> From: Carlos Costa Portela [mailto:[EMAIL PROTECTED]]
> We make two types of changes: simple ones (bugs and simple text
> changes) and others more complex (new apps in development). In our
> situation we feel we cannot use cvs in the classical way.
>
> We're really not sure how to make this work well, so if
> you could
> help point us in the right direction, we'd greatly appreciate it.
I'm hacking up web stuff as well using CVS to control my source code.
The way I do that, is that I have the trunk marking nothing in particular,
but have every developer on a branch.
In a scenario for 2 developers I would have:
changes...
Dev B ------------- merge...
Version 1.0 -----------| |-------------
| Dev A ------------- |
| | merge...
| |
main trunk -----| |----------- Main Trunk
After the final merge to the main trunk, I branch off Version 1.1,
branch off there the Dev A and B branches.
If now a bug occurs in the stable Version 1.0, I just branch off a Fix
trunk, propagate the changes there to the Version 1.0 and the main trunk
and have the Version 1.1 branches updated.
>From the developer's environment point of view, every developer has his
own web server, database scheme, unix account etc. So they can develop
whatever they need, and only the stable (code frozen, ran thru a
User Acceptance Test, etc.) is propagated to the production Web Server.
Any other questions, I'll be more than happy to provide more detailed
insight.
Regards,
Guus
