On Fri, 27 May 2005 kfogel@collab.net wrote:
> lundblad@tigris.org writes:
> > --- trunk/subversion/svnlook/main.c (original)
> > +++ trunk/subversion/svnlook/main.c Thu May 26 16:42:21 2005
> > @@ -538,7 +545,7 @@
> > apr_pool_t *pool)
> > {
> > apr_array_header_t *path_pieces;
> > - svn_error_t *err;
> > + svn_error_t *err, *err2 = NULL;
> > int i;
> > const char *full_path, *dir;
> >
> > @@ -558,7 +565,7 @@
> >
> > path_pieces = svn_path_decompose (dir, pool);
> > if (! path_pieces->nelts)
> > - return SVN_NO_ERROR;
> > + goto cleanup;
>
> See scenario (b) in my comments below.
>
Yeah. Good atch! ACtually, the first test (or the second) is bogus.
svn_path_decompose only returns 0 components in the empty path case. I've
remvoed the first test and made the second test return err. (And I added a
docustring as well. Didn't notice it was missing; I just stumbled upon
this routine while debugging the rest of this patch...)
Thanks for reviewing,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 27 22:52:26 2005