PEP 3121: Mark as Final

[encukou]

• Final implementation has been merged (including tests and docs)
• PEP matches the final implementation
• [n/a] Any substantial changes since the accepted version approved by the SC/PEP delegate
• Pull request title in appropriate format (PEP 123: Mark Final)
• Status changed to Final (and Python-Version is correct)
• Canonical docs/spec linked with a canonical-doc directive (or canonical-pypa-spec, for packaging PEPs)

(There have been changes since 3.0, and the 3.0 docs for this were in a different place, but the feature was implemented as specified.)

Helps #2872.

---

📚 Documentation preview 📚: https://pep-previews--3471.org.readthedocs.build/

[gvanrossum]

Is this still the spec people should follow? Or has it been superseded?

[encukou]

It's still valid except

• m_reload was repurposed (since 3.1 it was documented as "Currently unused, should be NULL."; now it's m_slots).
• PyType_Copy was never added. Should that be said somewhere?

Nowadays you should follow the docs around PyInit_modulename and PyModuleDef.

[gvanrossum]

In cases where reality has moved on beyond what the PEP specifies, but not by much, I think it's allowed (and sometimes useful) to add a note to the PEP along the lines of "in later versions this feature was changed like this: blah, blah". This helps people who read the PEP (out of habit, or with expectations from other communities, or because of StackOverflow links, etc.).

[encukou]

I agree. In this case though, I think the differences are minor enough to not even need the note.
The affected stuff was mentioned in the Specification section, but the PEP doesn't actually say what it'd do. It's very unlikely that anyone will draw any bad conclusions, even if they ignore the big "This PEP is a historical document" banner.