Eric Gillespie <epg@google.com> writes:
> Eric Gillespie <epg@google.com> writes:
>
> > In reviewing my resource URL log for this message, it now occurs
> > to me that I can probably come up with a test case for
> > authz_tests that will highlight the problem. What if I change
> > authz_open_directory to try to commit a change to both A/B/lambda
> > and A/B/E/mu? Wouldn't that cause it to do this:
> >
> > open_root A
> > open_directory A/B
> > open_directory A/B/E
> > open_file A/B/E/mu
> >
> > and thus trigger the authz failure when it tries to open_dir A/B/E?
>
> No, reading authz_open_directory more closely, it already would
> be doing an open_directory on A/B; the only reason it isn't is
> because A/B and A/B/E are at the same revision in the working
> copy. So that's what I need to change.
No, I just can't get it to fail via svn; I guess there's
something else different about the way svn_client is using ra
that I don't realize. Oh well, moving on to the test suite work...
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Oct 26 00:26:59 2007