Tamas Rudnai wrote:
> I usually create a development branch and commit everything there - even 
> the
> unstable or uncompilable ones - so I can roll back if I deleted something 
> by
> accident or can just make diffs to see what did I do... Then when 
> everything
> is ok I just merge the branch to the main trunk, so that main trunk is
> always the stable and trusted one.

That's a great idea. My rule is that in order for me to committ the code, I 
have to get a clean compile first. Most of the time it's easy, since the 
changes between committs are small and bugs are easy to find and fix. 
However, from time to time the changes can be substantial, span many files, 
and take several hours to complete until I have a clean compile. A 
development branch would allow me to committ even stuff that doesn't 
compile.

> This way is also good for bugfixes - creating a new branch for each bugfix
> and then merge back when the bug is fixed, so that I can see if the what
> effort made for the fix - if later on would turn out that the fix created
> many new bugs :-)

Nice!

> BTW: I use CVS and I do the same with local repository on external HD - 
> but
> I do not know why SVN is better or differrent? Also working with Perforce
> but to be honest can't see the advantage over CVS either - would be nice 
> to
> see some opinion.

As you can tell, I'm relatively new to version control systems, I've only 
used Subversion.

Vitaliy 

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist