Erik Bjäreholt

@johan-bjareholt I agree. I also notice that aw-watcher-jetbrains seems to not use the heartbeat API, which may make it unreliable. Edit: Looks like it is pretty unreliable and doesn't work...

I think in the Python version of query2 all functions that modify data do deepcopy's on input arguments in order to avoid mutating them. >querying would probably be insanely faster...

I think the server should respond with HTTP code 503, with a suitable error in the log. That way, watchers can queue events in memory, as appropriate. On Mon, Nov...

We should keep in mind that we might want to use the pragma [`secure_delete`](https://www.sqlite.org/pragma.html#pragma_secure_delete) or run a [`VACUUM`](https://www.sqlite.org/lang_vacuum.html) to ensure secure deletion of data.

> Maybe run VACUUM when aw-server is starting up? > Seems like a sane time to do it. If the operation is somewhat costly then I'd rather not as that...

>Maybe also make a unique identifier for each computer? I think we should at some point as people have already had issues with this (https://forum.activitywatch.net/t/remove-host-computer/358). Syncthing does it by generating...

>Fixed in #76 I appreciate it, but I kinda think that the solution you proposed was better (and is how Syncthing deals with nodes of different versions): >Maybe look at...

Great idea. We should talk about this IRL, but some points: - How do we keep it simple but still have the all the tools needed? - I'm not sure...

@johan-bjareholt I think something like: ```py window_events = query(...) # query and do AFK filtering for browser in browsers: # this for-loop would need to be expressed in the function...

>To call that "union" is very misleading in my opinion. A union should be a common ground, not exclude anything. For sure, I was just unsure what to call it....