Maintenance release. Your timeline, your data, and how the app works are all unchanged. If you don't use GLANCEvault sync, there is nothing here for you.
Sync
lifeGLANCE saves to the vault twice over: on a timer every minute, and again shortly after you make a change, so an edit reaches the vault even if you close the app straight away.
The timer knew how to back off after a failure: wait, retry, wait longer. The save-after-an-edit path did not. It went straight out every time. On a device with a poor connection, or one the server had stopped accepting, every single edit fired another request that was never going to arrive.
Both now share the same restraint. After a failed save the app waits before trying again, and edits you make in the meantime are held and sent when it resumes. Nothing is dropped. Only the timing changes.
Two things follow that you may notice:
- The Cloud Sync screen can now show "Can't reach the sync server. Retrying in 30s." at moments where it used to say nothing at all. This is not a new problem. It is an existing one finally being reported.
- If the server has stopped accepting this device's credential, lifeGLANCE now notices when you next make an edit, rather than waiting for the next scheduled sync.
Under the hood
- Updated to
@glance-apps/sync2.0.0, which moves backoff, quota and credential handling down into the sync primitives, so every path inherits them instead of each caller having to remember. - That release also flipped a default governing how much of a partly-finished download is allowed to remember its place. lifeGLANCE keeps its existing behaviour, now declared explicitly rather than inherited: a large sync over a poor connection still resumes from the last page it finished instead of starting over.
- Six new tests cover the paused-push and resume paths, including one run against the previous version to confirm the new protection is doing the work rather than leftover app code.