fix go-to definition / hover includes the lhs in operators #1517

[asukaminato0721]

Summary

Fixes #1517

The operator-definition path now only activates when the cursor is actually in the operator gap, instead of anywhere inside the enclosing expression, so hovering or go-to-def on the left 1 in 1 + 1 no longer resolves int.__add__.

Test Plan

update test

[Copilot]

Pull request overview

This PR fixes an LSP bug where go-to-definition and hover could incorrectly treat the left-hand operand of an operator expression as if the cursor were on the operator itself (e.g., hovering the 1 in 1 + 1 showing the operator dunder).

Changes:

• Tighten operator “cursor hit-testing” by requiring the cursor to be positioned between operands (or in the operator region) before resolving operator dunder definitions.
• Extend operator handling for chained comparisons to select the correct left/right operands for the specific operator under the cursor.
• Add regression tests for hover and definition on a binop LHS literal.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
pyrefly/lib/state/lsp.rs	Adds cursor-position gating for operator dunder/definition resolution to avoid matching operands as operators.
pyrefly/lib/test/lsp/hover.rs	Adds hover regression test ensuring LHS literal hover shows Literal[1], not an operator dunder.
pyrefly/lib/test/lsp/definition.rs	Adds go-to-definition regression test ensuring LHS literal doesn’t resolve to an operator definition.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[asukaminato0721]

can't fix this... i guess it's another issue

[Arths17]

Hey @asukaminato072, I was just looking over this. Resolving int.add when hovering an LHS literal was definitely an annoying bug for users, so thanks for that

and for the subscript range logic for c[0] is definitely a separate beast compared to binary operators. It's better to keep this PR focused on fixing the #1517 binop regression rather than expanding the scope.

[github-actions]

According to mypy_primer, this change doesn't affect type check results on a corpus of open source code. ✅