Comparing software-sdl and software-16-sdl showed many small differences
this makes both engines' code much more alike.
The software-16-sdl was especially buggy, hopefully, this should make
it just as stable as the software-sdl engine.
SVN revision: 63248
These options must be set on the player even if there's no file opened
yet. Thus, just postpone the related commands to when the player has
opened the file already.
SVN revision: 63234
Move the frame drop counter to the shared memory object, so the player
can check its value and not send new frames if we didn't "consume" the
previous one.
Also move the triple buffering pointer changing to the display function,
so it will only lock once when accessing the critical region.
SVN revision: 63220
Note: this function help keep a jump table so we reduce
the need to walk over the complete list to insert one
element. It's of course doesn't make it an O(log(n)) in
access time, but it increase it's cost more slowly.
With 10000 items, you can count around 50 pointers
dereferencing and with with 50000 items around 200 pointers
dereferencing.
Of course the comparison stay in O(log(n)).
SVN revision: 63213
Otherwise if we try to call file_set(<some file>) after file_set(NULL),
it will send a close command to the player for an already closed file.
Also make the example cycle through the list of movies, but still
passing through a NULL file when reaching one of the ends of the list.
SVN revision: 63208
These values need to be reset as soon as a new file is set, otherwise
they can give wrong information to evas about the size of pixels to
render.
SVN revision: 63205
Note: this help the driver to know that we don't it
to swizzle back the pixel from GPU format, because we
are going to replace all pixels.
SVN revision: 63201
Now the player will have the stdin and stdout available for anything
else, and there's no need to make line-buffering with ecore_exe anymore.
SVN revision: 63198
It now checks if was possible to get the shared memory, and returns true
or false. This will command will also be used later to return false for
file opening no matter what was the problem, and maybe the int parameter
will indicate the type of error.
SVN revision: 63197
The problem was that our rbtree implementation doesn't support chaining
duplicates and with no fonts, all the paragraphs had the same y coord (0).
Changing the paragraph index to work on line_no for sorting fixes it,
as line numbers are also correct when there are no fonts.
Thanks a lot to KaKaRoTo for letting me know about this issue.
SVN revision: 63188