> But I'm not sure if adopting a scheme like this, which is pretty much what > AdvanceMAME does, would make releases that much faster, since I still need > to make sure various targets build with gcc 3.3.x and 2.95.x and sometimes > have to fix problems in either the core or the xmame sources.
well, I'm not really intimate with the mame source in the same way you are, but barring major shifts in the way the source is organized (and new drivers), do very many new mame versions cause breakage in compilation? I'd imagine that there's not too many drastic changes between one revision to the next (again, barring new drivers)... but if you released a patch (which could also have it's own version number independant of mame.... though that could be confusing), the users could test if something breaks, and then report back. I'm all for regression testing, but I'd imagine you'd rather get a quick email saying "sparc port's got issues. here's the compile log", than spending a few hours compiling for each platform, and saying "yep. still works". with this in mind, you would only have to release every time there's a bugfix, or someone submitted new features. you have every right to be a slacker. this just allows you to do so. :) _______________________________________________ Xmame mailing list [EMAIL PROTECTED] http://toybox.twisted.org.uk/mailman/listinfo/xmame