fix: quarantine files on EXIT_NONZERO instead of aborting chunk

[shashwatMishra96]

When a worker process exits with a nonzero code (e.g. internal parse-limit abort on a pathological file), treat it the same as CRASH/HANG: quarantine the offending file via the existing two-consecutive-strikes mechanism and continue indexing the rest of the chunk.

Previously, EXIT_NONZERO caused the supervisor to abort the entire chunk, losing all indexed data for that directory. This is impactful for large-scale repos (Android AOSP: 253GB, 60M+ files) where machine-generated files routinely trigger internal limits.

The two-consecutive-strikes intersection mechanism already guards against false quarantines — a systemic nonzero exit (e.g. bad CLI arg) produces no recurring suspect, so the intersection is empty and the supervisor gives up (same behavior as before).

Tested on Android QSSI17 (253GB, 24 chunks) + QCM6490 vendor (244 chunks). Result: 11.35M nodes indexed, 0 data loss, 2 genuinely pathological files quarantined (checkpatch.pl, a 140K-line generated .ts file).

What does this PR do?

Checklist

• Every commit is signed off (git commit -s) — required, CI rejects
  unsigned commits (DCO, see CONTRIBUTING.md)
• Tests pass locally (make -f Makefile.cbm test)
• Lint passes (make -f Makefile.cbm lint-ci)
• New behavior is covered by a test (reproduce-first for bug fixes)

[github-actions]

Thanks for opening this — it has been seen, and it is queued.

This note is automated, but it is not a brush-off: it exists so you know where your PR stands instead of having to guess from silence.

Current review status: working through a backlog. 0.9.1-rc.1 is out, so the release freeze that held reviews is over — but it left a large queue of open pull requests behind it, and we are reading through them oldest-first. The background is in discussion #1144.

What that means for this PR, concretely:

• It will not be closed for inactivity. No stale bot touches pull requests here.
• It may still sit a while before a human reads it. That is on us, not on you.
• Older PRs are read first, so a recent one is not being skipped — it is behind a queue.

Things that will genuinely speed it up whenever review does happen:

• Keep it rebased on main — the tree is moving quickly right now, and a conflicting branch cannot be reviewed as the diff you intended.
• Get CI green, or say which failures you believe are pre-existing.
• Keep the change to one claim. Bundled features and refactors get split before they get merged, which costs you a round trip.
• Every commit needs a sign-off (git commit -s) — CI enforces DCO.

If this fixes a bug, a reproduction we can run is worth more than a description of the symptom.

Thanks for contributing, and sorry in advance for the wait.

[DeusData]

The design earns its merge: EXIT_NONZERO joining CRASH/HANG in the quarantine path is right, the two-consecutive-strikes intersection already guards innocents for systemic failures, and the AOSP-scale field validation (253 GB, 11.35M nodes, exactly 2 genuine quarantines) is the strongest evidence a PR in this area has brought. One requirement stands between it and merging: this repo's reproduce-first rule needs a regression test — a deterministic nonzero-exit worker on one file proving the chunk continues and the file lands quarantined with phase "error", plus the systemic-nonzero give_up case. One design thought to address in a comment (not necessarily code): a deterministic first-file nonzero exit progressively quarantining wrong files is a property the crash path already accepts — say so explicitly in the WHY comment so the next reader knows it's a considered trade.