Sam Lantinga <slouken@libsdl.org> [Mon, 30 Jan 2012 21:42:24 -0500] rev 203
Switched the SDL 2 autoconf macro to AM_PATH_SDL2
Sam Lantinga <slouken@libsdl.org> [Mon, 30 Jan 2012 20:10:01 -0500] rev 202
Updated Android Makefile for SDL2
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 13:24:55 -0500] rev 201
Updated framework project for SDL2
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 11:12:19 -0500] rev 200
Switched from SDL_net to SDL2_net
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 11:01:02 -0500] rev 199
Removed old unsupported project files
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 11:00:10 -0500] rev 198
Updated to version 1.2.9
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 10:59:38 -0500] rev 197
Updated to version 1.2.9
Sam Lantinga <slouken@libsdl.org> [Sat, 28 Jan 2012 10:30:55 -0500] rev 196
Fixed bug 1402 - Error compiling 1.2.8
Sam Lantinga <slouken@libsdl.org> [Fri, 20 Jan 2012 19:16:04 -0500] rev 195
Fixed bug 1388 - Debian patch(es): endian-ness issues
manuel.montezelo@gmail.com 2012-01-20 12:26:38 PST
2) Second bug report: "SDLNet_(Read|Write)(16|32) assume host endianness is
always LE"
http://bugs.debian.org/217221
"When SDL_DATA_ALIGNED is defined, SDLNet_Read16 etc. are defined as
byte-swapping macros. However, since network order is big endian, these
macros should not swap anything on big endian systems. Patch attached."
The accompanying patch "endian.patch" doesn't apply cleanly with 1.2.8 but it
does with 1.2.7. In any case it's also easy to understand: it removes
conditional execution of code (SDL_BYTEORDER != SDL_BIG_ENDIAN) and keeps only
the (SDL_BYTEORDER == SDL_BIG_ENDIAN) version.
Sam Lantinga <slouken@libsdl.org> [Fri, 20 Jan 2012 19:10:48 -0500] rev 194
Fixed bug 1388 - Debian patch: ARM support
manuel.montezelo@gmail.com 2012-01-20 12:26:38 PST
1) First Debian bug report: "bad aligned access considerations on ARM"
http://bugs.debian.org/212570
"ARM CPUs like my iPaq's fail on unaligned reads as well. See the following
code for instance: [...] Moreover, AFAIK unaligned reads are always decomposed
into two aligned reads on modern CPUs. I think it would be wise to drop them
completely and always set SDL_DATA_ALIGNED to 1."
The accompanying patch "arm.patch" is trivial to understand and apply.