Petr Kobalicek
Petr Kobalicek
Nanovg rendering is incorrect in this case. The caps must be straight, see this Blend2D rendered sample:  After applying the `x/y` translation the input path becomes: ``` moveTo(168, 37);...
I think asmjit misdetects MMX2 in this case, I would have to check the impl (you are maybe the first testing cult on such CPU it seems)
I think this should have been fixed in asmjit by https://github.com/asmjit/asmjit/commit/4163483b36bb88f86301a54065566e35987d52de
The current nesting approach also doesn't work well with native addons, which should be installed in a flat hierarchy by default I guess.
Closing, if this is a problem please reopen.
@travisdowns If you have time to run cult on any Intel hardware I would be interested in results. I have updated cult to test more stuff, also memory ops, etc......
I think this can be closed as I have already merged some fixes, can you verify / close?
Closing, I hope the question got answered, if not, please reopen
@ApsarasX I haven't forgotten about this one. I initially started hacking this feature, but then I have decided that this cannot just be a hacky feature. My current idea is...
I'm sorry, but no progress on this front - I don't have time capacity to work on this at the moment. I plan to generate AArch64 assembler similarly to how...