typing: disallow TypeVar with a single constraint; add test; update conformance outputs (#1015)

[AryanBagade]

Closes #1015

Summary

• Added ≥2-constraints validation for user-declared TypeVar(...) in pyrefly/lib/alt/expr.rs, matching the existing scoped-parameter rule in pyrefly/lib/alt/solve.rs.
• Error text: “Expected at least 2 constraints in TypeVar T, got 1”.
• De-duplicates diagnostics when both constraints and bound are present.

Tests

• Added test_typevar_single_constraint_is_error in pyrefly/lib/test/generic_restrictions.rs.
• All tests green: cargo +nightly-2025-06-20 test --all, python3 test.py.

Conformance

• Ran python3 test.py; committed updates:
  • conformance/third_party/conformance.exp
  • conformance/third_party/conformance.result
  • conformance/third_party/results.json

Tooling

• Formatted with the pinned nightly toolchain.
• Clippy scoped to -p pyrefly is clean; workspace clippy warnings in pyrefly_config are pre-existing and out of scope.

Checklist

• Feature branch: feat/typevar-min-2-constraints
• PR title references issue
• “Closes Disallow specifying only 1 constraint in a Type Var #1015” at the top
• Tests added and passing
• Conformance outputs updated
• CLA signed if prompted

[yangdanny97]

Thanks for the PR and the tests!

I understand the goal of wanting to avoid emitting two diagnostics during cases when a user has 1 constraint & a bound type, but I believe this case is going to be extremely rare, and not worth the increased complexity in the new logic.

Let me know if you disagree

[yangdanny97]

Thanks, this looks cleaner. We'll try to get this merged on Tuesday.

[facebook-github-bot]

@yangdanny97 has imported this pull request. If you are a Meta employee, you can view this in D81493328.

[AryanBagade]

Hi @yangdanny97 , I think my branch has fallen behind main, which caused the conflict in conformance/third_party/results.json and led to the internal linter showing as pending/failing. Would you like me to rebase, resolve the conformance files, and re-run tests, or will you handle the conflict on your end?

[rchen152]

Review automatically exported from Phabricator review in Meta.

[AryanBagade]

Review automatically exported from Phabricator review in Meta.

Got it, thanks for the approval. Since the conflict is only in conformance outputs, I’ll leave it to be regenerated on merge.

[rchen152]

Ah, sorry for missing your comment earlier! Yeah, we can rebase and regenerate the conformance output on our end.

[facebook-github-bot]

@yangdanny97 merged this pull request in 99bab4a.

[yangdanny97]

yeah no need to worry if the merge conflict is just in the generated files - those get automatically re-generated during the merge process