Fix Assert failure in PushOverrideSearchPath when trying to restore a search

path that specifies useTemp, but there is no active temp schema in the
current session.  (This can happen if the path was saved during a transaction
that created a temp schema and was later rolled back.)  For existing callers
it's sufficient to ignore the useTemp flag in this case, though we might
later want to offer an option to create a fresh temp schema.  So far as I can
tell this is just an Assert failure: in a non-assert build, the code would
push a zero onto the new search path, which is useless but not very harmful.
Per bug report from Heikki.

Back-patch to 8.3; prior versions don't have this code.

2 files changed, 16 insertions, 6 deletions

diff --git a/src/backend/catalog/namespace.c b/src/backend/catalog/namespace.c
index 4a5363c66f9..71ec8f8250a 100644
--- a/src/backend/catalog/namespace.c
+++ b/src/backend/catalog/namespace.c
@@ -13,7 +13,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/catalog/namespace.c,v 1.127 2010/08/05 15:25:35 rhaas Exp $
+ *	  $PostgreSQL: pgsql/src/backend/catalog/namespace.c,v 1.128 2010/08/13 16:27:11 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -2667,6 +2667,17 @@ GetOverrideSearchPath(MemoryContext context)
  *
  * We allow nested overrides, hence the push/pop terminology.  The GUC
  * search_path variable is ignored while an override is active.
+ *
+ * It's possible that newpath->useTemp is set but there is no longer any
+ * active temp namespace, if the path was saved during a transaction that
+ * created a temp namespace and was later rolled back.  In that case we just
+ * ignore useTemp.  A plausible alternative would be to create a new temp
+ * namespace, but for existing callers that's not necessary because an empty
+ * temp namespace wouldn't affect their results anyway.
+ *
+ * It's also worth noting that other schemas listed in newpath might not
+ * exist anymore either.  We don't worry about this because OIDs that match
+ * no existing namespace will simply not produce any hits during searches.
  */
 void
 PushOverrideSearchPath(OverrideSearchPath *newpath)
@@ -2700,11 +2711,8 @@ PushOverrideSearchPath(OverrideSearchPath *newpath)
 	if (newpath->addCatalog)
 		oidlist = lcons_oid(PG_CATALOG_NAMESPACE, oidlist);
 
-	if (newpath->addTemp)
-	{
-		Assert(OidIsValid(myTempNamespace));
+	if (newpath->addTemp && OidIsValid(myTempNamespace))
 		oidlist = lcons_oid(myTempNamespace, oidlist);
-	}
 
 	/*
 	 * Build the new stack entry, then insert it at the head of the list.
diff --git a/src/backend/utils/cache/plancache.c b/src/backend/utils/cache/plancache.c
index 95f010f682d..339c8c2c855 100644
--- a/src/backend/utils/cache/plancache.c
+++ b/src/backend/utils/cache/plancache.c
@@ -35,7 +35,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/utils/cache/plancache.c,v 1.35 2010/02/26 02:01:11 momjian Exp $
+ *	  $PostgreSQL: pgsql/src/backend/utils/cache/plancache.c,v 1.36 2010/08/13 16:27:11 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -513,6 +513,8 @@ RevalidateCachedPlan(CachedPlanSource *plansource, bool useResOwner)
 
 		/*
 		 * Restore the search_path that was in use when the plan was made.
+		 * See comments for PushOverrideSearchPath about limitations of this.
+		 *
 		 * (XXX is there anything else we really need to restore?)
 		 */
 		PushOverrideSearchPath(plansource->search_path);