ny00
A minor patch is attached, with the following few changes to testjoystick.c:
- Unused constant definitions have been removed.
- Output for all analog axes is drawn, even when there is an odd number of axes. (I have a controller with 5 analog axes.)
- Buttons are now drawn on two rows, so there's room for more. In fact, it has been used for testing a proposed joystick patch for Android, where large button ID numbers have been involved (20 and up). For more details see http://bugzilla.libsdl.org/show_bug.cgi?id=1700.
- A few adaptations have been done for the Android platform, assuming joystick support is ever applied to it. One of them is that the very first joystick (in the enumeration of all joysticks) is opened for testing, if there is any.
- It is now possible to quit from the calibration by pressing on a mouse button, tapping on a touchscreen or pressing/tapping on the "Back" button of an Android device. Technically, a press on a key identified by key code SDLK_AC_BACK results in that.
ny00
At the moment, the README.android file mentions this as one step in building an app:
"
5. Edit <project>/local.properties to point to the Android SDK directory
"
However, the local.properties file has recently been removed from the repository, so an update is desired. For instance:
"
5. Create <project>/local.properties and use that to point to the Android SDK directory, by writing a line with the following form:
sdk.dir=/Users/slouken/android-sdk-macosx
Phil Sampson
/Library/Frameworks/SDL2.framework/Headers/SDL_stdinc.h:345:28: Implicit conversion loses integer precision: 'size_t' (aka 'unsigned long') to 'unsigned int'
Ryan C. Gordon
With this function...
SDL_UpdateWindowSurfaceRects(SDL_Window * window, SDL_Rect * rects, int numrects);
...is there any reason rects isn't "const SDL_Rect *" ?
All SDL_* functions are always available as real symbols, so you can always
link against them as a stable ABI. By default, however, all the things that
might have dithered down to macros in your application are now force-inlined,
to give you the same effect as before and theoretically better performance,
but still solve the classic macro problems.
Elsewhere, we provide real functions for these things that simply wrap the
inline functions, in case one needs to have a real function available.
Also: this exposed bugs: SDL_abs() does something different if you had the
macro vs the libc function, SDL_memcpy() returns a void* in the function
but not the macro, etc.
tomaszewski.p
According to API description:
SDL_RenderSetViewport - Use this function to set the drawing area for rendering on the current target.
DirectFB renderer in DirectFB_UpdateViewport function just sets clip on surface, instead of moving drawing area.
Attached patch uses set clip rectangle to move {x,y} coordinates during draw and blit operations to be compatible with API description.
tomaszewski.p
DirectFB renderer does not support SDL_RENDERER_TARGETTEXTURE what makes it incompatible with GL or GLES(2) renderers.
Attached patched:
- adds support for SDL_RENDERER_TARGETTEXTURE,
- DirectFB RenderCopy converts destination rectangle from floats to ints,
- fixes whitespaces to be the same in whole file.
Lee Salzman
When using SDL_GL_CreateContext() to create a >= 3.0 version or core/forward-compatible context, internally glXCreateContextAttribsARB is used. Mesa in particular seems to be having trouble with this call and returning all sorts of errors, so it is dangerous to poll for the highest GL version by using calls to SDL_GL_CreateContext unless you are sure, a priori, that the call will suceed, defeating the point of its use.
X11 protocol errors are of the following form, with varying details depending on user, but the cause is always SDL_GL_CreateContext as above...
X Error of failed request: GLXBadFBConfig
Major opcode of failed request: 153 (GLX)
Minor opcode of failed request: 34 ()
Serial number of failed request: 215
Current serial number in output stream: 221
These sorts of errors can be temporarily filtered out by setting an X11 error handler to catch and ignore them, which is safe with respect to SDL_GL_CreateContext behavior because this function is allowed to return NULL to indicate failure.
A patch is attached to do this temporary filtering/catching of errors generated by trying to use glXCreateContextAttribs and friends...
Philipp Wiesemann
SDL's README files seem to contain multiple errors and mistakes. I attached a patch with changes and updates.
README:
* removed Windows CE because no more supported
README-SDL.txt:
* corrected spelling mistake
README.DirectFB:
* corrected spelling mistakes
README.MacOSX:
* corrected spelling mistakes
README.Platforms:
* changed Android version to match AndroidManifest.xml
README.Porting:
* added missing directories from list
* removed cdrom directories from list
README.android:
* updated required NDK revision
* add project.properties to list
* changed lower limit for to android-10 and removed upper
* added a statement why older devices not supported
* added correct dates to statements about OpenGL ES
* added info about Google's device numbers and date
* corrected spelling mistakes
README.gesture:
* corrected spelling mistakes
README.pandora:
* corrected spelling mistake
README.touch:
* changed that values are no in range 0..1
* updated the names of some functions
* updated the notes about usage
* corrected spelling mistakes
* added info that API changed near original author contact
tomaszewski.p
Recent changes in SDL_sysrenderer.h and SDL_sysvideo.h had no impact on directfb backend.
Attached patch:
- updates interface,
- resolves uninitialized variable reading,
- changes logging tio use SDL_Log API,
- updates configure to use DIRECTFBCONFIG variable instead direct call to directfb-config.
