Can't cvs import empty directory structure 

http://groups.google.com/groups?q=cvs+initial+directory+tree&hl=en&scoring=r&rnum=6&selm=535991b0.0106201413.7300fcf0%40posting.google.com

Newsgroups: comp.software.config-mgmt
Date: 2001-06-20 15:13:21 PST

I'm brand new to CVS, but have installed version 1.11.1p1 on RedHat 7.1. After initing the repository at /cvs/cvsroot I got all of the CVSROOT files OK. I have looked at section 3.1.3 ("Creating a directory tree from scratch") in the Cederqvist text, and I tried to do the same within my directory structure…

/cvs
 \
  + cvsroot
   \
    + CVSROOT
    |
    + nms
     \
      + eicore
       \
        + src

The book says to:

As root, and with CVSROOT=/cvs/cvsroot I did the following:

Regardless of whether I use nms/eicore, nms/src or nms as the argument I get the following error message:

"cvs [import aborted]: attempt to import the repository"

Can't cvs import empty directory structure 

>I get the following error message:
> "cvs [import aborted]: attempt to import the repository"

It sounds like you are in a directory within the repository, rather than in a working directory. i.e. a checked-out one, somewhere else. (It can be part of a checkout of the whole repository, or just a part.)

Erik

Can't cvs import empty directory structure 

The repository is the complete tree starting from the position started with lower case written cvsroot. I think you should rename your directories into things like /Repository and do CVSROOT=/Repository after that an "cvs init". The basic Idea of CVS is to allow concurrent development. This means many users (e.g. you) can checkout a copy from the repository into their work area (usualy their home dirs).

After that you get the following tree:

/Repository CVSROOT

If you do "cvs import … XYZ BRANCH START" you get the following:

/Repository CVSROOT XYZ

Using of CVS should only be done as non root which means cvs co, cvs import etc. Administration of the Repository have to be done as root.

Karl Heinz Marbaise