At one point yes, but preferrably only when it's "safe" to remove the GL backends so it will be a while out (e.g. after WebGPU becomes ubiquitous in browsers).
Or as an emergency measure when software rot of GL drivers and tooling on Linux and Android becomes critical.
Don't expect an automatic performance increase through a Vulkan backend though, at least not when compared to D3D11 or a good GL driver like NVIDIA's.
Also, when we're at a point where GL can be removed, it's probably time to start thinking about a proper sokol-gfx successor, since a lot of design compromises of the sokol-gfx API are caused by GL.
I'm also hoping to push out Vulkan support a bit more in order to wait for some of the recent desktop extensions which make Vulkan more bearable to trickle into Android (like VK_KHR_dynamic_rendering or VK_EXT_shader_object), but that's probably a bit optimistic.
At one point yes, but preferrably only when it's "safe" to remove the GL backends so it will be a while out (e.g. after WebGPU becomes ubiquitous in browsers).
Or as an emergency measure when software rot of GL drivers and tooling on Linux and Android becomes critical.
Don't expect an automatic performance increase through a Vulkan backend though, at least not when compared to D3D11 or a good GL driver like NVIDIA's.
Also, when we're at a point where GL can be removed, it's probably time to start thinking about a proper sokol-gfx successor, since a lot of design compromises of the sokol-gfx API are caused by GL.
I'm also hoping to push out Vulkan support a bit more in order to wait for some of the recent desktop extensions which make Vulkan more bearable to trickle into Android (like VK_KHR_dynamic_rendering or VK_EXT_shader_object), but that's probably a bit optimistic.