Most games register a music channel, and when there is no music,
they still stream silence (and run through all the Converter::flow
code!). Scan the buffer for that to pause the AudioTrack. Ugly, but
worth it - reduces CPU usage on many games and hence saves battery life.
Only works in situations without any registered channels (or all paused)
at the mixer (like on the launcher or GMM).
CPU usage before (Galaxy Tab):
~5% scummvm
~15% mediaserver
After:
~2% scummvm
0% mediaserver
;)
All backends that derive from BaseBackend are consoles and/or portables
without any connection from graphics modes to scalers.
The default implementation of resetGraphicsScale() is now a noop, which
seems correct for all current backends.
This fixes unwanted mode changes since 93fdcbfa.
This was probably set to ensure onCreateInputConnection() gets called,
but it is regardless of this flag. Now the soft keyboard doesn't show
up on startup on devices without physical keyboard, which prevents
unnecessary surface changes.
This fixes a segfault when trying to use the OpenGL backend with 16bpp games,
however this does not make Last Express nor Loom PC-Engine Japanese work for
me. They now fail with a "Pixel format not supported" error.
When doing glTexSubImage2D we do not need to assure that the source data has
the same width as the texture when we want to update the data, we just need
to assure that the width * Bpp cound matches the pitch.
If we do not do this, we might end up with a default alignment of 4, which will
fail (as in the graphics will be messed up) in case the screen resolution is
not divisible by 4.
Thanks to digitall for noticing this problem and finding out about
GL_UNPACK_ALIGNMENT.
This fixes messed up graphics, when SDL decides to add padding bytes to the
lines of an SDL_Surface. Formerly the code always calculated the pitch via
w*bpp, which of course does not work in all cases.
Sadly it seems the engines do not care whether their focus rect really fits
inside the game screen. To ease finding such instances (which might cause odd
clipping by the backend) I added a warning for them.
Only applies when using loadable modules.
This requires a second compile run or black voodoo linker tricks.
Not implemented at this time. Disable it, so we get DS build failure
reports from buildbot (which are useful with all the DISABLE_EVERYTHING
defines).
The resulting binary might be too big to be usable.
Rather than that I introduced a function which converts hardware screen
coordinates to overlay / game screen coordinates.
The logic which converts mouse movement events with hardware screen
coordinates to overlay / game screen coordinates is now inside notifyEvent.
This is still broken design, since one should not abuse an observer for that.
dhewg
Johannes Schickel
anotherguest