SDL 3.0 Adds Colorspace Concept, More Code Coming From Valve Around HDR

[smitty3268]

I thought a previous Phoronix article said the plan was to drop X11 support in SDL 3. I imagine that'd hurt uptake, given how long developers did things like building for 32-bit Windows for compatibility with the broadest base of users.

I believe the plan was to enable wayland by default (assuming it works well enough), not drop X entirely. Unless I missed something.

That said, I don't know that it will affect devs one way or another. The SDL api should be the same regardless, that's the point of abstracting it.

[ssokolow]

That said, I don't know that it will affect devs one way or another. The SDL api should be the same regardless, that's the point of abstracting it.

It would if they insist on making builds using SDL 2 and ignoring APIs/features added in SDL 3 for compatibility's sake.

[carewolf]

this really isn't in the purview of valve, HDR on the desktop is bottlenecked by wayland, and valve's and KDE's... lets call it a macgyver-esq solution is NOT suitable for general desktop usage.

Well MacGyvering something up is how development happens.

I believe we used to call it hacking

[Quackdoc]

Well MacGyvering something up is how development happens.

I believe we used to call it hacking

calling it a hack up job seems to have a nore negative connotation then what i mean. even if I don't agree with it. In the end its still good work, but it is clear and serious downsides. that simply cannot be addressed without a proper protocol.

[smitty3268]

It would if they insist on making builds using SDL 2 and ignoring APIs/features added in SDL 3 for compatibility's sake.

That will happen regardless of whether SDL 3 supports wayland or not. Changing the API for v3 means it will be slow to be adopted, regardless of what it provides or doesn't.

I assume someone will provide a 2to3 lib just like people did for v1 -> v2, anyway.

[ssokolow]

That will happen regardless of whether SDL 3 supports wayland or not. Changing the API for v3 means it will be slow to be adopted, regardless of what it provides or doesn't.

I assume someone will provide a 2to3 lib just like people did for v1 -> v2, anyway.

I'm not talking about supporting Wayland, I'm talking about dropping support for X11. If that happens any time soon, it's likely that you'll see game developers relying on people having such a 2to3 lib for the same reason it took so long to get 64-bit builds of many Windows games... ensuring compatibility with the broadest base of potential customers.

...and that'd lock them out from any SDL 3 features which can't be transparently retrofitted the way sdl12-compat does for PulseAudio, Pipewire, Wayland, and non-modesetting fullscreening.

[Alexmitter]

that still isn't feature complete with X11

And, X11 is also not feature complete with Win32 windowing, or OSX's display protocol, or Androids display protocol.
Some see it as a feature that any unprivileged process can read full input and screen output on X11, and Wayland will never gain this "feature" either.

[smitty3268]

I'm not talking about supporting Wayland, I'm talking about dropping support for X11. If that happens any time soon

I'm not sure how that even got brought up. I don't think anyone has ever suggested that will be happening and it would be a very dumb choice by the sdl developers.

it's likely that you'll see game developers relying on people having such a 2to3 lib for the same reason it took so long to get 64-bit builds of many Windows games... ensuring compatibility with the broadest base of potential customers.

Sure, to the degree that game devs even use sdl directly anymore. I think mostly they just rely on Proton to handle everything now.

[ssokolow]

I'm not sure how that even got brought up. I don't think anyone has ever suggested that will be happening and it would be a very dumb choice by the sdl developers.

While I've been having trouble finding the page again, I was certain I'd read a Phoronix post that said that, at least at the time, they wanted to do that.