Parse statements individually to prevent parser hangs

[nlothian]

Summary

• parse each clause or directive separately to avoid Earley parser state blow-ups
• allow direct parsing of clause and directive start rules for faster single-statement handling
• preserve PlDoc association by adjusting meta positions per statement

Testing

• uv run pytest tests/test_parser.py::TestBasicParsing::test_parse_fact_with_atom
• uv run python - <<'PY'
  import signal, sys, time
  from vibeprolog.parser import PrologParser

parser = PrologParser()
with open('library/lists.pl') as f:
content = f.read()

signal.signal(signal.SIGALRM, lambda s,f: sys.exit(1))
signal.alarm(30)
start = time.time()
try:
res = parser.parse(content)
print('parsed', len(res), 'in', time.time()-start)
finally:
signal.alarm(0)
PY

---

Codex Task

[kilo-code-bot]

✅ No Issues Found

1 files reviewed | Confidence: 95% | Recommendation: Merge

Review Details

Files: vibeprolog/parser.py (1 change)

Checked: Security, bugs, performance, error handling, code clarity

Summary: The change implements individual statement parsing to prevent Earley parser state explosion on large files. Code correctly handles position adjustments for PlDoc comments and maintains backward compatibility. No issues found.

[gemini-code-assist]

Code Review

This pull request introduces a significant performance and stability improvement by parsing Prolog statements individually, which is an effective strategy to prevent parser hangs on large files. The implementation is well-thought-out, especially in how it preserves PlDoc comment associations by adjusting metadata positions. I have one suggestion to further improve the robustness of the new parsing loop. Overall, this is a solid enhancement.

[kilo-code-bot]

✅ No New Issues

Changes since last review look good. Previous issue with statement relocation fallback has been addressed by raising RuntimeError instead.

[kilo-code-bot]

✅ No New Issues

Changes since last review look good. The RuntimeError for statement relocation failure is appropriate and addresses the previous fallback issue.

[kilo-code-bot]

✅ No New Issues

Changes since last review look good. The new commit adds Edinburgh radix token handling and resets module predicate tracking, both implemented correctly without introducing issues.

Recommendation: Merge