Open j6k4m8 opened 1 year ago
How about multiprocess rather than multithread here? Proposal:
Totally — love it. Each process can be super lightweight too since 99% of the lifecycle is waiting for an HTTP round-trip...
...actually pretty low-importance until #10 since git rate limits will prevent this from actually being useful to us.
would be nice if polling strat was separate for each paperremote. currently, if one paper is being actively edited, all papers get polled more frequently.
Configurably, we should multithread the polling server, multiplexing across documents (but not across edits in one document, in case one changes the context of another)