On Tue Mar 31 16:56:17 2026 +0000, Jacek Caban wrote:
In other words, if you pick examples that support your claim and ignore those that do not, you somehow prove something? Another way to look at it is that drivers may or may not expose those extensions as they see fit. This way, you do not need to cherry-pick "evidences". What?? I don't pick examples that support my claim, I am showing actual evidence that the vast majority of extensions are distinct between GL and GLES.
The driver feature flag that enables these extensions is the second field, and it shows that whenever driver supports the feature (or in the case of dummy_true: always), only one of the extension name is exposed, and that depends on and only on the nature of the context (GL or GLES context). Drivers *cannot* chose which one they expose, contrary to what you say. Now clearly you have decided that I'm wrong no matter what I say, and you don't even bother offering any concrete example to make your point so I don't think there's anything more for me to say on that topic and I will stop arguing. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10472#note_134545