testing-gaps

name	testing-gaps
description	Run coverage, inspect results, and identify missing test scenarios for a given source file. Use when analysing test coverage or finding untested branches.

Testing Gaps

Find untested branches and missing real-world test scenarios for a source file.

Workflow

1. Run text coverage for the target file
2. Identify uncovered branches from the output
3. Cross-reference with test file to understand what's tested
4. Report uncovered branches + missing real-world scenarios

Step 1: Run Coverage

Use cargo llvm-cov with --text output (not --html) for parseable results.

# Full coverage, filtered to target file
cargo llvm-cov test --text \
  --ignore-run-fail \
  --ignore-filename-regex '(_test.rs|\/test\/)' \
  2>/dev/null | sed -n '/TARGET_FILE\.rs:/,/^$/p'

Example for preferred_semver.rs:

cargo llvm-cov test --text \
  --ignore-run-fail \
  --ignore-filename-regex '(_test.rs|\/test\/)' \
  2>/dev/null | sed -n '/preferred_semver\.rs:/,/^$/p'

Step 2: Read Coverage Output

The text format shows per-line execution counts:

  42|     80|        .and_then(|range| ...)     # Hit 80 times
  74|      0|        Some(preferred)            # Never hit
 321|      0|        instance.mark_conflict(...) # Never hit

• count > 0 = covered
• count = 0 = uncovered branch
• ^N annotations on sub-expressions show partial coverage within a line

Extract only uncovered lines

... | grep -E '^\s+\d+\|\s+0\|'

Step 3: Classify Gaps

For each uncovered branch, determine:

1. What code path leads here? — Trace the if/else chain backward
2. What input would trigger it? — What package.json + config combination
3. Is it reachable? — Some branches may be defensive/unreachable
4. Is it worth testing? — Real-world scenario vs theoretical edge case

Categories

Category	Action
Real-world scenario never tested	Write a test
Edge case in existing logic	Write a test
Defensive branch (unreachable)	Note but skip
Dead code	Consider removing

Step 4: Identify Missing Scenarios

Beyond line coverage, look for missing combinations:

• Feature A tested, Feature B tested, A+B never tested together
• Only tested with one package, never with multiple
• Only tested with highestSemver, never lowestSemver
• Only tested with one range type (^), never others (>=, <=)
• Only tested without semver groups, never with

Step 5: Report

Structure findings as:

1. Uncovered branches — Specific lines, what triggers them, why untested
2. Missing real-world scenarios — Combinations and interactions not covered

Running Tests After Adding Coverage

just test                              # All tests
cargo test test_name -- --nocapture   # Specific test
just coverage                         # Regenerate coverage report