#273 wall-clock timeout in entry.sh

[VasilevNStas]

When baza.rb pushes a job to the server, entry.sh runs judges update to process the factbase. Previously there was no time limit — if judges hung (e.g. due to a bug, network issue, or unexpected factbase state), the container would run forever and baza.rb's polling of finished? would time out after 10 minutes with no clear indication of what went wrong.

This commit wraps the judges call in GNU timeout(1):

timeout ${JOB_TIMEOUT:-600} judges update --summary --max-cycles=3 --no-log ...

What this changes

Before	After
No time limit — container hangs forever	timeout enforces a hard ceiling
baza.rb sees only "job disappeared" after 10 min	Exit code 124 + "timed out" in stdout
Ops have to manually kill stray containers	timeout sends SIGTERM → SIGKILL automatically

Key details

• Default timeout: 600 seconds (10 minutes), matching baza.rb's polling ceiling
• Configuration: override via JOB_TIMEOUT environment variable
• Exit code: 124 (GNU timeout standard) propagates to the container, giving baza.rb a deterministic status instead of ambiguous "still running?"
• Diagnostics: "timeout: the command timed out" appears in stdout

Why not use the existing --lifetime flag?

The judges gem has a --lifetime option, but it is an internal Ruby timeout. An OS-level timeout is more robust — it catches any hang, including Ruby VM freezes, native extension deadlocks, and blocked I/O.

Compatibility

timeout(1) is part of GNU coreutils and is available in every Linux-based Docker image (all our CI images, yegor256/rultor-image, etc.). macOS users need brew install coreutils (gtimeout) for local testing, but production always runs on Linux.

Checklist

• bundle exec rubocop — 0 offenses
• bundle exec rake — all tasks pass
• HoC ≤ 133

@yegor256 please review

[VasilevNStas]

@yegor256 — this PR is part of a 3-PR series that addresses a critical reliability gap in the baza.rb ↔ entry.sh pipeline.

The problem: When baza.rb pushes a job, entry.sh runs judges update with no time limit, no convergence detection, and no signal handling. If judges hang or fail to converge, the container runs forever and baza.rb's polling of finished? times out after 10 minutes with no diagnostic information — the user sees a silent failure.