Feature request: bytes.decode

[ivirshup]

Feature request

I would like to do b"abc".decode("utf-8") in no python mode.

My use case is that h5py has decided to return text data as bytes. I would like to be able to quickly decode text data stored in structured arrays without having the loop be in python. I'm sure there are more common use cases for this as well.

I would hope that cpython's utf-8 decoding (https://github.com/python/cpython/blob/master/Objects/stringlib/codecs.h) could be reused here, but don't really know enough about how unicode strings work to comment more.

[esc]

@ivirshup thank you for submitting this, I have labelled it as a feature request.

[stuartarchibald]

This is technically feasible but quite tricky, requires unicode_decode_utf8 implementing https://github.com/python/cpython/blob/a0c603cb9d4dbb9909979313a88bcd1f5fde4f62/Objects/unicodeobject.c#L5103 as the caller of the noted codec libraries, which will also need translating.