Is it possible to ask sync_forever to stop its loop gracefully?

From glancing through the code, it seems that the only way is to cancel the async task -- but then I'd need to code all my callbacks to handle cancellations correctly, and that seems hard. Maybe there is a better way?

Would you consider a PR that would change that?

(edit: I meant, would it be ok if I prepare a PR that would add a function to gracefully stop sync_forever -- or would such change go against the philosophy of the library?)

okay i figured it out........ it's not simple but, you get the response from the sync callback and have to go in to the rooms list in it and get the timeline's prev_batch token and then use it in a room_messages call and it returns chunks of events at a time..

i have kind of a ridiculous question.......... one of my friends on here for some reason wants to change to a different account and wanted to try to delete all of his messages off of his current account he is on before he does that.. I know some other platforms have that feature but i don't think most matrix clients do..

i know it would be a lot of data transmission getting like the entire history from every room to try to automate the deletion of all of one's own past messages they sent........ is there a more efficient way to do it, to where the homeserver can only return back the events of a particular user?....... The API docs on the room_messages have a parameter called message_filter but it doesn't say how to use it any where except for message_filter (Optional[Dict[Any, Any]]) – A filter dict that should be used for this room messages request.

so, is there an efficient way to delete your own past messages from all of the rooms? or should i just tell him it is a ridiculous request and not even bother with trying to automate it?..

I don't think matrix homeservers probably have any type of API call like that to delete all user history data at the homeserver level

It's been discussed many a time before and it seems to be unwanted; it would enable more harm than good. There is, however, a proposal to allow redaction of multiple events at once using only one redaction event.