From: Mike Frysinger
Subject: [SDL] [patch] remove limits.h hack from linux joystick code
there is a small hack in the current linux joystick code to work around broken
cross-compilers by including more than one limits.h ... at the very least,
this is not arm-specific by any means
i'm of the opinion that libsdl should not contain such things ... people
should be fixing their cross-compilers, not bloating other's source trees
patch attached to clean the cruft
-mike
From: Mike Frysinger
Subject: [SDL] [patch] teach configure about uClinux
the current configure script doesnt know about the uclinux target so this
patch fixes that
-mike
Updated CWprojects.sea.bin with Anders F Björklun's CodeWarrior 5 and 6 projects.
In addition, added the following tweaks from Sam:
* Updated the version string to 1.2.10
* Removed Win32 targets from CodeWarrior 5 projects
* Cleaned up access paths
* Removed unnecessary macos_prefix.h
* Added all necessary StdCLib headers to Support:MacOS
* Added OpenGL 1.2 API headers and libraries to Support:MacOS
* Added testdyngl test program
* Removed OpenGL stub from all test programs except for testgl
So, here's a patch with a reimplementation of QZ_SetIcon() that does what I
described above. I apologize for the delay, I've been quite busy in the last
few days.
It appears to work here on 10.4.5 PPC in the limited testing that I've done;
I'll try to test it on 10.3.9 and 10.2.8 as well, but that might take another
week or so. Please test on i386.
Regarding alpha channels, per-surface alpha, and color keys, the same semantics
as for regular blits to an RGB surface should apply (for the final icon
composited onto the dock), unless I made a mistake - except in one pathological
case: if the icon surface has an alpha channel, its SDL_SRCALPHA flag is not
set (i.e. it has been explicitly cleared, since it's on by default for RGBA
surfaces), and it has a color key, plus an explicit mask was specified (instead
of the one autogenerated from the colorkey), then the color-keyed areas appear
black instead of transparent. I found no elegant way of fixing this, was too
lazy to implement the inelegant one, and decided that it isn't worth the effort
(but if someone disagrees, I can do it).
From: Eric Wing
Subject: Re: Updated tarball
Attached are additional cleanups to SDLMain.m for the files in src/
main/macosx. It looks like somebody already cleaned up most of the
issues. There is one about the class interfaces being empty which
didn't look easy to fix in a meaningful way so I left them.
From: Mike Frysinger
Subject: [SDL] [patch] fall back to using MAP_PRIVATE with mmap() in fbcon dri
trying to use MAP_SHARED with mmap() on uClinux (aka non-mmu) hosts nowadays
will simply fail since the kernel disallows it ... falling back to using
MAP_PRIVATE on these hosts is acceptable ... as such, ive attached a patch
for the fbcon driver that will fall back to using MAP_PRIVATE if mmap() with
MAP_SHARED failed
going by a grep of MAP_SHARED, the only other drivers that utilize this flag
are video/wscons, video/ps2gs, and sound/dmaaudio ... i dont think these
would appear on a non-mmu host so the patch i wrote is restricted to just
SDL_fbvideo.c ...
-mike
Currently there are two Classic-only drivers defined in "SDL_config_macos.h",
and the ones in "SDL_config_macosx.h" are Mach-O/framework-only (i.e. not CFM)
As a workaround, I disabled CDROM and JOYSTICK when building SDL for CFM
Carbon.
The AltiVec blitters don't compile, since they require __VEC__ to be enabled in
order for the compiler to understand "vector" and friends (i.e. do AltiVec)
But you don't want to turn AltiVec on globally, since then the code would only
run on a G4 (there are already runtime tests, before using the AltiVec
variants)
The solution here is to enable AltiVec locally, for the actual AltiVec code.
If SDL_OpenAudio() is passed zero for the desired format
fields, the following environment variables will be used
to fill them in:
SDL_AUDIO_FREQUENCY
SDL_AUDIO_FORMAT
SDL_AUDIO_CHANNELS
SDL_AUDIO_SAMPLES
If an environment variable is not specified, it will be set
to a reasonable default value.
From: Bob Ippolito
Subject: SDL trunk (r2346) and Mac OS X
As for all the Carbon warnings.. the two File Manager ones should be
easy to get rid of, the QuickDraw ones won't be so easy since that
requires actual refactoring.
PBUnmountVol -> FSEjectVolumeSync
FSClose -> FSCloseFork
From: Bob Ippolito
Subject: SDL trunk (r2346) and Mac OS X
The current state of the trunk doesn't quite compile on Mac OS X,
I've attached a series of patches that gets it to compile and kills a
few warnings.
sdl-trunk-r2346-dlcompat-warnings.diff:
The dlcompat thing is just loaded with incorrect type signatures ..
some of them have changed since 10.2 and others are just flat wrong.
This puts it in sync with the latest headers. People on 10.2 will get
the warnings instead of people with new kits.
If your mouse cursor is not within the SDL frame and you warp the mouse, you
get an enter notify, not a motion notify.
Therefore, SDL does not update its internal mouse state.
What's about calling SDL_PrivateMouseMotion even when getting an EnterNotify?
Regards,
Johannes
Here's a patch for SDL_fbevents.c that makes SDL fbcon recognize both left Alt
and right Alt (altgr). Without this, it's impossible to write a lot of
characters on a standard fi-latin1 mapped keyboard, like "@£${[]}\|~".
/Jonatan
