Re: Creating a new repository from existing project

[Ross Patterson]

On Monday 30 December 2002 07:56 pm, Hew Fen Phin wrote:
> Here's my concern, let say there's a version one and a version two, if I
> import version one followed by version two, will cvs be able to notice
> the difference between the two sets of files (although version two is
> based on version one, cvs does not have a previous record of that) and
> incrementally include version two instead of overwriting what was in
> version one. 

CVS never "overwrites" files, it always maintains incremental differences 
between their versions.  As to not knowing that version 2 is based on version 
1, that's not quite true.  When you check in version 2 with version 1 already 
present, CVS will learn the relationship.

> We are, in a sense, trying to manually recreating the whole
> project history in cvs (including branches etc.), is this possible ?

You should consider reproducing the historical record of your project as if 
it had been CVS-controlled from the start.  You presumably have a 
large-enough set of backups to work from (you *did* back up the project 
directories frequently, right?).  Create an empty repository and check in the 
restored contents of each backup in sequence.  That's your main trunk, 
growing and changing over time.   Any time a "branch" occurred (however you 
did that), create a branch in CVS and begin checking in the backups of that 
branch appropriately.  Any time you created a release of your software, 
create a CVS tag for it after checking that level in (or do whatever you plan 
to do for release identification in the future).  

For extra credit, continually reset your system clock to make it appear that 
the checkins occurred at the time of the backup, and you'll be able to use 
the appropriate dates in CVS commands :-)
-- 
Ross A. Patterson
Chief Technology Officer
CatchFIRE Systems, Inc.
5885 Trinity Parkway, Suite 220
Centreville, VA  20120
(703) 563-4164