Re: Fwd: Daemonizing Unison and why this is easier said than done
Greg Troxel <[email protected]> Fri, 13 Oct 2023 18:53:06 -0400
| Newsgroups | gmane.network.unison.devel |
|---|---|
| Message-ID | <[email protected]> |
Josh Marshall <[email protected]> writes: > Having my two computers sync as changes happen seems nice and > ostensibly easy to implement. Famous last words, I know. But why > would a Unison daemon be difficult to impractical? Are you aware of the various fsmonitor implementations? Leaving unison running (and the corresponding remote) seems to be routine for many already. So I'm not sure which piece you think is missing, and what it might do instead. See "-repeat watch" and src/fsmonitor*. Certainly the fsmonitor world could be better documented. I wonder if you mean support for closing stdin/stdout, using syslog for logging, controlling terminal dissociation, etc. This is fairly minor, and can be achieved with nohup(1) and redirection, plus some futzing around for logging. If so, that seems like a reasonable thing to add. You also might mean support for long-running processes on both sides that survive disconnection. That raises issues about partial processing while disconnected, vs only doing connected processing and queuing input. It also raises issues about ssh and reconnection in terms of future access to authentication, vs running over D-TLS or some other schemes. You didn't mention this, but unison operates under a single uid, and one could want synchronization of files belonging to multiple users. So far we haven't really crossed that line fully.