Skip to content

analyze_task.md

Latest commit

 

History

History
64 lines (42 loc) · 2.56 KB

analyze_task.md

File metadata and controls

64 lines (42 loc) · 2.56 KB

Analyze Task

The analyze_task is responsible for analyzing user-uploaded test cases. It determines whether the test case causes a crash, identifies the crash type, and collects relevant metadata. This task is the entry point for processing user-uploaded test cases.

Key Responsibilities

  1. Build Setup:
    Sets up the appropriate build (custom or regular) for the test case. If a specific revision is provided, it ensures the build matches the revision.

  2. Testcase Setup:
    Prepares the test case and its dependencies for execution. This includes setting up the file paths and initializing required environment variables.

  3. Crash Detection:
    Executes the test case and determines if it causes a crash. If a crash occurs, it collects the crash type, address, state, and stack trace.

  4. Security Analysis:
    Determines if the crash is a security issue and, if so, estimates its severity.

  5. Reproducibility Check:
    Tests whether the crash is reproducible or a one-time occurrence.

  6. Duplicate Detection:
    Checks if the test case is a duplicate of an existing one.

  7. Task Creation:
    Creates follow-up tasks such as minimization, regression testing, progression testing, and impact analysis.

Workflow

  1. Initialization:

    • Resets memory tool options and environment variables.
    • Fetches the test case and crash information from the backend.

  2. Build Setup:

    • Sets up the appropriate build based on the crash revision or the latest available revision.

  3. Test Execution:

    • Runs the test case and captures the crash output.
    • If no crash occurs, marks the test case as unreproducible.

  4. Crash Analysis:

    • Analyzes the crash stack trace and determines its type, address, and state.
    • Checks if the crash is a security issue and estimates its severity.

  5. Metadata Update:

    • Updates the test case and crash metadata with the collected information.

  6. Task Creation:

    • Creates additional tasks for further processing, such as minimization and regression testing.

Important Notes

  • Reproducibility:
    If the crash is not reproducible, the task retries on another bot to confirm the result.

  • Security Flag:
    The task determines whether the crash is a security issue and assigns a severity level if applicable.

  • Duplicate Handling:
    If the test case is a duplicate, it is marked as such, and no further tasks are created.

  • Error Handling:
    If the build setup fails or the test case is invalid, the task logs an error and retries or closes the test case.