Fabien Meghazi wrote:
Well, honestly I am all for dropping XMAME in favour of SDLMAME.
Now I have said officially what I used to say privately.
Personnaly I think going SDL is the way to go (if possible)
SDL is widely used on the emulation softwares. It's fast and well
documented. The two main advantages I see are 1) Less work maintaining
more than one driver for gfx output at code level 2) More
developpers with SDL knowledge could participate.
My primary concern for all emulators is vsync. Hans told me that it
was not a problem with SDL (using opengl) so I think DGA and xv are
not relevant anymore.
Now for the other part, I don't think a fork of sdlmame is a good
thing. Maybe the teams should merge if possible. A merge is time
consuming for all the decisions, make files, project structure,
versionning, etc ... but if a commitment can be made I think it will
be better for mame under unix. Well front-ends guys won't be happy of
_one_more_ change but this a bad for a good. (not sure if this
expression is ok in english ;-)
If you look at what happened with UAE, WinUae, E-UAE , you have to
admit that it's sad to see all that knowledge, know-how and skilled
developpers being splitted.
Don't you think ?
I fully agree, unfortunatly Arbee doesn't want effect support in SDLMAME
(he really doesn't!) actually I think he doesn't want any features which
would extend sdlmame beyond the official mame under windows
capabilities, I however would like to see atleast support for some
effects, even with opengl filtering pacman really looks better with
scale2x scaling.
So if we want an xmame / sdlmame with effect support we must fork, maybe
Arbee is willing to step down as an sdlmame maintainer to make way for a
new sdlmame based project under the xmame umbrella, because AFAIK
sdlmame was born out of necessity (Arbee and OG are mamedevs using a non
windows OS). Maybe we will need todo a small fork (IOW keep as close to
sdlmame as possible, but add some functionailty) maybe we merge by
adding a few wel placed ifdefs to the main sdlmame sources giving us the
nescesarry hooks into the sdlmame code and doing the rest in different
files (if you take a look at my scale2x patch you'll see that most of it
is in new files).
IOW we just have to wait and see, the first steps in "merging" xmame and
sdlmame would be:
1) The xmame maintainer (Laurent Desnogues) (and community) decide to \
go the SDL route instead of the Xlib route.
2) Laurent contacts Arbee about his intention and starts discussing
howto collaborate with sdlmame.
Regards,
Hans
_______________________________________________
Xmame mailing list
[email protected]
http://toybox.twisted.org.uk/mailman/listinfo/xmame
