Dale J. Chatham schrieb:
> Understood. Trouble is, one can only check out a directory and there
> are 1,000+ files in one directory and 5,000+ files in the other. It
> would be nice to be able to do an end run around the "supposed to do it"
Despite the discussion is a few weeks old now:
IIRC, it is part of Subversion's philosophy to consider diskspace cheap
at the client's side (which allows to do a diff without asking the
server). Thus, there is no alternative to a complete checkout of the
directories where you want to maintain changes. Maybe it is possible to
re-organize the directory structure, e.g. with a subdirectory "0" for
files 0 to 999, "1" for 1000 to 1999, and so on?
> Ryan Schmidt wrote:
>> On Jan 7, 2009, at 13:11, Dale J. Chatham wrote:
>>> Dale J. Chatham wrote:
>>>> So, the question is, given the tree, can I traverse the tree, checking
>>>> (from essentially a command line) for each file/directory so I will
>>>> know whether to add or import the file/directory.
>>>> There are literally thousands of files, so checking everything out,
>>>> then creating the files isn't something I want to do.
>>> It appears that the way to do this is to first check out the directories
>>> I'll be working with.
>>> Then, as I create files, check to see if it already exists. If it does,
>>> just write. If not, write the file, then an svn add.
If the assumed new files all satisfy a certain shell-style regular
expression (e.g. "*.cfg"), you can simply use this RE in the add command:
svn add *.cfg
Subversion will yield warnings for files already versioned, but the
operation should succeed.
>>> At the end of the process, check the whole mess in.
>> Sounds like a good plan to me! :)
>> That's how you're really meant to make changes to a repository: check
>> out a working copy, make your changes, then commit them.
To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_subversion.tigris.org].
Received on 2009-02-03 13:23:09 CET