libretro stella core vs stella2014 core

[alijani1]

Anyone have an idea on how we can get some of the fixes made here on stella2014 into the standard llibretro-stella core: https://github.com/libretro/stella

emuelec images do not have stella2014 cores installed by default and on many of the systems they are a pain to install manually.

any thoughts?

[alijani1]

Anyone was able to get this working with batocera or emuelec ?

[hizzlekizzle]

What are some of the things you would like to see? I was under the impression the up-to-date core was better in every way aside from speed.

[alijani1]

@hizzlekizzle, what I am specifically interested in is the fix that was done under issue 78

you can see the video on how great the fix is in lr-Stella2014 here

Problem is Batocera and emuelec dot allow us to use stela2014 core and they are limited to the base core which is https://github.com/libretro/stella and it does not contain this fix and I have had no luck finding anyone to fix it in that base project.

[thrust26]

The Stella team would really like to see that the mainline Stella would be better supported by the Libretro team. We are quite frequently getting requests regarding Libretro and the latest Stella and usually do not have the resources (or knowledge) to support them.

We would much prefer that development here switches to mainline Stella first. With optional back porting to Stella2014. Instead of vice versa as it currently is.

[hizzlekizzle]

I fully agree. Unfortunately, many of the downstream projects use the older fork for potato hardware (SBCs, embedded, etc.). Additionally, I think one of the primary hangups is just that we aren't even aware of libretro-focused issues unless they're reported here.

I wonder if there's a good workflow for getting our contributors' eyes on the upstream issues..? What about a libretro repo that holds nothing but a submodule pointing to the upstream repo, and you guys could direct users to report libretro-specific issues there? and if/when we fix them, they would necessarily have to be submitted upstream?

[thrust26]

Adding paddle support for Libretro (like requested above) and other controllers shouldn't be a major problem for anyone knowing Libretro. Of course the Stella team would provide any necessary support here.

[alijani1]

The Stella team would really like to see that the mainline Stella would be better supported by the Libretro team. We are quite frequently getting requests regarding Libretro and the latest Stella and usually do not have the resources (or knowledge) to support them.

We would much prefer that development here switches to mainline Stella first. With optional back porting to Stella2014. Instead of vice versa as it currently is.

First, am I correct that the main libreto stella core is here? https://github.com/libretro/stella

It looks like it's not been updated for over 3 years :(

@thrust26 , it looks like @hisslekizzle is willing to have contributors assist and was suggesting a possible workflow. any thoughts on his suggestion?

[thrust26]

I have no clue how the Libretro Stella projects are organized. There seem to be at least two branches. An old one, based on Stella 3.x, and one which is based on newer Stella. The Stella team concentrates on the computer version, plus RetroN77. And that's already a bit too much for us.

So if anyone wants to help us out, he (or she) is more than welcome. The best way to support mainline Stella is via Pull Requests. I have posted some brief instructions here. This also work for Libretro specific code. However we will not be able test the latter (for multiple reasons). So the responsibility is on the side of the people who provide the platform specific code.

But I really urge everybody to align at Libretro side first. IMO it makes no sense to work on multiple forks. Even less for us, with our limited capacities.

[hizzlekizzle]

AFAIK, we build and distribute from the upstream repo, so anyone wanting to work on the libretro integration can fork the upstream repo, make changes, we can discuss/build/test those changes from the fork and then you can PR them upstream.

[thrust26]

Sounds good to me.

[theflyingape]

Latest RetroArch git source now fails to render Stella screen while all other cores work perfectly fine.