Update docs & tests to reflect that unassigned OLD/NEW are now NULL.

For a long time, plpgsql has allowed trigger functions to parse
references to OLD and NEW even if the current trigger event type didn't
assign a value to one or the other variable; but actually executing such
a reference would fail.  The v11 changes to use "expanded records" for
DTYPE_REC variables changed the behavior so that the unassigned variable
now reads as a null composite value.  While this behavioral change was
more or less unintentional, it seems that leaving it like this is better
than adding code and complexity to be bug-compatible with the old way.
The change doesn't break any code that worked before, and it eliminates
a gotcha that often required extra code to work around.

Hence, update the docs to say that these variables are "null" not
"unassigned" when not relevant to the event type.  And add a regression
test covering the behavior, so that we'll notice if we ever break it
again.

Per report from Kristjan Tammekivi.

Discussion: https://postgr.es/m/CAABK7uL-uC9ZxKBXzo_68pKt7cECfNRv+c35CXZpjq6jCAzYYA@mail.gmail.com

doc/src/sgml/plpgsql.sgml

@@ -3849,7 +3849,7 @@ ASSERT <replaceable class="parameter">condition</replaceable> <optional> , <repl
       <para>
        Data type <type>RECORD</type>; variable holding the new
        database row for <command>INSERT</command>/<command>UPDATE</command> operations in row-level
-       triggers. This variable is unassigned in statement-level triggers
+       triggers. This variable is null in statement-level triggers
        and for <command>DELETE</command> operations.
       </para>
      </listitem>
@@ -3861,7 +3861,7 @@ ASSERT <replaceable class="parameter">condition</replaceable> <optional> , <repl
       <para>
        Data type <type>RECORD</type>; variable holding the old
        database row for <command>UPDATE</command>/<command>DELETE</command> operations in row-level
-       triggers. This variable is unassigned in statement-level triggers
+       triggers. This variable is null in statement-level triggers
        and for <command>INSERT</command> operations.
       </para>
      </listitem>

doc/src/sgml/release-11.sgml

@@ -1033,6 +1033,23 @@ Branch: REL9_3_STABLE [84261eb10] 2018-10-19 17:02:26 -0400
        </para>
       </listitem>
 
+      <listitem>
+<!--
+2018-02-13 [4b93f5799] Make plpgsql use its DTYPE_REC code paths for composite-
+-->
+
+       <para>
+        In PL/pgSQL trigger functions, the <varname>OLD</varname>
+        and <varname>NEW</varname> variables now read as NULL when not
+        assigned (Tom Lane)
+       </para>
+
+       <para>
+        Previously, references to these variables could be parsed but not
+        executed.
+       </para>
+      </listitem>
+
    </itemizedlist>
 
   </sect2>
@@ -2574,7 +2591,6 @@ same commits as above
       <listitem>
 <!--
 2018-02-13 [4b93f5799] Make plpgsql use its DTYPE_REC code paths for composite-
-
 -->
 
        <para>

src/pl/plpgsql/src/Makefile

@@ -27,7 +27,7 @@ DATA = plpgsql.control plpgsql--1.0.sql plpgsql--unpackaged--1.0.sql
 REGRESS_OPTS = --dbname=$(PL_TESTDB)
 
 REGRESS = plpgsql_call plpgsql_control plpgsql_domain plpgsql_record \
-	plpgsql_cache plpgsql_transaction plpgsql_varprops
+	plpgsql_cache plpgsql_transaction plpgsql_trigger plpgsql_varprops
 
 GEN_KEYWORDLIST = $(top_srcdir)/src/tools/gen_keywordlist.pl
 

src/pl/plpgsql/src/expected/plpgsql_trigger.out

+-- Simple test to verify accessibility of the OLD and NEW trigger variables
+create table testtr (a int, b text);
+create function testtr_trigger() returns trigger language plpgsql as
+$$begin
+  raise notice 'tg_op = %', tg_op;
+  raise notice 'old(%) = %', old.a, row(old.*);
+  raise notice 'new(%) = %', new.a, row(new.*);
+  if (tg_op = 'DELETE') then
+    return old;
+  else
+    return new;
+  end if;
+end$$;
+create trigger testtr_trigger before insert or delete or update on testtr
+  for each row execute function testtr_trigger();
+insert into testtr values (1, 'one'), (2, 'two');
+NOTICE:  tg_op = INSERT
+NOTICE:  old(<NULL>) = (,)
+NOTICE:  new(1) = (1,one)
+NOTICE:  tg_op = INSERT
+NOTICE:  old(<NULL>) = (,)
+NOTICE:  new(2) = (2,two)
+update testtr set a = a + 1;
+NOTICE:  tg_op = UPDATE
+NOTICE:  old(1) = (1,one)
+NOTICE:  new(2) = (2,one)
+NOTICE:  tg_op = UPDATE
+NOTICE:  old(2) = (2,two)
+NOTICE:  new(3) = (3,two)
+delete from testtr;
+NOTICE:  tg_op = DELETE
+NOTICE:  old(2) = (2,one)
+NOTICE:  new(<NULL>) = (,)
+NOTICE:  tg_op = DELETE
+NOTICE:  old(3) = (3,two)
+NOTICE:  new(<NULL>) = (,)

src/pl/plpgsql/src/pl_exec.c

@@ -891,11 +891,12 @@ plpgsql_exec_trigger(PLpgSQL_function *func,
 	/*
 	 * Put the OLD and NEW tuples into record variables
 	 *
-	 * We make the tupdescs available in both records even though only one may
-	 * have a value.  This allows parsing of record references to succeed in
-	 * functions that are used for multiple trigger types.  For example, we
-	 * might have a test like "if (TG_OP = 'INSERT' and NEW.foo = 'xyz')",
-	 * which should parse regardless of the current trigger type.
+	 * We set up expanded records for both variables even though only one may
+	 * have a value.  This allows record references to succeed in functions
+	 * that are used for multiple trigger types.  For example, we might have a
+	 * test like "if (TG_OP = 'INSERT' and NEW.foo = 'xyz')", which should
+	 * work regardless of the current trigger type.  If a value is actually
+	 * fetched from an unsupplied tuple, it will read as NULL.
 	 */
 	tupdesc = RelationGetDescr(trigdata->tg_relation);
 

src/pl/plpgsql/src/sql/plpgsql_trigger.sql

+-- Simple test to verify accessibility of the OLD and NEW trigger variables
+
+create table testtr (a int, b text);
+
+create function testtr_trigger() returns trigger language plpgsql as
+$$begin
+  raise notice 'tg_op = %', tg_op;
+  raise notice 'old(%) = %', old.a, row(old.*);
+  raise notice 'new(%) = %', new.a, row(new.*);
+  if (tg_op = 'DELETE') then
+    return old;
+  else
+    return new;
+  end if;
+end$$;
+
+create trigger testtr_trigger before insert or delete or update on testtr
+  for each row execute function testtr_trigger();
+
+insert into testtr values (1, 'one'), (2, 'two');
+
+update testtr set a = a + 1;
+
+delete from testtr;