On Tue, 21 Mar 2000, |}avid (opeland wrote:
We've also been developing stuff with CVS to help with large scale website
maintenance.
> This will only work if your repository is on the same machine as your webserver
> and if your webserver is only one one machine.
>
> For something more flexible, you essentially have to have a cronjob do periodic
> cvs updates.
>
> One of our projects does this. The website is huge and changes are pushed live
> 4 times a day. If changes aren't made by the time of an update, they have to
> wait until the next one.
Our approach is to use some ftp mirroring programs (Net::FTP perl stuff).
The uploads are triggered from a CGI interface and end up as a CVS export
to a /home/publish area.
> What turned out to be more of a problem is making sure files get comitted,
> added, and deleted as necessary where people are editing.
our users haven't seen it yet (just coming out of alpha) but we anticpate
problems here too.
> Typical HTML person
> doesn't know UNIX and can't understand how CVS works and typically gets
> confused due to CVSs cryptic error messages and subtle behavior (no slight on
> CVS; it's a programmer's tool). We built a web interface that abstracts CVS's
> necessities away from the HTML person.
yep sounds like our HTML ppl all right :) For this first tryout we wrapped
and boiled the user-controlled side of CVS down to Check IN/OUT, Preview
and Publish.
cheers
Gavin
_______________
Online Engineer
