Brian Behlendorf wrote:
On Mon, 24 Nov 2003, Aaron Mulder wrote:
The truth is, if the whole thing is generated from source files in
CVS, so long as the actual source code is backed up, the web site CAN be
recovered in a relatively painless fashion -- just restore the CVS backup
and build the site. Am I missing something?
Others have made the point about the infrastructure team not wanting to
have to learn N groups' build systems in case we need to restore from CVS.
I wanted to add that even if you're storing 190 MB of generated content,
the actual amount of generated content that *changes* each time you do a
rebuild should be small, a proportion of the scale of other code changes
that happened or other documentation changes in the corresponding amounts
of time. If every generated page ends up being different from the last
time the build was run for something spurious like adding a timestamp,
perhaps that could be re-examined.
AFAIK the main issue is about the Javadocs, sourcedocs, etc. IMHO these
artifacts can also be removed from the CVS stuff, as having them
restored in case of an accident is not really necessary.
It would be better if these could be generated dynamically though, as
checking in 190MB of stuff is not that nice in any case.
--
Nicola Ken Barozzi [EMAIL PROTECTED]
- verba volant, scripta manent -
(discussions get forgotten, just code remains)
---------------------------------------------------------------------
