Release v0.5.0 · eyelock/ynm

#106 345e5f4 Thanks @eyelock! - A new sync.auto setting (off by default) syncs distributed memory in the background from the agent's Stop hook. When it is on and a distributed mount has writes the remote has not seen, the hook starts ynm sync --quiet detached: a full fetch, merge and push, never the personal store, and the turn does not wait for it. It runs at most once per sync.intervalMinutes per repository (default and minimum 5, because the Stop hook opens the store once per 5 minutes per session; at most 1440), and a write made inside the interval goes out at the first Stop after it. Errors never reach the session.

A successful ynm sync now records what the remote holds in the remote-tracking refs for the notes and documents it pushed, so unpushed writes can be detected locally with no network.

#107 bc22cd3 Thanks @eyelock! - A --mount naming a mount that does not exist is now refused with exit 2 and the mounts there are: no mount "projet" (have: personal, project, team). Before, recall, context, list, export, review, reindex, sync and wiki build matched nothing and exited 0, so a typo, or --mount project run outside an initialised repository, looked like an empty store.

Install: see https://eyelock.github.io/ynm/#/how-to/install (Homebrew ynm standalone or ynm-slim, direct download, Docker).

What's Changed

Back-merge release 0.4.2 into develop by @eyelock in #105
feat: sync distributed memory from the Stop hook when sync.auto is on by @eyelock in #106
fix: an unknown --mount is refused, not read as an empty store by @eyelock in #107
Release 0.5.0 by @eyelock in #108

Full Changelog: v0.4.2...v0.5.0