Doc: improve description of plpgsql's FETCH and MOVE commands.

We were not being clear about which variants of the "direction"
clause are permitted in MOVE.  Also, the text seemed to be
written with only the FETCH/MOVE NEXT case in mind, so it
didn't apply very well to other variants.

Also, document that "MOVE count IN cursor" only works if count
is a constant.  This is not the whole truth, because some other
cases such as a parenthesized expression will also work, but
we want to push people to use "MOVE FORWARD count" instead.
The constant case is enough to cover what we allow in plain SQL,
and that seems sufficient to claim support for.

Update a comment in pl_gram.y claiming that we don't document
that point.

Per gripe from Philipp Salvisberg.

Discussion: https://postgr.es/m/172155553388.702.7932496598218792085@wrigleys.postgresql.org

Author: tglsfdc

doc/src/sgml/plpgsql.sgml

@@ -3409,13 +3409,16 @@ FETCH <optional> <replaceable>direction</replaceable> { FROM | IN } </optional>
 </synopsis>
 
     <para>
-     <command>FETCH</command> retrieves the next row from the
+     <command>FETCH</command> retrieves the next row (in the indicated
+     direction) from the
      cursor into a target, which might be a row variable, a record
      variable, or a comma-separated list of simple variables, just like
-     <command>SELECT INTO</command>.  If there is no next row, the
+     <command>SELECT INTO</command>.  If there is no suitable row, the
      target is set to NULL(s).  As with <command>SELECT
      INTO</command>, the special variable <literal>FOUND</literal> can
-     be checked to see whether a row was obtained or not.
+     be checked to see whether a row was obtained or not.  If no row is
+     obtained, the cursor is positioned after the last row or before the
+     first row, depending on the movement direction.
     </para>
 
     <para>
@@ -3467,11 +3470,25 @@ MOVE <optional> <replaceable>direction</replaceable> { FROM | IN } </optional> <
 
     <para>
      <command>MOVE</command> repositions a cursor without retrieving
-     any data. <command>MOVE</command> works exactly like the
+     any data. <command>MOVE</command> works like the
      <command>FETCH</command> command, except it only repositions the
-     cursor and does not return the row moved to. As with <command>SELECT
+     cursor and does not return the row moved to.
+     The <replaceable>direction</replaceable> clause can be any of the
+     variants allowed in the SQL <xref linkend="sql-fetch"/>
+     command, including those that can fetch more than one row;
+     the cursor is positioned to the last such row.
+     (However, the case in which the <replaceable>direction</replaceable>
+     clause is simply a <replaceable>count</replaceable> expression with
+     no key word is deprecated in <application>PL/pgSQL</application>.
+     That syntax is ambiguous with the case where
+     the <replaceable>direction</replaceable> clause is omitted
+     altogether, and hence it may fail if
+     the <replaceable>count</replaceable> is not a constant.)
+     As with <command>SELECT
      INTO</command>, the special variable <literal>FOUND</literal> can
-     be checked to see whether there was a next row to move to.
+     be checked to see whether there was a row to move to.  If there is no
+     such row, the cursor is positioned after the last row or before the
+     first row, depending on the movement direction.
     </para>
 
     <para>

src/pl/plpgsql/src/pl_gram.y

@@ -3219,11 +3219,11 @@ read_fetch_direction(void)
 	{
 		/*
 		 * Assume it's a count expression with no preceding keyword.
-		 * Note: we allow this syntax because core SQL does, but we don't
-		 * document it because of the ambiguity with the omitted-direction
-		 * case.  For instance, "MOVE n IN c" will fail if n is a variable.
-		 * Perhaps this can be improved someday, but it's hardly worth a
-		 * lot of work.
+		 * Note: we allow this syntax because core SQL does, but it's
+		 * ambiguous with the case of an omitted direction clause; for
+		 * instance, "MOVE n IN c" will fail if n is a variable, because the
+		 * preceding else-arm will trigger.  Perhaps this can be improved
+		 * someday, but it hardly seems worth a lot of work.
 		 */
 		plpgsql_push_back_token(tok);
 		fetch->expr = read_sql_expression2(K_FROM, K_IN,