Plans for the future? (OpenGL being deprecated by Apple)

[knuxyl]

Apple has halted development and progress on OpenGL drivers for their systems, providing it only as a legacy feature. Considering they dropped 32-bit support completely, it would be unreasonable to assume OpenGL won't eventually be dropped. It's not a guarantee, but Apple's intentions are to create their own environment, they don't even natively support Vulkan.

The only way in the future to access official support of these APIs would be the use of an AMD gpu, and we have yet to see an M2 Mac Pro, so we don't even know if they will be supported in the future, or if Mac is planning on completely dropping AMD from its lineup (all indications point to they probably will with next mac pro release).

There are 3rd party implementations of Vulkan on Mac by translating to Metal, so I believe Vulkan could be a route to go.

I would prefer though that love be rewritten in Rust and use WGPU as the graphics backend. (leaving of course the lua side the same as it is). This would future proof the engine without breaking compatibility with mobile or web. This would give us access to all gfx APIs. This of course would require a significant amount of work, but I believe the end result would be worth it.

So what are yalls thoughts on this? Should love just ride OpenGL until EOL on Mac, and then become obsolete, or should love start moving towards a new backend, and even possibly a new language?

[slime]

If you want to use Rust, you don't need to stick with love - there are plenty of other frameworks and engines and libraries out there, some of which are inside Rust's ecosystem.

As for graphics backends on Apple platforms, love 12 already has a working Metal backend alongside the existing OpenGL backend. https://github.com/love2d/love/issues/1567 - there are no future hypotheticals about starting, it's already done.

We're also working on a Vulkan backend (which may be able to use MoltenVK to run on Apple platforms) which is largely complete now.

This of course would require a significant amount of work, but I believe the end result would be worth it.

As far as I know there would be little practical benefit to rewriting all of love's C++ code in Rust, despite being a massive undertaking. What are you imagining will change if that's done?

[ddabrahim]

As for graphics backends on Apple platforms, love 12 already has a working Metal backend

Wow, I was not expecting that. I knew about Vulkan is in progress but had no idea about Metal

Is that mean M1 Macs also getting native support?
So when Love 12 is going to be officially released?

Thanks a lot

[knuxyl]

If you want to use Rust, you don't need to stick with love - there are plenty of other frameworks and engines and libraries out there, some of which are inside Rust's ecosystem.

As for graphics backends on Apple platforms, love 12 already has a working Metal backend alongside the existing OpenGL backend. https://github.com/love2d/love/issues/1567 - there are no future hypotheticals about starting, it's already done.

We're also working on a Vulkan backend (which may be able to use MoltenVK to run on Apple platforms) which is largely complete now.

This of course would require a significant amount of work, but I believe the end result would be worth it.

As far as I know there would be little practical benefit to rewriting all of love's C++ code in Rust, despite being a massive undertaking. What are you imagining will change if that's done?

No I was saying have the engine be in Rust, but the developer still code in lua.

If love already supported Metal and vulkan is in development, then everything I said is meaningless, I had no idea that was already done. Thanks!

[slime]

Is that mean M1 Macs also getting native support?

M1 macs have native support already in love 11.4.

So when Love 12 is going to be officially released?

We don't have an official release schedule, but you can see the remaining nice-to-have things that aren't done yet here (although we don't guarantee we'll complete all of them before releasing 12.0): https://github.com/love2d/love/milestone/1
Out of those, support for RTL languages is probably the biggest.

No I was saying have the engine be in Rust, but the developer still code in lua.

I understand, I just don't know why that would be useful.