Turn back on the absolute path protection for the moment.

It is masking a bug that I'm tracking down in the SNAPC FULL - FILEM interations

Also make sure to cleanout the filem structure before asking for another
checkpoint file when not storing the files in place.

This commit was SVN r16109.

orte/mca/filem/rsh/filem_rsh_module.c

@@ -386,7 +386,7 @@ int orte_filem_base_rsh_rm(orte_filem_base_request_t *request)
 static int orte_filem_rsh_query_remote_path(char **remote_ref, orte_process_name_t *peer, int *flag) {
     int ret;
 
-#if 0 /* Some debugging */
+#if 1 /* JJH: Some debugging */
     /* If it is an absolute path */
     if( *remote_ref[0] == '/' ) {
         *flag = ORTE_FILEM_TYPE_DIR;

orte/mca/snapc/full/snapc_full_global.c

@@ -978,6 +978,11 @@ static int snapc_full_global_gather_all_files(void) {
                 exit_status = ret;
                 /* Keep getting all the other files, eventually return an error */
             }
+
+            tmp_argc = 0;
+            opal_argv_delete(&tmp_argc, &filem_request->remote_targets, 0, 1);
+            tmp_argc = 0;
+            opal_argv_delete(&tmp_argc, &filem_request->local_targets, 0, 1);
         }
     skip:
         /* Do a bit of cleanup */