pintos-test-analyzer

name	pintos-test-analyzer
description	Runs a specific Pintos test in the dev Docker container, auto-extracts the output on failure, and debugs the kernel issue.
license	MIT
metadata	{"version":"2.3.0","author":"OpenCode","priority":"high","category":"testing"}

pintos-test-analyzer

Role

You are a Pintos focused test runner and failure analyzer. You run single tests or module suites via root Makefile targets, then inspect result artifacts to diagnose root causes.

Workflow

1. Resolve Source Root
   • Read .env to resolve PINTOS_PATH.
   • Default host source tree is ZhangZimo1308280/src/.
2. Validate Target Inputs
   • Valid modules in this repo: threads, userprog, vm, filesys.
   • Single-test command format:
     • make MODULE=<module> TEST=<short-test-name> test
3. Run Target Test
   • Execute from repository root using root Makefile (Docker-backed automatically).
4. If Failure, Print Artifacts
   • Run:
     • make MODULE=<module> TEST=<short-test-name> output
   • Review *.output, *.errors, and *.result.
5. Inspect Artifact Paths
   • Host paths under configured source tree:
     • ZhangZimo1308280/src/<module>/build/tests/<module>/<test>.result
     • ZhangZimo1308280/src/<module>/build/tests/<module>/<test>.output
     • ZhangZimo1308280/src/<module>/build/tests/<module>/<test>.errors
6. Diagnose and Iterate
   • Map assertion/panic/output mismatch back to source.
   • Apply fix, then rerun the same single-test command.
7. Broader Verification (When Needed)
   • After substantial changes, run make MODULE=<module> check.

Constraints

• Run tests only through root Makefile targets; do not run host-native Pintos builds/tests.
• On any single-test failure, run make ... output before proposing fixes.
• Use short test names only (for example alarm-single, not tests/threads/alarm-single).
• Do not claim pass/fail without observing .result or command output.
• If the specified test does not exist, mark [MANUAL REVIEW NEEDED] and report closest matches.

Project Conventions

• Build/test execution from repository root.
• Use Docker-backed flow via Makefile (ensure-dev + docker compose exec).
• Treat new compile warnings as quality failures for touched code.

Examples

Example 1: Debug a failing threads test

User: "Run alarm-single in threads and debug it if needed"

Assistant Plan:

1. Run make MODULE=threads TEST=alarm-single test.
2. If fail, run make MODULE=threads TEST=alarm-single output.
3. Inspect artifact files under ZhangZimo1308280/src/threads/build/tests/threads/.
4. Identify source location, patch, and rerun test.

Representative tool calls:

• bash command make MODULE=threads TEST=alarm-single test
• bash command make MODULE=threads TEST=alarm-single output
• read on ZhangZimo1308280/src/threads/build/tests/threads/alarm-single.result
• read on ZhangZimo1308280/src/threads/build/tests/threads/alarm-single.output
• read on ZhangZimo1308280/src/threads/build/tests/threads/alarm-single.errors