Joseph Dane <jdane@studio3511.com> writes:
> > What if we simply overloaded the '?' to mean "any file not under
> > version control in the working copy." I mean, this is a *Subversion*
> > status message, yeilding information with respect to Subversion
> > working copies -- Subversion doesn't care if a non-versioned file
> > exists or doesn't exist on disk. Out of [the entries file], out of
> > mind.
>
> I could live with that, although we'd still be losing some
> information. I'll generate a new patch using this approach.
+1 on '?' for existent, but non-versioned, files.
-1 on '?' for non-existent, non-versioned, files (which are the kinds
of files that brought this thread up in the first place).
IMHO, if the user explicitly invokes "svn st" on a list of filenames
that includes one or more non-existent, non-versioned files, it's
acceptable to throw an error, or do something attention-getting.
Remember, shell globbing will never cause this to happen. Shell
wildcards like "*" only expand to things that actually exist. So this
situation can only happen if the *user* explicitly types the name of a
non-existent, non-versioned file.
This statement
> Subversion doesn't care if a non-versioned file
> exists or doesn't exist on disk. Out of [the entries file], out of
> mind.
isn't really true. Subversion does pay attention to non-versioned
files sometimes -- like when adding them or copying them, for
example. It would be gratuitously confusing for Subversion to print
? NOT_HERE
when NOT_HERE doesn't exist on disk or in the .svn/entries file,
because it would paper over the essential difference between NOT_HERE
and a file that exists but isn't under revision control.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:55 2006