Update to newest "golang.org/x/mobile/gl"

[maja42]

I'm currently investigating my options for writing an OpenGL application for both desktop and web and this repository seems the be the best candidate - good work!

It is a while since it was forked from "golang.org/x/mobile/gl" and it seems that there were quite some major changes - adding WebGL support wasn't one of them.
One change I like particularly is the render-thread: Applications use a glctx-object that exposes the OpenGL functionality, but the calls to that context are only pushed into a channel and executed later by a dedicated render thread. This improves multithreading / go-routine support greatly.

I'm not able to use "golang.org/x/mobile/gl" due to the lack of WebGL support (I don't need to support mobile). And I also don't like the fact that desktop applications require the ANGLE dlls instead of using the OpenGL bindings from go-gl - but I can live with that.

Is it possible for goxjs to adapt those features or to be updated to the newest version of golang.org/x/mobile/gl?
Which changes would be required to do so?

[dmitshur]

Thanks for your interest in this package.

The upstream golang.org/x/mobile/gl has indeed made some internal changes since the fork point, the major one being the render-thread you mentioned.

I don't currently have plans to work on backporting this feature for goxjs/gl, because for my needs, the current implementation works sufficiently well. Before the render-thread can be implemented, it would need to be evaluated. If you're interested in doing that, I recommend doing it on a fork and letting me how it goes.

The only short term plans I have for goxjs/gl development right now is the addition of WebAssembly support, as can be previewed on the wasm branch.

[maja42]

Thanks for the fast answer. I'm currently working on a small engine and decided to use x/mibile/gl for now.

But once I'm confident enough in that project to add WebGl support, I wouldn't mind switching to goxjs and to add the missing features there.

x/mobile seems to have way too much overhead if "only" OpenGL is needed. It's more like an experimental framework than a library.

What is your opinion about a dedicated glContext interface instead of global package functions? And what is your attitude regarding breaking changes or releasing a v2?

[dmitshur]

Thanks for the fast answer. I'm currently working on a small engine and decided to use x/mibile/gl for now.

But once I'm confident enough in that project to add WebGl support, I wouldn't mind switching to goxjs and to add the missing features there.

Sounds like a good approach.

If you want to make big changes to goxjs/gl, I recommend you work on a fork and feel free to share updates. I'm unlikely to have time to review big PRs.

x/mobile seems to have way too much overhead if "only" OpenGL is needed. It's more like an experimental framework than a library.

x/mobile/gl was created primarily for accessing OpenGL on mobile devices (iOS and Android). Desktop support was added later, mostly to aid development.

What is your opinion about a dedicated glContext interface instead of global package functions?

It's reasonable to consider when making a new API, but it's a large change, and one needs to evaluate the benefits and costs. At this time, I don't have a strong desire to make such a change.

And what is your attitude regarding breaking changes or releasing a v2?

Breaking API changes to fix bugs in the API can be considered, but otherwise I'd like to keep the API stable. v2 is fine, but I don't have plans to work on one myself soon.