Jacob Atzen wrote:
>It is not possible to check out a single file. The finest level of
>checkouts you can do is directory level. You may export single files by
>using 'svn cat' but then it's no longer a working copy.
So then it's useless, since if it isn't a working copy, there's no way
to commit the changes?
>You can use checkout to checkout named directories, with and without
>subdirectories. The manual describes this at:
But as David Ripton said, -N is broken. So until it if fixed, I won't
bother using Subversion.
>I would suggest you make a local configuration file to tell your
>application the paths it needs and then not put this under version
>control. And then have your developers checkout whatever part of the
>repository they need to do their work and view the changes.
But how can they "view" those changes if their URL paths are all wrong
due to their working copy being outside the web hierarchy or not even in
a web directory?
>I don't understand how you today can manage a live
>site where people are editing the files on top of each other. Aren't you
>having problems with people editing files on top of eachother? And the
>site being unavailable due to programming work being done on the files?
No one ever edits files on top of each other. Did you not read my post?
We are using RCS. One person checks out a file with RCS (locking it). He
or she edits it, then checks it back in (unlocking it).
The site is never unavailable due to programming work being done on the
site, because the programming work is done on our development machine,
not our production machine.
I'm thinking we should just stick with RCS until Subversion makes more
sense. If the -N thing is broken, Subversion's pretty useless to me.
>I hope this has made the world of Subversion a little less confusing for
A little bit, yes, thank you.
Carole E. Mah
Brown University Computing & Information Services
To unsubscribe, e-mail: firstname.lastname@example.org
For additional commands, e-mail: email@example.com
Received on Sun Jan 23 20:30:21 2005