Snapshots are the first step into testsuites and/or continuous builds.
For example, i daily use and develop on a Fink packaged Caudium, and
it's so easy to update it to current bleeding edge when a snapshot's
available.
Off course this could be done from a CVS repo, but that would be more
long and complicated on my side. More, reading README.cvs, i guess it
wouldn't be side effect free, ie some people could get different
snapshot from the same CVS source depending on some autoconf "magic".
My last argument would be that's it's way easier for testing: snapshot
2008XXYY is more specific than "get a cvs/svn snapshot number 42 or from
yesterdayy night". It's also easier to get a revision from the past.
I hope i've made my point :)
Archive powered by MHonArc 2.6.16.