From: Jens Leuschner
Subject: Small patch for libSDL
In newer XFree86 releases in the German keyboard the "AltGr"-Key no
longer produces the keysym "Mode_switch", instead it produces
"ISO_Level3_Shift".
I have created a small patch against SDL-1.2.6 (which cleanly applies to
the current CVS-SDL too) that makes SDL recognize ISO_Level3_Shift as
AltGR key.
From: Thomas Jarosch
Subject: Re: [SDL] DSPF_RBG15 in src/video/directfb/SDL_DirectFB_video.c
In my local DirectFB 0.1.19 copy, DSPF_RGB15 is an alias for DSPF_ARGB1555.
They seem to have it removed in DirectFB CVS.
As DSPF_RGB15 was only an alias, the attached patch
can safely be applied to SDL and won't brake older DirectFB versions.
From: Max Horn
Subject: Small libtool patch (framework related)
here is a small patch for ltmain.sh. It corrects a problems with the
libSDL.la file, which prevented applications using a version of libtool
w/o framework support (read: all currently released libtool version) to
not being able to link with SDL anymore. I also notified the libtool
maintainers, as the same problem exists in libtool CVS.
From: Stephane Marchesin
Subject: Re: [SDL] 3DNow! detection on a P4 system
Well... I found another bug : 3dnow was detected on P2/P3 cpus. So I
took one more look at the code and found a mistake in
CPU_getCPUIDFeaturesExt :
The condition for having extended cpuinfo is that when we query extended
cpuinfo we get a result >= 0x80000001. So we must exit if eax < 0x80000001.
The attached patch does that.
From: Max Horn
Subject: Auto hide mouse & other changes
the attached bug adds the auto-hide-mouse feature I talked about
earlier. Turned out it was a lot simpler than I thought, simply by
using our existing code :-). I actually spent much more time on fixing
various bugs in the code and correcting (IMO) some behavior (although,
due to the lack of real specs for SDL, it's probably arguable what
'correct' means...).
* adds auto (un)hiding of mouse depending on whether it is in- or
outside the game window
* computation of course coordinates is correct now (it often and
reproducible got out of sync with the old code, since the NSEvent
window was in some cases *not* our window anymore, so locationInWindow
returned wrong results)
* added a method which at any time returns the mouse coords, relative
to our window
* fixed handling of lost/gain input/mouse/app focus "events"
From: Max Horn
Subject: SDL_HasAltiVec; BUGS file
the attached patch adds SDL_HasAltiVec to SDL CVS. Note that at this
point, this only works on MacOSX (and maybe darwin). I don't know how
to properly add a test for e.g. Linux/PPC at this point. I found an
email which might help in doing so:
http://zebra.fh-weingarten.de/~maxi/html/mplayer-dev-eng/2003-01msg00783.html
However, since I have no way to test on a non-OSX PowerPC system, I am
not comfortable blindly adding such code... I just hope that if
somebody from the Linux/PPC (or FreeBSD/PPC, or whatever) community
notices this, they'll jump up and provide a patch for us ;-)
From: Anders_F_Bj?rklund
Subject: [SDL] Dynamic OpenGL lib support for Mac
Here's a patch that adds LoadLibrary and GetProcAddress
to the Carbon macintosh driver (for Mac OS 9 and Mac OS X):
http://www.algonet.se/~afb/SDL-1.2.6-macdynamicgl.patch
It just calls the corresponding function from SDL_loadso.
It also fixes one Mac bug in SDL_loadso.c, that made it fail
always when loading a library, and fixes the screen update
after receiving an update event - which caused the OpenGL
context to be overwritten by a blank window by UpdateRect...
From: Max Horn
Subject: Re: Again Audio CD patch
Am 04.01.2004 um 22:38 schrieb Sam Lantinga:
>
> Okay, I fixed the buffering problems by simply using a 4 second buffer
> instead of a 1 second buffer. However, using your code I can't play an
> entire CD - the playback stops after the first song.
>
Found the problem: FSReadFork returns eofErr when the file is finished.
However, we check its return value for errors, and if anything but
noErr occurs, the reader thread aborts its current iteration. That is
bad, because it aborts before it can ever set the flag which tells that
the file is over (also, any remaining data which FSRead did return is
lost - so you'd not hear about to 4 seconds from the end of the file.
Furthermore, the computed data size was 8 bytes to high (I forgot to
account for the fact that the size of an (A)IFF chunk always contains
the chunk header & size fields, too). This is enough to make it work.
However, the end condition is rather fragile, so I tuned some other
things to be pessimistic (check for <= 0 instead of == 0, when eofErr
is encountered enforce mReadFilePosition == mFileLength). You never
know...
The attached patch fixes the issue for me.
From: Pete Shinners
Subject: [SDL] directfb cursor patch
here is a quick and dirty patch that allows directfb to change the
visibility of the cursor. this display backend still has no way to change
the actual cursor shape, but being able to set the cursor visibility is a
simple change that should offer some benefits.
the code was largely based on the cursor code for other sdl backends. and
this documentation.
http://directfb.org/documentation/DirectFB_Reference/IDirectFBDisplayLayer_SetCursorOpacity.html
From: "Andy Pfiffer"
Subject: [SDL] patch: PocketPC 2000 diffs for release_1_2_6
I was making a backup of my current workspace of SDL for PocketPC,
and thought I would send out my diffs.
Apologies in advance if the patch has become whitespace mangled.
These diffs are relative to CVS tag release_1_2_6, and contain (I think)
all changes that have been mentioned on the mailing list in the last
few months.
I only have PocketPC 2000 & an ARM-based device (iPaq 3635), and I
only use Embedded Visual Tools 3.0, so I can't say for sure if it
breaks other PocketPC 200[023] builds.
nanosleep because nanosleep may not be portable to all systems
using SDL with the ALSA backend. This may be a moot point with
the switch to blocking writes anyway...
Date: Sat, 27 Dec 2003 21:47:36 +0100
From: Michel Daenzer
To: Debian Bug Tracking System
Subject: [SDL] Bug#225252: [PATCH] ALSA fixes
Package: libsdl1.2debian-all
Version: 1.2.6-2
Severity: normal
Tags: patch
For SDL 1.2.6, the ALSA backend was changed to call snd_pcm_open() with
SND_PCM_NONBLOCK. That's a good idea per se, however, it causes high CPU
usage, interrupted sound and stuttering in some games here. Taking a nanosleep
whenever snd_pcm_writei() returns -EAGAIN fixes this, but I think it's more
efficient to use blocking mode for the actual sound playback. Feedback from the
SDL and ALSA lists appreciated.
The patch also fixes the default ALSA device to be used.
From: Max Horn
Subject: SDL: video/quartz cleanup
while doing some experimental changes in the quartz code, I was annoyed
by having to recompile that one big .o file over and over again. So I
decided to finally realize one TODO: properly splitting the code over
multiple files :-).
With two exceptions, I didn't make code changes, only rearranged files
and added new headers. Since there are several new files, making a
patch didn't work out so well, so I decided to just send you all the
new & modified files.
The one source change I made is related to showing/hiding the mouse. I
renamed cursor_visible to cursor_should_be_visible and cursor_hidden to
cursor_visible; I think that makes reading the code easier.
Then I added two new functions: QZ_ShowMouse and QZ_HideMouse. They
help manage cursor_visible (the former 'cursor_hidden'). Finally I
replaced the Carbon ShowCursor/HiderCuror calls by [NSCursor hide] and
[NSCursor unhide]. The API docs are not conclusive, but it might be
that with those the "cursor_visible" (former 'cursor_hidden') hack may
not be necessary anymore; however so far I didn't test this hypothesis,
so I left that in.
The other change was to remove in_foreground and use [NSApp isActive]
instead: Manually keeping track of whether we are in the foreground is
error prone. This should work better in some corner cases.
To: Ryan C. Gordon
Subject: SDL: DirectFB videodriver segfault
Date: Sun, 21 Dec 2003 20:40:14 +0100
Hi Ryan,
attached is a patch to fix a segfault in DirectFB_VideoQuit.
When SetVideoMode never gets called (f.e. wrong command line arguments
for "testsprite"), this->screen->hwdata is NULL.
Cheers,
Thomas
From: "Andrew Bachmann"
Subject: recent build changes
To ensure that I submitted all the changes to you I attempt to rebuild SDL from a clean cvs checkout. Unfortunately it failed with this error:
/boot/develop/libsdl/SDL12/src/audio/baudio/SDL_beaudio.cc:35: SDL_BeApp.h: No such file or directory
From: Stephane Marchesin
Subject: Re: [SDL] compile problem with sdl <-> alsa
The attathced patch does the trick (although there will have to be code
changes when the alsa 0.9 api get deprecated).
On a side-note, the good news is that using alsa 1.0.0rc1 solves the
previous issues I had with my ac97 via sound chipset.
