Obtain table locks as soon as practical during pg_dump.

For some reason, when we (I) added table lock acquisition to pg_dump,
we didn't think about making it happen as soon as possible after the
start of the transaction.  What with subsequent additions, there was
actually quite a lot going on before we got around to that; which sort
of defeats the purpose.  Rearrange the order of calls in dumpSchema()
to close the risk window as much as we easily can.  Back-patch to all
supported branches.

src/bin/pg_dump/common.c

@@ -105,6 +105,17 @@ getSchemaData(int *numTablesPtr)
 		write_msg(NULL, "reading schemas\n");
 	getNamespaces(&numNamespaces);
 
+	/*
+	 * getTables should be done as soon as possible, so as to minimize the
+	 * window between starting our transaction and acquiring per-table locks.
+	 * However, we have to do getNamespaces first because the tables get
+	 * linked to their containing namespaces during getTables.
+	 */
+	if (g_verbose)
+		write_msg(NULL, "reading user-defined tables\n");
+	tblinfo = getTables(&numTables);
+	tblinfoindex = buildIndexArray(tblinfo, numTables, sizeof(TableInfo));
+
 	if (g_verbose)
 		write_msg(NULL, "reading extensions\n");
 	extinfo = getExtensions(&numExtensions);
@@ -183,11 +194,6 @@ getSchemaData(int *numTablesPtr)
 		write_msg(NULL, "reading type casts\n");
 	getCasts(&numCasts);
 
-	if (g_verbose)
-		write_msg(NULL, "reading user-defined tables\n");
-	tblinfo = getTables(&numTables);
-	tblinfoindex = buildIndexArray(tblinfo, numTables, sizeof(TableInfo));
-
 	if (g_verbose)
 		write_msg(NULL, "reading table inheritance information\n");
 	inhinfo = getInherits(&numInherits);