Ensure that whole-row junk Vars are always of composite type.

The EvalPlanQual machinery assumes that whole-row Vars generated for the
outputs of non-table RTEs will be of composite types.  However, for the
case where the RTE is a function call returning a scalar type, we were
doing the wrong thing, as a result of sharing code with a parser case
where the function's scalar output is wanted.  (Or at least, that's what
that case has done historically; it does seem a bit inconsistent.)

To fix, extend makeWholeRowVar's API so that it can support both use-cases.
This fixes Belinda Cussen's report of crashes during concurrent execution
of UPDATEs involving joins to the result of UNNEST() --- in READ COMMITTED
mode, we'd run the EvalPlanQual machinery after a conflicting row update
commits, and it was expecting to get a HeapTuple not a scalar datum from
the "wholerowN" variable referencing the function RTE.

Back-patch to 9.0 where the current EvalPlanQual implementation appeared.

In 9.1 and up, this patch also fixes failure to attach the correct
collation to the Var generated for a scalar-result case.  An example:
regression=# select upper(x.*) from textcat('ab', 'cd') x;
ERROR:  could not determine which collation to use for upper() function

1 files changed, 9 insertions, 2 deletions

diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 79328c99797..75236c76a1f 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -2059,8 +2059,15 @@ transformWholeRowRef(ParseState *pstate, RangeTblEntry *rte, int location)
 	/* Find the RTE's rangetable location */
 	vnum = RTERangeTablePosn(pstate, rte, &sublevels_up);
 
-	/* Build the appropriate referencing node */
-	result = makeWholeRowVar(rte, vnum, sublevels_up);
+	/*
+	 * Build the appropriate referencing node.  Note that if the RTE is a
+	 * function returning scalar, we create just a plain reference to the
+	 * function value, not a composite containing a single column.  This is
+	 * pretty inconsistent at first sight, but it's what we've done
+	 * historically.  One argument for it is that "rel" and "rel.*" mean the
+	 * same thing for composite relations, so why not for scalar functions...
+	 */
+	result = makeWholeRowVar(rte, vnum, sublevels_up, true);
 
 	/* location is not filled in by makeWholeRowVar */
 	result->location = location;