forums.marpirc.net

http://marp.retrogames.com/inp/vai_giga ... mame62.zip

http://marp.retrogames.com/index.cgi?mo ... 00&tourn=0

Doesn't playback - loses sync on stage 2. NVRAM, maybe?

8898370151540 with default nvram.
300110610890 with no nvram.
win98.

Some alpha mame notes, since this recording is an alpha mame one:

http://mahlemiut.marpirc.net/alphamame-062-2.zip

I thought alphamame was supposed to ignore the nvram file? Most frames were at 100% thanks vai! But after the recording is over, the frame% print out goes haywire. Proly could stop the print out once the playback is over, or maybe print out on the F11 line the "summary" of average frame rates recorded after playback is over.

Not entirely sure on how to detect that playback has ended, but I would hazard a guess towards using an OSD feof function.

And no, I haven't disabled NVRAM usage, just in case games come along that won't startup without NVRAM.

oh yeah, forgot about cgangpzl and others still maybe needing nvram. Would it be better to err on the side of no nvram for the less of two evils? the fread() should fail on the eof too, you'd hope that was already detected, but proly not easy to get that info to the place where you print out the F11 text.

osd_fread() and osd_fwrite() work with the record and playback pointers, so I figure that it's the same for the EOF function, which I am about to search for. Shouldn't be a problem as the record and playback pointers are already there to detect whether or not you're recording or playing back.

...and it now works. no more nutty recorded speed display when playback ends.

Now, I'll just add Kale's fix and I look at another release soon.