Fix #3: gov why shows historical failures incorrectly

[cmtonkinson]

Fixes #3

Problem

The gov why command was showing tasks with historical failures even after they had successfully completed.

Reproduction Steps

1. Start a project with tasks executing
2. Run gov stop -w to kill workers (increments Attempts.Failed)
3. Run gov start to restart
4. Tasks complete successfully (state changes to merged, etc.)
5. Run gov why - incorrectly shows the completed tasks

Root Cause

The whyTaskKind() function in main.go was checking if task.Attempts.Failed > 0 to determine if a task should appear in output. This matches any task that has ever failed, not just tasks that are currently failing.

When workers are killed:

• task.Attempts.Failed is incremented
• Task transitions to blocked state
• When task resumes and succeeds, it transitions to success states
• But Attempts.Failed is never reset - it remains as historical record

The old logic showed any task with historical failures, even if it had since recovered.

Solution

Removed the Attempts.Failed > 0 check from whyTaskKind(). Now the function only shows:

1. Tasks currently in TaskStateBlocked (actively blocked/failing)
2. Planning tasks that caused supervisor failure

This preserves the Attempts.Failed counter for observability while fixing the incorrect gov why output.

Changes

main.go:

• Removed if task.Attempts.Failed > 0 { return "failed" } check
• Tasks with historical failures but current success states no longer appear

main_test.go:

• Updated test fixtures to use state: "blocked" instead of state: "triaged" with failed: 1
• Tests now correctly verify that only currently-blocked tasks appear in output

Testing

All tests pass:

go test ./...

Manual verification shows completed tasks with historical failures no longer appear in gov why output.