[cvsnt] Re: CVSNT for Web Development?

[Bo Berglund]

On Sun, 15 Aug 2004 13:16:01 -0600, "Siegfried Heintze"
<siegfried at heintze.com> wrote:

>I'm presently developing web pages in my office on my desktop machine in the
>directory C:\Program Files\Apache Group\Apache2\cgi-bin because that is
>where Apache HTTPD executes perl scripts.
>
>Now it seems logical to me to use the import command on this directory. But
>I also need to continue working in this directory to see my results. So now
>what do I do? Delete the entire directory and use the checkout command to
>recreate it? Yikes.... OK....

You can tell CVSNT to create the sandbox when you import and you can
disable the creation of the vendor branch as well. When you do this
you will import to CVS and convert your project folders to proper
sandbox folders at the asme time. Very convenient and only available
with CVSNT.

>So is the recommended procedure to make the sandbox and the WebServer
>directory the same directory?

I have done it in a slightly different way:
I have created a sandbox for the web files on my own PC like described
above. This sandbox is set to be used by my Apache as a web files
folder so my own Apache will show the webpages as I develop them, no
need for any extra steps.

On the CVS server itself I also have Apache running and there I have
created a sandbox with the web files by checking out the module to a
folder that Apache is configured to serve.
Then I have added a call to a batch file in the postcommit script that
triggers for each commit done to this particular module. The batch
script moves to the checked out webfiles sandbox and issues a cvs
update command there. The effect here is that whenever a developer
commits files to the module they get updated on the CVS server Apache
as well and you can immediately check them from anywhere on the
internal LAN.
>
>So I make some changes and commit.
>
>Well now, suppose I need to go over to Marty's office and work with together
>with Marty on the Marty account and modify those perl scripts. So I check
>them out on Marty's machine and make changes and use WebDAV or FTP to copy
>from my sandbox to the directory in C:\Program Files\Apache
>Group\Apache2\cgi-bin on the desktop machine of my office until the web
>pages display the way I want them to. Then I do a "cvs commit".

NO! You do not FTP or copy any files, you check the pages directly on
Marty's computer because you have installed Apache also there as
above.
Or you make the changes and commit to the server from Marty's PC, then
check the pages by looking at the CVS server Apache.

>
>What do I do when I return to my office? "cvs update"? Will that merge the
>old files with the new? That is not what I want. Do I have to release my
>sand box and check it out again? This seams strange and awkward.

No need for any such strange operations! Each sandbox is kept by
itself and changed files are sent to the CVS server and after that any
update from another PC will get him the new file contents.
>
>Assuming the answer is yes, let say I do a lot more work from my office
>machine that is hosting the site.
>
>Now assume Marty does no work on these files until I return to his office.
>Do I release the sand box again and check out again?

NO!

>
>Is it normal to release sandboxes frequently?

NO!



/Bo Berglund