Scott Wolchok
Scott Wolchok
selective build requires constexpr op names, so we can't refactor core parts of ops into template functions until/unless we come up with another way to do selective build.
> A variation of this has been implemented in [pytorch/pytorch#55685](https://github.com/pytorch/pytorch/pull/55685) As with the `TensorRef` in the proposal, `MaybeOwned` is hobbled by the Itanium ABI requirement to pass it by reference...
https://github.com/search?q=repo%3Apytorch%2Fexecutorch%20vbfdotq_f32&type=code
The answer to this question depends on your exact compilation setup; you are essentially asking "what is the path to the file that `#include ` includes?" What is the underlying...
It looks like you used `--`, so everything after that is taken literally, not as a flag, including your `-g`.
I think moving the flags before the double hyphen should do it: `fastmod -g '!base.css' -g '*.{tsx,ts,js,jsx,css,pcss}' -- '--palette-(dark)-([\w-]+)' '--theme-${1}-palette-${2}' ./src`
> it's kinda ugly and has the potential to hit the command line length limit The [`xargs`](https://en.wikipedia.org/wiki/Xargs) Unix tool should solve this problem.
See https://github.com/pytorch/executorch/issues/6691#issuecomment-2619823071 . In short, PyTorch officially deprecated Intel Macs back in January 2024. ExecuTorch is PyTorch, so I imagine we will follow suit.