Jonathan Wilson wrote:
>We have close to 1450 programs in this one directory, which will be
>moved into sub directories as it becomes more organised. I know this is
>ridiculous but like everything it takes time to fix. A cull is planed
>to remove a lot of the redundant code, but first it has to get into SVN
>so we'll have a complete backup & historical copy. Then we can sort out
>files by moving them into the correct subdirectory. When completely
>re-organised it will be branch into new project. But.....
>
>In the mean time development must continue, our environment has 12
>programmers. With SVN this would mean each one would require their own
>complete copy of the source (1450 programs). This is inefficient as it
>wastes file-system space (Programmers * Source Disk Usage). Then
>there's the issue of finding a program amongst the other 1450. Again
>this is our own fault, but I'm quite surprise such a highly rated SCCS
>like SVN can't handle a single file checkout.
>
"Disk space is cheap, bandwidth is expensive/unreliable" is a principle
which has guided the design of Subversion. Each programmer will
experience a delay when they first check out the repository but after
that you will find that commits and updates happen very quickly. The
benefits of having all the files on your local hard drive is usually
worth it.
Another approach you may consider is using WebDAV with
Autoversioning[1]. Every modern OS allows mounting of a WebDAV share.
Depending on your OS, you can either edit the files you want in place,
or copy one file down, edit it, then copy it back. The problem with this
is that your log messages will be just a generic autoversioning message.
[1] http://svnbook.red-bean.com/en/1.1/apc.html
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jan 26 16:17:00 2005