Sun Apr 13 20:53:32 EDT 2008


recently i ran into the rsync paper, about how rsync does its
magic. quite interesting: it uses 2 layers of hashing: a fast
incremental hashing, and a block hashing. the first one is used to
scan for blocks that are the same, the second one to verify they are.

anyways.. rsync. i'm using it in the following modes:

  * mirror / backup (make an identical copy)

  * merge (like copy, but doesn't delete)

  * in-place copy (for transferring large files with possible
    connection drops)