Pre-Commit Quality Review

Review the staged changes in this commit for code quality issues before committing.

Objectives

• Fast: Complete review in < 30 seconds
• Focused: Only review staged Ruby files
• Actionable: Clear pass/fail with specific fixes
• Non-interactive: Works in headless mode
• Expert-driven: Apply Ruby best practices from industry experts

Expert Panel

Apply principles from these Ruby/software design experts:

1. Sandi Metz - POODR principles, small methods/classes, SRP, DRY
2. Jeremy Evans - Sequel best practices, transaction safety, performance
3. Kent Beck - Simple design, revealing names, Command-Query Separation
4. Avdi Grimm - Confident Ruby, null objects, tell-don't-ask, meaningful returns
5. Gary Bernhardt - Functional core/imperative shell, fast tests, immutability

Review Process

1. Get staged files:

   git diff --cached --name-only --diff-filter=ACM | grep '\.rb$'
   

2. For each staged Ruby file, check:

   • Get the full diff: git diff --cached <file>
   • Review only the added/modified lines (lines starting with +)

3. Look for critical issues (❌ BLOCK):

   Sandi Metz violations:

   • ❌ Missing frozen_string_literal: true in new files
   • ❌ Methods > 15 lines (SRP violation, too complex)
   • ❌ Classes > 200 lines (god object)
   • ❌ Obvious code duplication (DRY violation)

   Jeremy Evans (Sequel) violations:

   • ❌ Raw SQL strings instead of Sequel dataset methods
   • ❌ Database writes without transaction wrapper
   • ❌ N+1 queries (multiple queries in loops)

   Kent Beck violations:

   • ❌ Overly complex solutions (nested conditionals > 3 levels)
   • ❌ Command-Query Separation violation (methods that both mutate and return calculated values)

   Avdi Grimm violations:

   • ❌ Defensive nil checks everywhere (should use null objects)
   • ❌ Implicit nil returns in public methods
   • ❌ Bare raise or rescue without exception class

   Gary Bernhardt violations:

   • ❌ I/O operations mixed with business logic (should separate)
   • ❌ Mutable state in value objects
   • ❌ Database/file I/O in unit tests (makes tests slow)

   Testing:

   • ❌ New public methods without corresponding tests

4. Look for medium issues (⚠️ WARNING):

   Sandi Metz:

   • ⚠️ Methods 10-15 lines (consider extracting)
   • ⚠️ Classes 100-200 lines (approaching god object)
   • ⚠️ Method parameters > 3 (use parameter object)

   Jeremy Evans:

   • ⚠️ Missing indexes on foreign keys
   • ⚠️ Connection not properly managed

   Kent Beck:

   • ⚠️ Poor naming (unclear variable/method names, abbreviations)
   • ⚠️ Methods doing multiple things (and/or in method name)

   Avdi Grimm:

   • ⚠️ Law of Demeter violations (chained method calls like foo.bar.baz.qux)
   • ⚠️ Asking instead of telling (lots of getters instead of sending commands)

   Gary Bernhardt:

   • ⚠️ Business logic scattered in imperative shell
   • ⚠️ Missing value objects (primitives passed around)

   General:

   • ⚠️ Missing documentation for public APIs
   • ⚠️ Tight coupling between modules

Output Format

Provide a clear summary with expert attributions:

# Pre-Commit Quality Review

## Files Reviewed
- file1.rb (23 lines changed)
- file2.rb (8 lines changed)

## Status: ❌ BLOCK / ✅ PASS / ⚠️ WARNING

### Critical Issues ❌ (must fix before commit)

**Sandi Metz violations:**
- file1.rb:1 - Missing frozen_string_literal: true
- file2.rb:15-38 - Method too long (24 lines, max 15)

**Jeremy Evans violations:**
- file1.rb:42 - Raw SQL string instead of Sequel dataset method
- file1.rb:50 - Database write without transaction wrapper

**Avdi Grimm violations:**
- file2.rb:10 - Implicit nil return in public method

### Medium Issues ⚠️ (consider fixing)

**Sandi Metz:**
- file3.rb:20-32 - Method is 13 lines (consider extracting)

**Kent Beck:**
- file2.rb:5 - Variable name `x` is unclear (revealing intent)

**Avdi Grimm:**
- file3.rb:45 - Law of Demeter violation: user.account.settings.theme

### Recommendations

1. Add `frozen_string_literal: true` to file1.rb (Sandi Metz)
2. Convert raw SQL at file1.rb:42 to Sequel dataset (Jeremy Evans)
3. Wrap database write at file1.rb:50 in transaction (Jeremy Evans)
4. Extract file2.rb:15-38 into smaller methods (Sandi Metz)
5. Return explicit value instead of nil at file2.rb:10 (Avdi Grimm)
6. Rename `x` to describe what it contains (Kent Beck)
7. Extract file3.rb:20-32 for better readability (Sandi Metz)
8. Use tell-don't-ask at file3.rb:45 (Avdi Grimm)

## Suggested Actions

[If BLOCK] Run: git reset HEAD <files> && fix issues && git add <files>
[If WARNING] Consider fixing before commit or create a follow-up task
[If PASS] Commit looks good! 🚀

Exit Behavior

• Return clear judgment: BLOCK, WARNING, or PASS
• Be specific: Include file:line references for all issues
• Be helpful: Suggest concrete fixes
• Be fast: Don't over-analyze, focus on obvious problems

Quick Reference Checklist

For each added/modified line in the diff, scan for:

Sandi Metz:

• frozen_string_literal at top?
• Methods < 15 lines? (warn at 10+)
• Classes < 200 lines? (warn at 100+)
• No duplicate code?
• Parameters <= 3? (warn if more)

Jeremy Evans:

• Sequel datasets not raw SQL?
• DB writes in transactions?
• No N+1 queries in loops?

Kent Beck:

• Nested conditionals <= 3?
• Clear, revealing names?
• Methods do one thing?
• Command-Query Separation?

Avdi Grimm:

• Null objects instead of nil checks?
• Explicit returns (not implicit nil)?
• Specific exception classes?
• Law of Demeter (max 2 dots)?

Gary Bernhardt:

• Business logic separate from I/O?
• Value objects immutable?
• Tests avoid I/O?

Success Criteria

The review is complete when:

• All staged Ruby files have been checked
• Clear BLOCK/WARNING/PASS verdict is given
• All critical issues have file:line references with expert attribution
• Concrete fix suggestions are provided with expert rationale