were 'consensing' to the S_and/S_or operator.
report(LOG_DEBUG, "expr_eval_immediate: bottom_up and_or: full scan: " PF_EXPR,
PA_EXPR(expr_single));
+ /* It would be nice to pretend that all 'veto' decisions already made in the child
+ * had to set our 'result' to the correct value. But 'consent' decisions don't set
+ * us and the behaviour of auto-set from the child in 'veto' case may get changed
+ * in the future versions.
+ * So we rather don't depend on it.
+ * This overhead doesn't change altgorithmic complexity anyway.
+ */
result_parent = consent;
for (expr_child = expr_parent->u.and_or.child_first; expr_child; expr_child = expr_child->next)
{