The test coverage reports were showing misleading function coverage data where all individual files displayed 0.0% function coverage, even though functions were actually being executed during tests and the overall summary showed 90.5% function coverage.

Problem

The coverage scripts relied on lcov --list to display per-file coverage statistics, but this command has a known issue where it incorrectly shows 0.0% function coverage for individual files even when the underlying coverage data contains valid function execution information.

Before:

### Overall Coverage
- Lines: 76.9% (877/1141)  
- Functions: 90.5% (76/84)  ← Correct overall

### File Coverage
| File | Function Coverage |
|------|-------------------|
| src/jsrt.c | 0.0% |        ← Incorrect per-file
| src/console.c | 0.0% |     ← All showing 0%
| src/runtime.c | 0.0% |

This created confusion as the individual file percentages clearly didn't add up to the 90.5% overall coverage.

Root Cause

Investigation revealed that while lcov --list was showing 0.0% function coverage, the actual .info file contained correct function execution data:

• FNF:28 (28 functions found) and FNH:27 (27 functions hit) for console.c
• FNDA:420,jsrt_console_log showing actual function call counts
• genhtml correctly processed this data and showed accurate function coverage

The issue was specifically with how lcov --list displays function coverage percentages in the per-file view.

Solution

Modified both coverage scripts to parse the .info file directly instead of relying on lcov --list:

1. Direct parsing: Extract SF: (source file), FNH:/FNF: (functions hit/found) entries directly
2. Accurate calculation: Compute function coverage as (functions_hit / functions_found) * 100
3. Consistent methodology: Both badge and comment scripts now use identical parsing logic

Result

Function coverage now displays accurate, realistic values:

### Overall Coverage  
- Lines: 76.9% (877/1141)
- Functions: 90.5% (76/84)

### File Coverage
| File | Function Coverage |
|------|-------------------|
| src/jsrt.c | 100.0% |
| src/console.c | 96.4% |
| src/runtime.c | 73.3% |
| src/timer.c | 100.0% |

The per-file percentages now properly reflect the actual function execution and align with the overall coverage summary, eliminating user confusion about the coverage data accuracy.

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.