Skip to content

Bug: actionable error (no crash) when scan/build root looks like a whole filesystem #258

Description

@vibgrate-team

Context

Walking an accidental root (/, a full OS image tree, or an enormous unpack directory) can hang, OOM, or fail opaquely. Local-first tools should fail closed with an actionable error (what happened + what to pass instead) rather than crashing or silently thrashing the machine.

Related: ignore/exclude docs (#162). This ticket is specifically about hard safety rails and error quality at the root boundary.

What to do

  1. Detect clearly unsafe / runaway roots (e.g. filesystem root, obvious OS image layouts, or walk budgets exceeded).
  2. Exit non-zero with a deterministic, actionable message: suggest a project subdirectory, --ignore patterns, or raising an explicit override flag if one exists.
  3. Add a focused test that asserts the error path (no flake, no wall-clock dependence).

Acceptance

  • Unsafe root → clear error + non-zero exit; no crash dump as the primary UX
  • Message tells the user what to do next (narrow the path / ignore / override)
  • Deterministic tests; DCO; pnpm test green

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions