I don't understand the justification for allowing these errors, beyond "well, it used to do that." As we well know, just cuz it *used* to do something, doesn't make it good/right :-P
On Fri, Mar 5, 2010 at 19:06, <rhuij...@apache.org> wrote: > Author: rhuijben > Date: Sat Mar 6 00:06:56 2010 > New Revision: 919674 > > URL: http://svn.apache.org/viewvc?rev=919674&view=rev > Log: > * subversion/libsvn_client/ra.c > (svn_client__open_ra_session_internal): > Ignore all errors on retrieving wc information. This matches > the behavior of this internal function in 1.6. This should fix > most errors on the buildbots, caused by running the testsuite > below a 1.6 working copy. > > Modified: > subversion/trunk/subversion/libsvn_client/ra.c > > Modified: subversion/trunk/subversion/libsvn_client/ra.c > URL: > http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_client/ra.c?rev=919674&r1=919673&r2=919674&view=diff > ============================================================================== > --- subversion/trunk/subversion/libsvn_client/ra.c (original) > +++ subversion/trunk/subversion/libsvn_client/ra.c Sat Mar 6 00:06:56 2010 > @@ -322,11 +322,19 @@ > if (base_dir) > { > const char *base_dir_abspath; > + svn_error_t *err; > > SVN_ERR(svn_dirent_get_absolute(&base_dir_abspath, base_dir, pool)); > - SVN_ERR(svn_wc__node_get_repos_info(NULL, &uuid, ctx->wc_ctx, > - base_dir_abspath, FALSE, > - pool, pool)); > + err = svn_wc__node_get_repos_info(NULL, &uuid, ctx->wc_ctx, > + base_dir_abspath, FALSE, > + pool, pool); > + > + if (err && (err->apr_err == SVN_ERR_WC_NOT_WORKING_COPY > + || err->apr_err == SVN_ERR_WC_PATH_NOT_FOUND > + || err->apr_err == SVN_ERR_WC_UPGRADE_REQUIRED)) > + svn_error_clear(err); > + else > + SVN_ERR(err); > } > > return svn_error_return(svn_ra_open3(ra_session, base_url, uuid, cbtable, > cb, > > >
