Apply normalization consistently in VLenBytes

[folded]

None and 0 are treated like a 0 length string when computing lengths, and the same normalization should be applied to the value passed to PyBytes_AS_STRING. If this is not done, an assertion is hit in the python runtime (when compiled in debug mode).

TODO:

• Unit tests and/or doctests in docstrings
• Tests pass locally
• Docstrings and API docs for any new/modified user-facing classes and functions
• Changes documented in docs/release.rst
• Docs build locally
• GitHub Actions CI passes
• Test coverage to 100% (Codecov passes)

[martindurant]

Like on line 231? It seems like l = 0 in this case, there is nothing to copy, and these operations are wasted. I might be wrong...

[folded]

If you don't pass a bytes object, you will hit this assert when calling PyBytes_AS_STRING: https://github.com/python/cpython/blob/40fff90ae3d46843bb9d27c6a53ef61c861a3bb4/Include/cpython/bytesobject.h#L21

it would be equivalent to skip the PyBytes_AS_STRING and memcpy entirely if l = 0. Shall I update the PR to do that instead?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.91%. Comparing base (a8f6efb) to head (d3a8fc0).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #570   +/-   ##
=======================================
  Coverage   99.91%   99.91%           
=======================================
  Files          59       59           
  Lines        2334     2334           
=======================================
  Hits         2332     2332           
  Misses          2        2