On Fri, Nov 30, 2012 at 12:51 PM, Eric S. Raymond <esr_at_thyrsus.com> wrote:
> I'm not sure what the entire right design fix for Subversion is here, but
> I *am* sure you guys should be paying attention to this now so you can have
> the fix ready and deployed by the time forge evolution makes it urgent, which
> I would say will be on the close order of three years out. Sooner, if I
> actually get to pay concentrated attention to the problem.
>
> So think: What would it take to divorce identity-for-attribution from
> identity-for-authentication, making the former user-settable the way
> DVCSes do? The challenge, of course, is doing it upward-compatibly.
>
> I apologize for not being able to make any concrete suggestions here;
> outside of the dumpfile format I don't know your code and protocols
> well enough.
I'm not sure it's something we want to change for everyone. I suspect
you're the first person that's ever raised any complaints about this.
This is really a philosophical difference between a centralized
version control system and DVCS.
Depending on your repository access setup there are ways around this.
If the server side knows the value you want to put in svn:author
instead of the authenticated user name then it's pretty trivial to fix
with a hook script.
In my opinion the thing to do here is to allow auto revision
properties (possibly on a per repo/server) basis. Then we could pass
along some user configured extra value with each commit and server
admins could decided to put a hook script in place that puts it in
svn:author if they wanted.
If that's something we want to do as a project or not I don't know.
Received on 2012-11-30 22:05:51 CET