Plompy 0.3.1
Recovery and clock-safe workflows
This is a historical release. Its notes describe this version, not the current setup. See the latest stable release before starting a new installation.
This version predates Plompy Local. It covers the Discord engine; the separate local MCP gateway and starter kit arrive in 0.13.0.
GitHub notes and assets for this version may use the earlier Szpont name. Historical filenames and source links are preserved.
What changed#
Recovery and ordering fixes#
- Retry scoped memory commits after a transient Git failure without leaving staged content behind. Preserve operator staging; tracked deletions and renames remain supported.
- Keep same-thread FIFO order independent of wall-clock adjustments using a durable acceptance ledger. Continued workflows still yield fairly to other threads.
- Measure active execution budgets with a monotonic process clock, including failure accounting. Wall-clock corrections no longer refund or prematurely exhaust the remaining budget.
- Preserve the original timeout/cancellation cause when a gracefully stopping Codex process emits a later generic error.
- Report the installed browser package version during MCP initialization instead of a stale hardcoded version.
Upgrade boundary#
- Public SQLite schema advances from 1 to 2. Legacy jobs retain their previous
(created_at,id)ordering; new intake receives an immutable sequence. - Reject incompatible private-deployment databases before mutation. Back up state before upgrade; a binary rollback requires restoring its matching pre-upgrade database snapshot.
- Add offline regressions for transient commit failures, operator staging, clock corrections, continuation/recovery ordering, migration atomicity, variant mismatch and actual MCP initialization metadata.
Downloads and source#
This GitHub release has no separately uploaded download assets. Its source and published distribution details are linked from the release page.
Original GitHub release · Source at v0.3.1 · Changelog at this tag
Before you upgrade#
Back up private state and credentials, review the version's migration boundaries, and retain a compatible rollback snapshot. A release does not update an existing deployment automatically.