Shells - non-blocking?

[David Bovill]

Roughly how would this work?

Lets give an example a "svn commit" - the idea is to do a regular svn
backup without interupting development scripting for instance - so I
want an asynchronous backup interface rather than having to wait a
second or two every few minutes for results to come back from the
online repository.

A shell script or shell command... can this be opened generally as a
process? Or is it as i thought - only something that is occasionally
coded into the application / code / script / shell command - (
whatever you call it :) by the author of the program?

And what would be the relation betwenn writing to that process and the
params you would send on the command line?