Home Page Forums MAME MAME SVN builds - new driver flood

I don't know if it's a great idea to enable drivers we don't have a CPU core for - it's just gonna crash and make regression testing impossible.

I tend to agree... or at least do something so it doesn't crash.

BTW, I'd love to work on adding H8S to the H8 core right now but SVN is busted.


...which Google says means the server is out of disk space.

This is a limited time offer before I have to go back to ignoring MESS, so hopefully it gets fixed soon ;-)

You can PM or send to me if you like..

well, cybiko v1/v2 run for at least 10 seconds without problems (but of course they're not executing any code, so they only show a blank screen); cybikoxt acts like fmtowns was acting a few days ago: it starts, it runs a couple of seconds, and finally it gets an exception due to unimplemented opcode

as long as we ignore its output error, regression testing should not be harmed by its presence

Well, like I said, I'm happy to fix that if someone gets SVN working again =) I just did a lot of H8 work for System 23 in MAME (you'll see it in the next u release).

just in case svn does not come back soon, you can use the latest revision you have: the driver was always compiled and only the messdriv.c entries were commented out. uncomment them and you can try those systems with any updated CPU core you may have

(notice that the only changes I have done to enable cybiko are the following
http://git.redump.net/cgit.cgi/mess/commit/?id=eda30fc4edf9af9e90de5eea9452dd9c911aabe9
a couple of defines for CPU_H8Sxxx = CPU_H83002, and a couple of #if 0 in address maps... I suggest you to keep only the tag fix flash1->flash2 of all those changes)

SVN back. Apparently server did not have disk full but svn server process was crashed/hung.

LN

Cool, thanks LN. It's indeed working

It's not actually missing opcodes, it's that it puts the stack on internal RAM that's in a different place from the 3002 so an RTS goes berzerk very soon