Why are you forced to use Eclipse?

As far as I know directories can't be deleted in CVS, so how
should Eclipse offer you that functionality as a frontend?

Regards, Lothar

mr_nice_1973@yahoo.com schrieb:

This problem is based on CVS not eclipse!
If you use a diffrent source repository it might be possible to use
differend "UN-user friendly" program.

CVS is no desinged to delete folders, only files are tagged with
versions.

If you are able to maintain CVS, you will know how to get rid of
unnessecery folders. It is not eclipse!

My advice is to read ore about CVS.
http://www.andamooka.org/reader.pl?section=cvsbook
http://cvsbook.red-bean.com

Cheers Dirk

You have never used EMACS? WordStar? MS Word?

Eclipse is bad mainly because of the documentation or more lack
thereof.  It is hump to get over, then it is not half bad.

What you can do is to have Eclipse delete any empty directories it
retrieves from the CVS repository. It does this by using the -P option
for the CVS checkout/update commands. You can set this within Eclipse's
properties for Team->CVS under the "Files and Folders" tab, there should
be an option for Prune empty directories.

Most teams I've been on come up with some sort of policy for dealing
with this. This is what I've most often done in the past:

 For the most part, we rely on the -P option (or Eclipse setting) to
prune empty directories, and this is fine for most things. Typically, we
don't purpose create directories under CVS that are not intended to have
anything committed to them. These, we create through our build system.
We typically like to be able to pull back at least the last three
releases from our repository. If you have deleted directories from your
repository that contained files that were part of one of these releases,
then there's going to problems...

We track the directories that have nothing in them, and periodically
review directories from older builds that we no longer require *instant*
access to. Before making any changes, we will create an tarball of the
existing repository, and archive it off somewhere. Then, we physically
remove/move any older directories we want from the repository, and then
delete any tags/branches that include one of those directories from the
repository. Access to the older repositor(y/ies) can easily be done by
opening the tarball in another location.

Then developers can either perform surgery within their local sandboxes
to remove the directories that no longer exist within CVS, or they can
delete and checkout the project again.

None of this is particularly difficult to do or manage -- its just a bit
of a nuisance, and once your baseline has stabilized, it doesn't occur
all that often. You might consider subversion as an alternative...

One last thing... in the early stages of project development when many
of the biggest changes to project structure are most likely, and no
releases have been tagged, it's not uncommon in my experience to delete
directories/files from the repository often. Usually this ends up not
being a problem -- occasionally, I'll delete something that I wish I
didn't. When you *do* decide on deleting a file/directory from the
repository, it's critical that it be communicated to others on the team,
so they are aware of it, and can plan for it accordingly.

Hope this helps,

Rob