test: synchronize process timeout fixture

[steipete]

Root cause

The descendant-cleanup regression started its 500 ms kill deadline immediately after launching its shell fixture, then read the PID file only after the timed operation returned. Under scheduler pressure, the timeout could terminate the shell before it spawned the child and wrote pids, producing NSCocoaErrorDomain 260. The temporary directory cleanup was not early; the file had never been created.

Fix

• move the descendant process-group regression to the ProcessTimeout owner tests
• have the fixture persist both PIDs and emit a readiness byte through a pipe
• close the parent pipe writer and keep the descendant from inheriting the readiness pipe
• begin the timeout only after the parent receives readiness and parses both PIDs
• add the 0.13.5 reliability changelog entry

Proof

• exact regression: 20 consecutive runs passed
• full suite: 501 tests passed
• lint: passed with existing warning baseline only
• production arm64 CLI and arm64e/arm64/x86_64 helper build: passed
• live binary: imsg --version returned 0.13.5 and imsg status --json returned a valid SIP-safe payload
• pre-commit autoreview at P2 scope: clean

[clawsweeper]

ClawSweeper status: review started.

I am starting a fresh review of this pull request: test: synchronize process timeout fixture This is item 1/1 in the current shard. Shard 0/1.

This placeholder means the worker is alive and reading the current context. I will edit this same comment with the actual review when the claws are done clicking.

Crustacean status: shell secured, claws on keyboard, evidence pebbles being sorted.