Re: Proposal for $Revision$ keyword amendment

On 2005-10-12 12:28:27 -0400, John Peacock wrote:
> Vincent Lefevre wrote:
> > But I think that this number doesn't make much sense in practice.
> > And this is not Molle's vision:
> >
> > The keyword reflects the newest (highest) "last comitted revision" of
> > any item in the current working copy.
>
> Except that, in practice, the technique Greg is suggesting would be
> accurate most of the time. Only if someone committed after you
> committed, but before you updated, would the "last comitted
> revision" differ.

No, Greg said:

  Go to your svn working copy and look at .svn/entries. See the
  "revision" field in the XML entries (for the directory itself and
  for any files whose revision differs from the directory). That's the
  number which would be substituted in for $UpdateRev$ in my vision.

Except in rare cases, the "revision" field does not correspond to the
last committed revision amongst the items in the current working copy:
it is usually higher, because things may have been committed in other
directories (that are not descendants).

> If Molle _really_ has to have a direct measurement of the state of
> the entire WC, then svnversion is the only way to accomplish that
> (specifically because it deals with the WC as a whole, which the
> other Subversion CLI apps do not do).

Yes, "svnversion -c" has the wanted behavior.

-- 
Vincent Lefèvre <vincent_at_vinc17.org> - Web: <http://www.vinc17.org/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/>
Work: CR INRIA - computer arithmetic / SPACES project at LORIA
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org