On Fri, 17 May 2002, Wayne Davison wrote: > so feel free to tell me exactly where I've missed the boat.
[Replying to myself... hmmm...] In my description of the _new_ protocol, my references to a generator process are not really accurate. The current generator process is forked off after the initial file-list session figures out what files need to be checked for differences, and it then churns out rolling checksums for the sender process. The "generator" in my previous description is really just a receiver-side scanner process (that looks for files that need to be check-summed). So, either the new receiver process would handle the checksum generation itself, or we'd need a 3rd process on the receiver side to generate the checksum data (and it would need a pipeline into the sender). As a first step in investigating this further, I'm looking into librsync to see if it might be easy to create a simple sender/receiver duo using this library. If anyone knows where some decent documentation on librsync is, please let me know (I'm looking for it now, but the tar doesn't appear to come with any decent docs). I was wondering if librsync manages to implement the protocol without forking off a separate generator process... ..wayne.. -- To unsubscribe or change options: http://lists.samba.org/mailman/listinfo/rsync Before posting, read: http://www.tuxedo.org/~esr/faqs/smart-questions.html
