Re: csup, cvsup (Re: [HEADS UP] packages not supporting destdir)

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Mon, 02 Apr 2012 02:00:28 +0400
    From:        Aleksej Saushev <asau%inbox.ru@localhost>
    Message-ID:  <87pqbrru2r.fsf_-_%inbox.ru@localhost>

  | Does it have any advantages over rsync?

rsync is OK for what it does, I guess, but is a resource killer, and
slow.   Using cvsup I can update the NetBSD repository (I do a local repo
copy, so using cvs isn't an option) from a server in Japan or
Europe (I'm in Thailand) in about 1/10 of the time it takes me to
rsync the repo from one local system to another, and with no visible
effect on the client sysstem (I use rsync locally as I've never even
attempted to find out what would be needed, or how I would, run a
cvsup server).

  | Also, if csup is "the preferred way to update sources within the FreeBSD,
  | NetBSD and OpenBSD projects and more" as claimed in description, why do
  | we not have it in base system?? Am I missing anything?

We don't have cvsup because it requires a modula 3 compiler, and as has
been mentioned, csup still seems to have problems (cvsup just works for
me as long as I can keep locating a suitable server to pull from - since
they're not project maintained, from time to time one just vanishes).
Given that, and how network & system resource friendly it is, I prefer
to just keep using it.

kre

ps: cvsup is fast because it knows what a cvs file is, and how updates
get made to them, whereas rsync just sees "some binary blob that is
different" when a cvs file has been altered.