Fix updater relaunch deadlock

[bifrost-proxy]

Summary

• complete the termination persistence flush without depending on the blocked MainActor
• defer updater-triggered termination until the installer task returns
• add an isolated process-exit regression and run it in CI/release workflows

Validation

• swift build
• swift run Pesty --verify-localization
• swift run Pesty --verify-history-settings
• swift run Pesty --verify-updater
• scripts/test_update_termination.sh .build/debug/Pesty
• isolated seed + two restart UI regression
• scripts/test_deletion_sync.sh .build/debug/Pesty
• git diff --check
• git diff --cached --check

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ee8dccdccc

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Tolerate ps failure after the child exits

When Pesty exits normally and Bash reaps it before this poll, ps -p "$test_pid" finds no process and returns nonzero; because pipefail and errexit are enabled, the assignment terminates the script before the empty-state check, wait, and result assertion. This makes the newly mandatory CI and release checks fail precisely when termination succeeds; tolerate the expected ps failure before examining state. This prevents reliable verification of the updater's required old-PID exit behavior.

AGENTS.md reference: AGENTS.md:L468-L470

Useful? React with 👍 / 👎.