Mark Hammond

> @mhammond Testing with about:sync I can see how it's hard to reproduce this issue. Switching to the about:sync tab triggers a sync most of the time (about 1 out...

> One question, is there a debounce period (where we don't start any timers) right after the 5 second timer expires and the sync happens? Nope, the 5 seconds is...

I agree with all of the above. In addition, we should prune history and history metadata. Ideally that could take some hints from the free storage space.

My only concern here is *when* to run it. Places only has 1 write connection, so whenever we do this, other operations which write to places will either be blocked,...

We landed an improved version in https://github.com/mozilla/application-services/issues/5011 and https://github.com/mozilla/application-services/pull/5057 - the CHANGES addition in that PR has: --8

(Oh, I meant to say that we copied the desktop maintenance strategy, both in terms of the small number of visits pruned, the target sizes specifications, etc, primarily because it...

> Thought about two more issues worth considering: > > * I think we'd want to use a dedicated writer for this though, so we don't interrupt/cancel the wrong operation...

It's certainly true that we are never calling it today, so ramping up slowly should in theory have *some* small benefit. Another challenge is that a world where this has...

The next pywin32 build will have 3.7 as the min supported version.

You can get wheels of recent builds via the github "actions" page - eg, https://github.com/mhammond/pywin32/actions/runs/3302591696 and download the artifacts. But yes, the intention is to close this once I've validation...