Warn if search space for inferred strategy is small (no variation)

[jobh]

Fixes #3653.

[Zac-HD]

Thanks for another great PR @jobh! Quite a few comments below sorry, but I think we'll get through them pretty quickly 🚀

[Zac-HD]

I think SmallSearchSpaceWarning should only be emitted if signature(typ) has parameters other than self (without caring about the name).

[jobh]

I think SmallSearchSpaceWarning should only be emitted if signature(typ) has parameters other than self (without caring about the name).

Agreed. I did actually consider this, but skipped it since I didn't know off-hand how to write a simple-yet-robust check for this. I'm sure I can figure it out, but maybe you already know?

[Zac-HD]

We have an internal get_signature() helper, then check whether the .parameters attribute is empty.

[jobh]

Thanks, done. A few lines above it is indicated that get_signature.parameters might also have an item for return, but I couldn't see that behaviour in a quick test. So I'm ignoring that possibility, just checking for empty .parameters.

hypothesis/hypothesis-python/src/hypothesis/strategies/_internal/core.py

Line 1255 in 4a22076

and k != "return"

[jobh]

Thanks for the review @Zac-HD! I think all issues you raised have been addressed, with only the question above potentially remaining.

[Zac-HD]

Yeah, we can ignore that - it's only possible if you've manually constructed and assigned an illegal signature, because using return as an argument name is of course a syntax error. We'll be able to drop that check at the 3.10 EOL in 2026, thanks to python/cpython#92065.

[Zac-HD]

Pushed some final small tweaks, and I think we're ready to merge 🚀

[Zac-HD]

@jobh you should find an invite in your email to join the Hypothesis team. There's no obligation to accept, nor to do anything having accepted, but rather an offer to continue hanging around and contributing to the project in whatever way you wish to - and the right to click "merge" on PRs when you think they meet the standards laid out in our contributor guides.

Regardless, thanks again for the amazing stuff you've done over the last week or two!

[jobh]

@jobh you should find an invite in your email to join the Hypothesis team. There's no obligation to accept, nor to do anything having accepted, but rather an offer to continue hanging around and contributing to the project in whatever way you wish to - and the right to click "merge" on PRs when you think they meet the standards laid out in our contributor guides.

Regardless, thanks again for the amazing stuff you've done over the last week or two!

Thanks for the kind words @Zac-HD , and thanks for the invite. I will accept, but I don't know yet how active I will be — I haven't actually started using it yet, but it's been a fun experience so far.