T-SQL Rule Implementer

Implement a new linting rule from scratch using TDD methodology.

Workflow

1. •Understand the requirement: Clarify what SQL pattern should be detected and why
2. •Check for existing rules: Search src/TsqlRefine.Rules/Rules/ to avoid duplicating existing rules
3. •Study a similar rule: Read the most similar existing rule and its tests for patterns to follow
4. •Write tests first in tests/TsqlRefine.Rules.Tests/{Category}/{RuleName}RuleTests.cs
   • •Include SQL that should trigger violations (positive cases)
   • •Include SQL that should NOT trigger violations (negative cases)
   • •Check expected diagnostic count, message, and position
5. •Implement the rule in src/TsqlRefine.Rules/Rules/{Category}/{RuleName}Rule.cs
   • •Prefer AST-based (visitor pattern) over token-based
   • •Use helpers from src/TsqlRefine.Rules/Helpers/ — never duplicate common logic
   • •See .claude/rules/rules-development.md for templates and helper reference
6. •Register the rule in src/TsqlRefine.Rules/BuiltinRuleProvider.cs
7. •Add sample SQL in samples/sql/{rule-id}.sql
8. •Run specific tests: dotnet test --filter "FullyQualifiedName~{RuleName}Tests"
9. •Run full suite: dotnet test src/TsqlRefine.sln -c Release
10. •Smoke test via CLI: echo "<test SQL>" | dotnet run --project src/TsqlRefine.Cli -c Release -- lint --stdin --output json

Naming

Rules

• •Always write tests before implementation (TDD)
• •All tests must pass before considering the task done
• •Use FrozenSet/FrozenDictionary for static lookup collections