-
Notifications
You must be signed in to change notification settings - Fork 5.6k
Apple software renderer retry #2571
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
This window hint allows choosing between hardware and software renderers (where available) for the created OpenGL or OpenGL ES context. Fixes glfw#589.
…they should work headless
…ftware-renderer-retry
If it's still useful to know, I can confirm that this works for me on macOS Apple Silicon. It is painfully slow as the software renderer appears to use only a single core. Sadly it also seems to have a mesh render bug with a few of my examples. |
Correctly uses accelerated graphics for me on my M2 Macbook Air 👍 |
I can confirm the exact same build as Werwolv runs in a virtual machine with the "Apple Software Renderer" renderer (while a regular glfw release would not work) |
Thanks for PR @HinTak I'm using it for running my tiny GLFW project inside of VM with macOS Ventura 13.6.9 x86_64 and it works quite well there. I've recently started messing with setting |
I explained at the beginning that this is mainly about Apple github CI #2570 . I can't remember why I have WGL or EGL code in there. As for headless Linux, I have already got an alternative: https://github.com/kyamagu/skia-python/blob/17a97a0e547c42acac6d9d6e0ec91960685d89a9/.github/workflows/ci.yml#L215 . There is a EGL issue (and even a stale EGL branch) here, I believe. Linux / Mesa related comments should go there instead. |
There is a (stale) EGL branch https://github.com/glfw/glfw/tree/egldevice . |
This is quite confusing... Not gonna lie 👀 Sorry, I simply looked at the diffs and since they show that PR touches way more code and adds more hints I though it's not only about what mentioned in the description (and not only about Apple platform itself). I believe I know what is the source our of confusion, see commit history: Sorry once again for not noticing this in the first place. I'm not near my workstation at the moment (writing this from phone). I'll try to move my comments to proper PRs per your request, as soon as I can. In the meantime could you take a look at the commits? It really seems like the old commit was merged into this by mistake. |
Look, can you comment on the other issue instead? #1592 . I am going to block you if you keep making unrelated comments. |
@HinTak, Your PR includes changes for more than just Apple Software Rendering. Maybe this was unintentional and a result of merging upstream, or maybe it accidentally pushes changes you didn't intend to be pushed? Either way, it is important to review the changes before they are pulled. I suggest, if you can, simplifying the changes down to only the ones relevant to the PR (e.g., Apple Software Rendering). I hope that explains that confusion. |
@ColleagueRiley it was intentional. Explained in #1592 . It is offensive to click on "review" when not qualified to do so. Please stop now. |
@sleeptightAnsiC You're right that this PR touches much more code than that responsible for the issue it seems to be concerned with. That's a problem, and in its current state, it doesn't seem take the full set of use-cases for GLFW into consideration. In other words, it's not ready for integration into GLFW. Your feedback and comments are appropriate here. Everyone commenting here is qualified to do so. |
Not interested in flame wars. Please stop now. |
Implement automatic software renderer re-try, and run glfwinfo post-build in the headless environment.
Fixes #2080 and #2570