[uae] Re: ToDo list
- From: Richard Drummond <evilrich@xxxxxxxxxxxxxx>
- To: uae@xxxxxxxxxxxxx
- Date: Fri, 12 Mar 2004 16:16:05 -0500
Hi Jens
On Friday 12 March 2004 09:03 am, Jens wrote:
> can you release a ToDo list of your work on UAE please?? I would like
> to see at which you program at present on UAE.
To do: all the things I haven't done yet. ;-)
Apart from general bug-fixing and cleaning up, how about these for some short
term goals (in random order):
1) Re-write the GTK+ GUI. It's a horrible mess and is missing lots of
features. I started doing this and got side-tracked by various things . . .
2) AHI emulation. I'm still trying to figure out the best way to do this. I've
looked at using the SDLmixer library as a cross-platform solution, but I
don't like it much.
3) Get SCSI emulation working on OS X. (Still trying to track down a cheap
SCSI burner for my Mac . . .)
4) Tidy up the bsdsocket emulation - it's a horrible mess, too. Being lazy,
I'd like to re-use as much of the Win32 version of possible (rather than the
current implementation). Advantages: the Win32 version is much more complete
and stable and doesn't require the stub bsdsocket.library. I've done some
experimentation along these lines, but not got far with it yet.
5) Re-write the JIT direct memory code to use to use mmap() etc. instead of
shmget(). The current implementation using shmget() limits direct memory to a
maximum of 32 MB (the largetst shm segment you can allocate with glibc). So
if you allocate more than 32MB of ZIII memory, for example, the JIT can't use
direct memory and so takes a significant performance hit. (Plus using mmap()
is more flexible, and may enable us to do some cool things further down the
line. ;-)
6) Support for hardfiles larger than 2GB. Should have done this ages ago.
7) OpenGL rendering of native screens. Will allow better performance on some
platforms (e.g. OS X) and will mean we can use some of WinUAE's gfx filters.
8) Any other little bits and pieces still to get working: snapshots, CD32
controller emulation, Catweasel support, other things I've forgotten.
Longer term goals include:
1) Port the JIT engine to PPC. This is in progress, but still a long way off.
2) PowerUp/WarpUp emulation on PPC hosts. I've done some exerimentation to
prove this is feasible. Requires item 5, above.
3) GNUstep/Cocoa GUI. I'm still getting to grips with Objective C and
GNUstep/Cocoa. Why GNUstep? It's fairly compatible with Cocoa and is easier
for me to work with. My Mac is too slow for doing real development on.
4) Ethernet emulation. The idea here is to create an ethernet tunnel to the
host using the Linux/BSD Tun/tap device or similar, and thus allow fuller
networking support. (This is how Bochs, Mac-on-Linux, etc support
networking.)
5) Any other ideas that I can think of.
Any comments?
Cheers,
Rich
Other related posts:
