BLD: only use PyObject_HEAD if PyObject isn't opaque - #20492

Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.

• Do the proposed changes actually accomplish desired goals?
• Do the proposed changes follow the Astropy coding guidelines?
• Are tests added/updated as required? If so, do they follow the Astropy testing guidelines?
• Are docs added/updated as required? If so, do they follow the Astropy documentation guidelines?
• Is rebase and/or squash necessary? If so, please provide the author with appropriate instructions. Also see instructions for rebase and squash.
• Did the CI pass? If no, are the failures related? If you need to run daily and weekly cron jobs as part of the PR, please apply the "Extra CI" label. Codestyle issues can be fixed by the bot.
• Is a change log needed? If yes, did the change log check pass? If no, add the "no-changelog-entry-needed" label. If this is a manual backport, use the "skip-changelog-checks" label unless special changelog handling is necessary.
• Is this a big PR that makes a "What's new?" entry worthwhile and if so, is (1) a "what's new" entry included in this PR and (2) the "whatsnew-needed" label applied?
• At the time of adding the milestone, if the milestone set requires a backport to release branch(es), apply the appropriate "backport-X.Y.x" label(s) before merge.

This doesn't feel like it would actually work if you tried to run it to me.

Just as an example:

The cast here is dodgy. You've allocated a Python object so that object starts with the PyObject head. So the cast to an object without the head ignores that structure. The correct line would be:

PyObject *self_obj = alloc_func(...); // as before
self = (IterParser*)PyObject_GetTypeData(self_obj, type);

(Assuming no inheritance... With inheritance type needs to be exactly IterParserType.) You'd return self_obj.

And you should be very suspicious of functions taking IterParser* since they almost certainly need to take PyObject* and then lookup the type data from that.

The size needs to be negative in the opaque case.

If the comments above aren't sufficiently obvious then I can make a PR with one file in a working state. The other files should be pretty straightforward once you've need the pattern. Or I can elaborate.

Thanks @da-woods for reviewing this. Indeed I did not attempt to validate the validity of my patch at runtime, as I only focused on getting past a build-time error, but this approach seems naive in retrospect. If you have time for it, I would indeed appreciate seeing a PR from you to fix a single file; I'd feel much more confident that we can get the rest of them right without assistance then :)