On 6/13/07, Arnauld Van Muysewinkel <arnauldvm@gmail.com> wrote:
> 2007/6/13, Andreas Mahel <andreas.mahel@gmx.net>:
> > Arnauld Van Muysewinkel wrote:
> > >
> > >
> > > On another hand, I still don't understand why SVN is preventing this...
>
> [SNIP]
> > An alternate use case for your problem might be called "Extend an
> > existing log message". This might be considered less harmful (by
> > admins);
>
> Indeed, that's my use case. It's OK as long as no log message can be
> *changed*, which is important for compliance reasons.
Why's that? Your repo admin can save the old message before replacing
it with the new one. That'll provide him/you with the audit trail
required.
> > however it appears that it hasn't been needed badly enough to
> > persuade a developer to implement it. And looking at the effort I
> > estimate behind designing this right (not just as a quick hack), I guess
> > there are far more pressing issues...
>
> For sure.
>
> > And I wouldn't blame subversion for the fact that your repo admin won't
> > allow functionality which is there.
>
> Certainly not. I was just naively thinking that it was probably more
> difficult to prevent an empty commit than to authorize, but I
> certainly lack knowledge of the internals of svn to judge.
HTH,
Erik.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jun 13 16:59:26 2007