> -----Original Message-----
> From: Hyrum K. Wright [mailto:hyrum_at_hyrumwright.org]
> Sent: dinsdag 5 mei 2009 22:53
> To: Greg Stein
> Cc: dev_at_subversion.tigris.org
> Subject: Re: svn commit: r37590 - trunk/subversion/libsvn_wc
>
> On May 5, 2009, at 3:44 PM, Greg Stein wrote:
>
> > On Tue, May 5, 2009 at 22:19, Hyrum K. Wright
> > <hyrum_at_hyrumwright.org> wrote:
> >> ...
> >> +++ trunk/subversion/libsvn_wc/adm_ops.c Tue May 5 13:19:56
> >> 2009 (r37590)
> >> ...
> >> @@ -2445,8 +2451,12 @@ svn_wc_remove_from_revision_control(svn_
> >> svn_error_t *err;
> >> svn_boolean_t is_file;
> >> svn_boolean_t left_something = FALSE;
> >> + svn_wc__db_t *db = svn_wc__adm_get_db(adm_access);
> >> const char *full_path = apr_pstrdup(pool,
> >>
> >> svn_wc_adm_access_path(adm_access));
> >
> > I know this isn't part of your change, but this seems dumb. By
> > definition, the baton's path will last longer than this function. Is
> > there a reason to make a copy of this string?
>
> I don't know / haven't looked at it too deeply.
>
> >> + const char *local_abspath;
> >> +
> >> + SVN_ERR(svn_path_get_absolute(&local_abspath, full_path, pool));
> >>
> >> /* Check cancellation here, so recursive calls get checked early.
> >> */
> >> if (cancel_func)
> >> @@ -2460,10 +2470,9 @@ svn_wc_remove_from_revision_control(svn_
> >> svn_node_kind_t kind;
> >> svn_boolean_t wc_special, local_special;
> >> svn_boolean_t text_modified_p;
> >> - svn_wc__db_t *db = svn_wc__adm_get_db(adm_access);
> >> - const char *local_abspath;
> >>
> >> full_path = svn_dirent_join(full_path, name, pool);
> >> + SVN_ERR(svn_path_get_absolute(&local_abspath, full_path,
> >> pool));
> >
> > You did this above.
>
> Sure, but the value for full_path has changed, so this will yield a
> different result.
But an entryname/basename joined to an absolute path will always return an
absolute path, so this step is not necessary.
The only case where joining a basename to an absolute path doesn't give you
an absolute path is when the basename is '.' or '..', but these values are
never valid as entrynames.
Bert
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2078486
Received on 2009-05-06 10:43:47 CEST