> Perhaps not a bug. I would submit that this behaviour does not match
> most peoples definition of recursion. To preserve the existing
> functionality (if thats important) perhaps the "-R" flag could be
> activated for "add" meaning "recurse all subdirectories even ones
> currently under version control".
>
> It also makes "svn add" appear to operate quite differently than "svn
> commit" which does recurse all directories.
...but... svn commit doesn't expect a list of one or more targets, either.
It operates on the meta-concept of the "working copy" as a whole (as defined
by the entries file, I think). Another issue is the "atomic unit"-ness of
the directory, so every subdirectory can potentially be treated as a working
unit of its own.
I still think the cleanest approach (at least at the moment) is a wrapper
script like I described before. Of course, I'm going to think that way
since I've already done it. :-/
>
> If I'm not mistaken, doesn't "cvs add" recurse all
> directories (its been
> a while)?
Couldn't say - not a CVS user.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Jul 8 18:28:57 2004