On 6/22/06, Lars Marowsky-Bree <[EMAIL PROTECTED]> wrote:
> linux-ha CVS committal > > Author : sunjd > Host : > Project : linux-ha > Module : lib > > Dir : linux-ha/lib/lrm > > > Modified Files: > lrm_msg.c > > > Log Message: > remove the filter as Andrew's advice > =================================================================== > RCS file: /home/cvs/linux-ha/linux-ha/lib/lrm/lrm_msg.c,v > retrieving revision 1.26 > retrieving revision 1.27 > diff -u -3 -r1.26 -r1.27 > --- lrm_msg.c 29 May 2006 11:55:53 -0000 1.26 > +++ lrm_msg.c 22 Jun 2006 10:29:05 -0000 1.27 > @@ -1,4 +1,4 @@ > -/* $Id: lrm_msg.c,v 1.26 2006/05/29 11:55:53 andrew Exp $ */ > +/* $Id: lrm_msg.c,v 1.27 2006/06/22 10:29:05 sunjd Exp $ */ > /* > * Message Functions For Local Resource Manager > * > @@ -63,11 +63,8 @@ > > if (g_hash_table_lookup(merged, key)) { > return; > + } > > - } else if(strncmp(key, "CRM_meta_" /*CRM_META*/, 9) == 0) { > - /* Never repopulate CRM meta attributes */ > - return; > - }Andrew, is this really safe now? Will this really prevent stale meta attributes from showing up still?
yes. the attributes that are passed to the op are now: attrs supplied when the resource was added + attrs supplied with the op previously it was: attrs passed to the previous op + attrs supplied with the op. this was bad because: a) the previous op could have included notify fields b) the previous op is constructed in the same way, so the crud accumulates _______________________________________________________ Linux-HA-Dev: [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
