On Tue, 27 Aug 2024 at 09:51, Rémi Denis-Courmont <r...@remlab.net> wrote: > > > > Le 27 août 2024 10:00:08 GMT+03:00, Marton Balint <c...@passwd.hu> a écrit : > > >IMHO the patch itself is harmless (simple, easy to follow), and if it helps > >somebody (no matter how unupported its use-case is), then why the hell not. > >But I can understand your point of view, that we should not bother with it, > >if it is unsupported in the first place, so I will just drop this, did not > >feel strongly about it anyway. > > Can't we just fail configure or compilation hard if __FINITE_MATH_ONLY__ is > true? > > If MPV or whichever downstream wants to get this flag working, it's on them > to do the hard work (and it's a lot harder than this patch).
Whoa, don't call out mpv like that. There are no plans to do any of that. -ffast-math should be considered harmful unless proven otherwise for given codebase. mpv itself depends on infinites. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
