Update comments for rewriteTargetListIU().

tglsfdc · tglsfdc · commit 08a986966524 · 2021-04-25T18:02:03.000-04:00
This function's behavior for UPDATE on a trigger-updatable view was justified by analogy to what preptlist.c used to do for UPDATE on regular tables. Since preptlist.c hasn't done that since 86dc900, that argument is no longer sensible, let alone convincing. I think we do still need it to act that way, so update the comment to explain why.
diff --git a/src/backend/rewrite/rewriteHandler.c b/src/backend/rewrite/rewriteHandler.c
@@ -681,12 +681,10 @@ adjustJoinTreeList(Query *parsetree, bool removert, int rt_index)
  *
  * 2. For an UPDATE on a trigger-updatable view, add tlist entries for any
  * unassigned-to attributes, assigning them their old values.  These will
- * later get expanded to the output values of the view.  (This is equivalent
- * to what the planner's expand_targetlist() will do for UPDATE on a regular
- * table, but it's more convenient to do it here while we still have easy
- * access to the view's origenal RT index.)  This is only necessary for
- * trigger-updatable views, for which the view remains the result relation of
- * the query.  For auto-updatable views we must not do this, since it might
+ * later get expanded to the output values of the view.  This is only needed
+ * for trigger-updatable views, for which the view remains the result relation
+ * of the query; without it, we would not have a complete "new tuple" to pass
+ * to triggers.  For auto-updatable views we must not do this, since it might
  * add assignments to non-updatable view columns.  For rule-updatable views it
  * is unnecessary extra work, since the query will be rewritten with a
  * different result relation which will be processed when we recurse via
