cglm incorrectly assumes that SSE2 is also available when SSE1 is

[aryalaadi]

this causes builds for i686 to fail

related issue

[recp]

Hi @aryalaadi,

You are right we must check these separately. Actually most of existing operations are f32 which is supported by SSE. There are some exceptions like _mm_castsi128_ps and maybe more... We can check where SSE2 is not supported and fallback some intrinsics to SSE if possible.

Added to TODOs, but a PR would indeed bring the fix more quickly

[recp]

@aryalaadi I tried to fix the issue at #412 any help, feedbacks would be awesome to make the work bug-less

[recp]

The issue is fixed at v0.9.4: https://github.com/recp/cglm/releases

Thanks for reporting this.