GNU Arch

TJ Hunter tjhunter at gmail.com
Sat Mar 5 12:28:22 MST 2005


The live servers don't auto check out. For us, we _can't_ break the
live server, so any new code has to go through testing on the dev
server before it's approved to get pushed to the live server.

We just check out a repository on the live server and cvs update when
necessary. You'll want to looking into branches and all that stuff. If
you find there's an emergency bug fix to push to the live system in a
file that you've done a lot of development on recently, you don't want
to put all the new changes up to the live system yet, just the bug
fix. So you can branch the code so that you can fix the old version
and commit that, then when you do a cvs update on the live server, it
will only get the bug fix. Someone else could probably explain this
better than I, though. :) (google is a good place to start.)

Like Peter said, to trigger and auto-update you can edit the loginfo
file in the CVSROOT. I can send you an example when I get back to work
on monday if you haven't found one already.

I think the best reason to have a web server running on your local
machine is that you can save your changes in a text file, switch to
your browser and test your changes right away. You don't have the
extra steps of commiting it, or copying it over to another web server.
Development goes a lot quicker. Especially when you're debugging a
bunch of var_dumps in php.

-TJ


On Fri, 04 Mar 2005 16:02:07 -0700, Eric Jensen <eric at emstraffic.com> wrote:
> That's a very appealing idea.  I have been pushing for us programmers to
> use Linux for our workstations recently, maybe I can add this to my
> arsenal.
> 
> How do you then get the code to your live servers?  Do they
> auto-checkout as well?  How do you trigger that?  Treat me as if I have
> only spent a few days getting a CVS system slapped together. ;)
> 
> Eric Jensen
>



More information about the PLUG mailing list