On Wed Apr 15 14:35:27 2026 +0000, Rémi Bernon wrote:
Fwiw I'm not against it, and I would like to move forward as well. Then now that I have been forced to spend more time looking over GLES, I can see that there's various things that would need to be done differently. I had opened https://gitlab.winehq.org/wine/wine/-/merge_requests/10582 for that purpose, it's been sitting around for a week now, and I wanted to get it in first, and with additional changes before I can start picking changes from this one. I find the claims that you are somehow forced to work on GLES contexts unfounded. While I mostly mentioned GLES in !9982, there were other cases as well, and it is clear to me that you did not fully understand the impact of the proposed changes. Your initial version would have hidden a large number of extensions. Most of that was already addressed in !9987, although not without further issues that were resolved during the review process, with the usual amount of your pushback. Regardless, you kept doubling down on the few remaining extensions.
Throughout this, I have only been suggesting that we keep the status quo and make sure we do not hide extensions without a good reason. Nothing more, nothing less. This is not controversial, just standard practice in Wine development. All it took was just a few lines of code that I wrote anyway. The continued obsession on filtering harmless extensions or requiring full GLES context support is not me forcing anything on you. It is also a bit ironic to complain about an MR being held for a week in an MR that you have been holding for a month. !10472 could be implemented on top of this MR. That would be both simpler and would avoid a regression from !10656 by using a cleaner design instead of the current ugly workaround. More generally, I had this drafted since last year and likely would have upstreamed it much earlier if not for all the drama. So what exactly are the things that "need to be done differently" for your GLES work? I would expect it to be straightforward to extend opengl_extension with whatever is needed. Drivers should already provide a different set of extensions when a GLES context is used. Handling of core GLES functions may need some changes, but that is true regardless of this MR. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10349#note_136291