3 Nov
2023
3 Nov
'23
7:03 a.m.
On Fri Nov 3 05:03:57 2023 +0000, Conor McCarthy wrote:
The issue with `double` opcode variants gets uglier for comparison instructions, so normalisation should go upstream before that. I'll make a draft MR. MR 442 deals with double instructions. This binop MR doesn't emit double instructions but it will have trivial conflicts in a few places. MR 442 should probably go upstream before 436 (`CAST`) because it will simplify that one.
-- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/409#note_50792