[GLLUG] CVS newbie

Jeremy Bowers jerf at jerf.org
Thu Sep 9 13:43:57 EDT 2004


Mark Szidik/mlc wrote:
> I am new to CVS, I want to use it in a bit of a less traditional way - for
> tracking website changes.   Does anyone do that?  What I am confused about
> is where the web server should look to for it's copy of the site.  Does it
> look in the CVROOT's repository location?  That seems to make sense to me
> since any updates to the repository will get to the website immediately.
> Or does it work better to occasionally checkout the site to the location
> where the web server looks for the files?

As Ben said, the latter.

Depending on where this is going, you may also want to consider 
"export"ing the repository to the web server, rather than "checking it 
out". Checking out the directory gives you something that you can e dit 
in place, and check in later, which includes the CVS directories that 
CVS uses for bookkeeping. On general principles, you probably don't want 
people accessing those files on your webserver by adding CVS to the URL.

There usually isn't anything *horrid* in there, but you may reveal the 
name of your CVS repository and the list of all files being managed 
under CVS... I don't think passwords are ever stored there (I always use 
SSH), but I'm not 100% certain, so *I* wouldn't depend on it.

I qualified this with "depending on where it is going" because in Apache 
it is easy to deny access to those directories; I don't know if other 
webservers make it so easy. Still, why fuss with denying access to 
things when you can make it so those things simply aren't there to be 
accessed?


More information about the linux-user mailing list