Proposal: rename _PyInterpreterFrame struct as _Py_framedata

I think the new PR gets the code to a pretty readable state:

  • the _Py_frame name helps make it clear that the new struct is an implementation detail of the existing Python frame concept, rather than introducing a new concept
  • _PyFrame functions take either PyFrameObject* or _Py_frame*, so the name still aligns with the type even when the function has shifted to working with the new struct
  • the removal of the f_ prefix from the _Py_frame fields that had it should be enough to solve the “quick orientation when resolving a merge conflict” problem that had me looking into this issue in the first place (my ultimate aim with this PR is to make it easier to sync the PEP 558 working branch with all the Python 3.11 performance improvements)

And for the local variable names, we can now say we tried renaming them, and decided the level of code churn was too high for the potential ambiguity reduction the change could have provided.

4 Likes