Ph. Marek wrote:
And regarding #3:
I'd use the same repository for all similar clients. Just because for a
new machine you simply do a "svn cp URL/machine1 URL/machine2", change a
few files (name, IP address) and go! Without having to copy some big
amount of data.

Furthermore you can do something like "svn diff URL/machine1 URL/machine2"
- which is not as easy with separated repositories.

And as soon as issue 2286 is done in subversion
(http://subversion.tigris.org/issues/show_bug.cgi?id=2286) it's even more
appealing to keep all machines in a single store.

Interesting idea, I'm on the fence whether it's a good or bad idea to touch old revs files in a FS repository. While I doubt that anyone was mounting old revs off of read-only media, it was an option with FS. Probably not an important enough reason to keep old revs as "no touchie".

And yes, that feature would indeed make it more appealing to keep machines in a single store.

I hadn't thought about the ability to do a "svn diff" between URLs. We've only been using SVN for about 9 months in production (12-15 personally) and for a lot of things we're still in the VSS+SourceOffSite mentality. Although I'm pretty sure that you can do a svn diff between repositories by fulling defining the URL?


It seems like it would be better to use separate repositories
(along with individual SSH keys for each client machine).
You can use individual keys per machine - you just have a bigger
authorized_keys.
And the rights per-machine can be restricted on a svnserver-base too, IIRC.

Hmmm... guess I'll have to go looking at SVN's permission scheme again. It's something that having separate repositories kept me from having to deal with. (I could simply give the individual user account on the SVN server the proper permissions to a single repository.)

OTOH, there are possibly some other advantages to using individual repositories:

- Easy to dump/load an individual repository without affecting other machines.

- Easy to use "du -hsc *" in /var/svn to see how much repository space each machine is using.

I'm on the fence now... I'll probably stick with individual repositories since most of our machines have different purposes (we don't have more then 2 machines that do the same thing).

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to