new mame2010 folder structure

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd No I don't think cheats have ever worked on lr-mame2010. I noticed on the Android MAME that there is also a 'cheat' folder with a 'output.xml' file (probably the last selected game/cheat). Maybe libretro needs that 'cheat' folder available to output the xml file ? I don't have time to test this at the moment though.

Thanks this is all helpful to me. I've only been working with this core for a couple of months.

maxbeanz

Creating a 'cheat' folder doesn't work.

markwkidd

@maxbeanz I made some progress tracking down the problem with contra.zip and the others but no fix yet. I need to stop for now but I'll be tracking this bug here: https://github.com/libretro/mame2010-libretro/issues/101

estefan3112

@markwkidd Happy to report that MAME2010 now successfully runs on both Raspberry Pi (from source) and Ubuntu Retropie. Also highscore support works, yippie! Very eager now to check this core further out.

maxbeanz

@markwkidd I tried a few more games today and found some more that don't boot :( I'll post them here in case you can spot any running themes.

Hypersports (hyperspt.zip)
Pacman 25th Anniversary (25pacman.zip)
Pac-Mania (pacmania.zip)
Renegade (renegade.zip)
Robotron 2084 (robotron.zip)
Rolling Thunder (rthunder.zip)
Smash TV (smashtv.zip)
Terminator 2 (term2.zip)

markwkidd

@maxbeanz Thank you for your notes. I've been working to track down the specific regression that is causing this bug with the hopes of fixing it by the end of the weekend.

On a brighter note, the patch to enable RetroPad controls for Players 3 and 4 in mame2010 did get merged today. My suggested 4-player test romset is simpsons.zip if anyone wants to try it out.

Normally I would be testing 4-player support with NBA Jam Tournament Edition v2.0, but tragically that romset is affected by this same regression that maxbeanz reported. :argh: :argh:

Edit: I had a mistaken understanding of where this regression was introduced. I am pretty sure that it has nothing to do with the new folder structure and everything to do with some bulk changes to the code by Twinaphex last week that removed the last bits of C++ from mame2010. For the curious I have detailed my process as an issue and I'm waiting to hear back from him: https://github.com/libretro/mame2010-libretro/issues/101

markwkidd

This morning Twinaphex merged by fix for the mame2010 segementation error with Contra, NBA Jam and many others.

@maxbeanz I'd be grateful as always to you or anyone else has follow up reports on how those titles are working with the new patch in RetroPie after recompiling. It's working well on my desktop system at lest.

maxbeanz

@markwkidd I can confirm that all the games I listed above now boot correctly except for 'Pacman 25th Anniversary (25pacman.zip)', not sure why that doesn't work.

Also, I noticed that the '.hi' files are appearing in the 'hi' directory. Does that mean that I don't need to add a 'hiscore.dat' file ?

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd I can confirm that all the games I listed above now boot correctly except for 'Pacman 25th Anniversary (25pacman.zip)', not sure why that doesn't work.

Woohoo!

I assume from the way you wrote this that 25pacman.zip was working before the recent updates. I'll take a look at that one. Possibly a different regression than the C++ to C conversion segfault which was fixed today.

Also, I noticed that the '.hi' files are appearing in the 'hi' directory. Does that mean that I don't need to add a 'hiscore.dat' file ?

That is correct. I have recently updated mame2003 and mame2010 so that they have a built-in hiscore.dat. If you want to use a traditional external file (for example if you are trying to fix or add new hiscore memory maps) there is a new core option for both cores that reads from an external hiscore rather than the internal.

maxbeanz

@markwkidd Thanks for all your updates on this. It seems the only thing that didn't work was the cheats but I suppose we can add lots of credits for most games as a substitute.

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd Thanks for all your updates on this. It seems the only thing that didn't work was the cheats but I suppose we can add lots of credits for most games as a substitute.

I'd suggest that that enabling mame2010 cheats might be a good bounty. In addition to the mainline MAME 0.139 cheat system I think that MAME4Droid has cheat support so there would be a few sources to draw on.

I've got my personal financial and time resources committed to other libretro MAME issues at present and I'm starting to ramble on but I bet it wouldn't be hard to come up with a suitable $50-$100 bounty if someone would chip in the first little bit.

P.S. If anyone reading this thinks they might want a bounty for cheat support, please start a fresh, descriptive, new github issue like for the new mame2003 vector rendering bounty and attach the bounty to that: https://github.com/libretro/mame2003-libretro/issues/332

maxbeanz

@markwkidd Just a quick question, I won't bother opening a new thread, is the mouse supposed to work (lr-mame2010) in gun games etc. or is that only possible in 2003 ?

markwkidd

@maxbeanz there is single mouse support once you turn on the "Mouse support" core option.

What a zany option. I just submitted a PR for it to be on by default

I'm not in a position to test multi-mouse support at the moment. This core was created long before there were any libretro cores that supported multi-mouse but it's possible that things may just work since MAME definitely has that underlying capability.

If anyone tries multi-mouse with mame2010, please share your results. I'm a broken record but there may need to be a github issue if it multi-mouse isn't working at the moment :)

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd I can confirm that all the games I listed above now boot correctly except for 'Pacman 25th Anniversary (25pacman.zip)', not sure why that doesn't work.

Also, I noticed that the '.hi' files are appearing in the 'hi' directory. Does that mean that I don't need to add a 'hiscore.dat' file ?

I see that 25pacman was not added until MAME 0.139u1 and therefore it's not in my MAME 0.139 set to test.

When you have a chance could you post a github issue about this regression for mame2010 with a log?

maxbeanz

@markwkidd said in new mame2010 folder structure:

@maxbeanz said in new mame2010 folder structure:

@markwkidd I can confirm that all the games I listed above now boot correctly except for 'Pacman 25th Anniversary (25pacman.zip)', not sure why that doesn't work.

Also, I noticed that the '.hi' files are appearing in the 'hi' directory. Does that mean that I don't need to add a 'hiscore.dat' file ?

I see that 25pacman was not added until MAME 0.139u1 and therefore it's not in my MAME 0.139 set to test.

When you have a chance could you post a github issue about this regression for mame2010 with a log?

This maybe why it doesn't work ! I can't be certain that it worked in lr-mame2010, I might be confusing it with Android MAME. Weird really because MAME4Droid uses set 0.139u1 so i'm not sure if that's what the Libretro version uses.

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd said in new mame2010 folder structure:

@maxbeanz said in new mame2010 folder structure:

@markwkidd I can confirm that all the games I listed above now boot correctly except for 'Pacman 25th Anniversary (25pacman.zip)', not sure why that doesn't work.

Also, I noticed that the '.hi' files are appearing in the 'hi' directory. Does that mean that I don't need to add a 'hiscore.dat' file ?

I see that 25pacman was not added until MAME 0.139u1 and therefore it's not in my MAME 0.139 set to test.

When you have a chance could you post a github issue about this regression for mame2010 with a log?

This maybe why it doesn't work ! I can't be certain that it worked in lr-mame2010, I might be confusing it with Android MAME. Weird really because MAME4Droid uses set 0.139u1 so i'm not sure if that's what the Libretro version uses.

I'm not sure which of the 0.139* versions mame2010 is derived from either, frankly.

maxbeanz

Does anyone know if lr-mame2010 is actually 0.139u1 ? I found a changelog here and one change said that it fixed the road transparency in Thunderceptor. I checked that game and there's no transparency, so does that mean mame2010 is 0.139u1 ? If it is then the '2010 reference set' is not correct because it doesn't contain 'pacman 25th anniversary'. But that pacman game doesn't seem to boot ! Arrgg ! Does anyone know from the code what set it uses ? Is it the same as MAME4droid ?

markwkidd

Looks like mame2010 began as MAME 0.139: https://github.com/libretro/mame2010-libretro/blob/cfb1dba3444686af3b18a7902c071c5b83c756fa/src/version.c

Whereas MAME4Droid0.139u1 began as MAME 0.139u1: https://sourceforge.net/p/mame4droid/code/HEAD/tree/Reloaded/trunk/src/version.c

That doesn't meant that mame2010 didn't have the mame 0.139u1 updates applied to it later on however, just that it didn't start that way.

maxbeanz

@markwkidd Yeah, I cannot figure out whether lr-mame2010 needs the 0.139 dat or the 0.139u1 dat.

Again, off topic. I enabled the mouse in the RetroArch menu, I can move the mouse but the mouse buttons don't work, I can't even map them. I tried Mame2003 and the same thing, the mouse buttons don't work. Have I missed something that changed in Retroarch ?

EDIT: As i'm using a mouse/keyboard combo I can use the Ctrl and Alt keys as the mouse buttons. No idea if that's the way it is but still stumped as to why the 'actual' mouse buttons don't work.

markwkidd

@maxbeanz said in new mame2010 folder structure:

@markwkidd Yeah, I cannot figure out whether lr-mame2010 needs the 0.139 dat or the 0.139u1 dat.

Again, off topic. I enabled the mouse in the RetroArch menu, I can move the mouse but the mouse buttons don't work, I can't even map them. I tried Mame2003 and the same thing, the mouse buttons don't work. Have I missed something that changed in Retroarch ?

EDIT: As i'm using a mouse/keyboard combo I can use the Ctrl and Alt keys as the mouse buttons. No idea if that's the way it is but still stumped as to why the 'actual' mouse buttons don't work.

I'm not experiencing that symptom here but I'm on a different platform with a different input driver. That sounds like a bug to me :(