Both ExecMakeFunctionResultSet() and evaluation of simple expressions
need to be done in the per-tuple memory context, not per-query, else
we leak data until end of query. This is a consideration that was
missed while refactoring code in the ProjectSet patch (note that in
pre-v10, ExecMakeFunctionResult is called in the per-tuple context).
Per bug #14843 from Ben M. Diagnosed independently by Andres and myself.
Discussion: https://postgr.es/m/
20171005230321.28561.15927@wrigleys.postgresql.org
* function itself. The argument expressions may not contain set-returning
* functions (the planner is supposed to have separated evaluation for those).
*
+ * This should be called in a short-lived (per-tuple) context.
+ *
* This is used by nodeProjectSet.c.
*/
Datum
{
TupleTableSlot *resultSlot = node->ps.ps_ResultTupleSlot;
ExprContext *econtext = node->ps.ps_ExprContext;
+ MemoryContext oldcontext;
bool hassrf PG_USED_FOR_ASSERTS_ONLY;
bool hasresult;
int argno;
ExecClearTuple(resultSlot);
+ /* Call SRFs, as well as plain expressions, in per-tuple context */
+ oldcontext = MemoryContextSwitchTo(econtext->ecxt_per_tuple_memory);
+
/*
* Assume no further tuples are produced unless an ExprMultipleResult is
* encountered from a set returning function.
}
}
+ MemoryContextSwitchTo(oldcontext);
+
/* ProjectSet should not be used if there's no SRFs */
Assert(hassrf);