-
Notifications
You must be signed in to change notification settings - Fork 0
Closed
Description
Problem
Zero automated tests for the core unleashed logic. The only tests in the repo are for sentinel.py. Cannot verify:
- Pattern detection works (with ANSI codes, different variations)
- Arrow keys pass through correctly (all keys in VK_MAP)
- Surrogate pair normalization handles emoji correctly
- Console mode is set/restored properly
- ANSI stripping preserves word boundaries
- Mirror filtering catches spinner noise
Current state
tests/test_sentinel.py— 11KB, comprehensivetests/has empty directories: unit/, integration/, e2e/, security/, etc.- pytest is in dev dependencies but not used for core code
Recommendation
Priority test cases:
strip_ansi()with real TUI output samples- Pattern matching with ANSI codes interspersed
- Surrogate pair normalization (emoji edge cases)
- Mirror filtering (spinner dedup, blank line removal)
- Cooldown logic timing
Related
- test(unit): Add tests for unleashed-danger.py #1 (unit tests for danger variant)
- test(integration): Complete unleashed.py pending scenarios #2 (integration tests for unleashed.py)
Source
Deep audit (2026-02-12).
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels