.)


> It would also be interesting if you could enable the VIDEL_DEBUG in the
> hostscreen.c file to see if the Dprintf(("Pixel format: ...")) call
> prints something different when you switch to fullscreen mode.

Yes, the shift counts would tell did the format change between fullscreen
& windowed mode.


Btw: at least on Linux, SDL seems to use a separate window for fullscreen
and window mode even if the window resolution & bitdepth is same.  Same
may be the case with OSX.


	- Eero
-------------- next part --------------
A non-text attachment was scrubbed...
Name: hostscreen.diff
Type: text/x-patch
Size: 5409 bytes
Desc: not available
URL: <https://lists.berlios.de/pipermail/hatari-devel/attachments/20110318/0ddbf025/attachment.bin>