fix(lint/useJsxKeyInIterable): fix a false positive when key is in the return statement, but not any variable declarations

[dyc3]

Summary

This PR fixes a false positive where the rule would flag variable declaration statements even if the return statement had a component with the key prop set.

fixes #3020

Test Plan

Added snapshot test

[codspeed-hq]

CodSpeed Performance Report

Merging #3031 will degrade performances by 6.04%

_{Comparing dyc3:05-31-fix_lint_usejsxkeyiniterable_fix_a_false_positive_when_key_is_in_the_return_statement_but_not_any_variable_declarations (9f2def5) with dyc3:05-31-fix_lint_usejsxkeyiniterable_fix_a_false_positive_when_key_is_in_the_return_statement_but_not_any_variable_declarations (b3f9a2d)}

Summary

❌ 1 regressions
✅ 91 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	dyc3:05-31-fix_lint_usejsxkeyiniterable_fix_a_false_positive_when_key_is_in_the_return_statement_but_not_any_variable_declarations	dyc3:05-31-fix_lint_usejsxkeyiniterable_fix_a_false_positive_when_key_is_in_the_return_statement_but_not_any_variable_declarations	Change
❌	dojo.js[cached]	8 ms	8.5 ms	-6.04%

[ematipico]

Are the names of your branches automatically generated? They are very long! 😆

[anonrig]

looks good after applying @ematipico's recommendation

[dyc3]

Are the names of your branches automatically generated? They are very long! 😆

Yeah. I use graphite on my end to stack branches. It helps keep my PRs smaller (in general) to make it easier to review them, without necessarily having to wait for reviews to continue working on stuff.