Patch by Matthew Smaling

 Matthias      2011-03-20 23:07:02 PDT

On X11, SDL 1.3 does not generate a mouse wheel event. Instead, button
down/buttton up events are generated by SDL. After looking at the code in
SDL_x11events.c, I assume this is due to the fact that X11 does not have a
dedicated mouse wheel event.

I did a little research on the behavior of mouse wheel events on X11 systems.
Apparently, mouse wheel events generate a button down/button up event with the
same time, i.e. with exact same timestamp.

Attached you can find my changes to SDL_x11events.c, which generates SDL mouse
wheel events for those button down events that have a button release event
immediately following it (for the same button, and with the same timestamp).

I did have to make an assumption: As standard X11 implementations know only 5
buttons, I have mapped Button4 to "wheel up" (i.e. +1), and Button5 to "wheel
down" (i.e. -1).

Note that if you include this patch, no SDL button down/up events will be
generated on X11 platforms for mouse wheel events (which is probably a
significant change for those that have programmed their code to work with
them).

- Always use a UIViewController, even if window is not resizable.
- Let non-resizable windows still flip over, so user can hold device with the
correct orientation, but upside down, if that's more comfortable.
- Don't set the UIScreen unless we're forced to, as it resets some state.
- Minor correction with conventions for -[self init] tapdance.

--HG--
extra : rebase_source : d1b861f1174282eccb34f5efd183b03f6efcc2fa

Fixes strange rendering after rotating the device.

--HG--
extra : rebase_source : 6962fd0710f95a81773157507aca653f9a3e1115

Frank Zago to SDL

There's also a patch to adapt one test for the nds, and that adds several
makefiles to compile these tests. Whenever you apply it, could you completely
remove the test/nds-test-progs/sprite and test/nds-test-progs/sprite2
directories ?

Frank Zago to SDL

I've cleaned up a few bugs in the nds code. A few more tests now pass.
There's still a few things to do, but overall I think it's starting to be in a
good shape.

The patch also includes a bug fix for SDL_ConvertSurfaceFormat() (gcc warning).

Implemented thread priority as the 'nice' value on Linux.  High priority threads require root permissions (you shouldn't give your game root permissions though!)

--HG--
rename : test/testhread.c => test/testthread.c

Something inside setStyleMask mucks with the view responder chain, which prevents the listener from hearing the right mouse down events.  We just reset the listener after changing the style to fix this.

Fixed issue when calling SDL_BlitScaled() directly with src or dst rectangles that were out of bounds.

Gleb Natapov to sdl

If application installs SIGINT/SIGTERM signal handler with
sigaction(SA_SIGINFO) syscall before initializing SDL, after
initialization
of SDL signal handler will be reinstalled without SA_SIGINFO flag which
brings havoc when the signal handler is called. The breakage is done by
SDL_QuitInit()/SDL_QuitQuit() function. They use signal() to detect that
signal handler is already installed even in sigaction() is available.