Separate parse-analysis for utility commands out of parser/analyze.c

(which now deals only in optimizable statements), and put that code
into a new file parser/parse_utilcmd.c.  This helps clarify and enforce
the design rule that utility statements shouldn't be processed during
the regular parse analysis phase; all interpretation of their meaning
should happen after they are given to ProcessUtility to execute.
(We need this because we don't retain any locks for a utility statement
that's in a plan cache, nor have any way to detect that it's stale.)

We are also able to simplify the API for parse_analyze() and related
routines, because they will now always return exactly one Query structure.

In passing, fix bug #3403 concerning trying to add a serial column to
an existing temp table (this is largely Heikki's work, but we needed
all that restructuring to make it safe).

src/backend/parser/Makefile

@@ -2,7 +2,7 @@
 #
 # Makefile for parser
 #
-# $PostgreSQL: pgsql/src/backend/parser/Makefile,v 1.44 2006/05/27 17:38:45 tgl Exp $
+# $PostgreSQL: pgsql/src/backend/parser/Makefile,v 1.45 2007/06/23 22:12:51 tgl Exp $
 #
 #-------------------------------------------------------------------------
 
@@ -14,7 +14,7 @@ override CPPFLAGS := -I$(srcdir) $(CPPFLAGS)
 
 OBJS= analyze.o gram.o keywords.o parser.o parse_agg.o parse_clause.o \
       parse_expr.o parse_func.o parse_node.o parse_oper.o parse_relation.o \
-      parse_type.o parse_coerce.o parse_target.o scansup.o
+      parse_type.o parse_coerce.o parse_target.o parse_utilcmd.o scansup.o
 
 FLEXFLAGS = -CF
 

src/backend/parser/README

@@ -1,20 +1,21 @@
 This directory does more than tokenize and parse SQL queries.  It also
-creates Query structures for the various complex queries that is passed
+creates Query structures for the various complex queries that are passed
 to the optimizer and then executor.
 
 parser.c	things start here
 scan.l		break query into tokens
-scansup.c	handle escapes in input
+scansup.c	handle escapes in input strings
 keywords.c	turn keywords into specific tokens
 gram.y		parse the tokens and fill query-type-specific structures
-analyze.c	handle post-parse processing for each query type
+analyze.c	top level of parse analysis for optimizable queries
 parse_clause.c	handle clauses like WHERE, ORDER BY, GROUP BY, ...
-parse_coerce.c	used for coercing expressions of different types
+parse_coerce.c	handle coercing expressions to different types
 parse_expr.c	handle expressions like col, col + 3, x = 3 or x = 4
-parse_oper.c	handle operations in expressions
+parse_oper.c	handle operators in expressions
 parse_agg.c	handle aggregates, like SUM(col1),  AVG(col2), ...
 parse_func.c	handle functions, table.column and column identifiers
 parse_node.c	create nodes for various structures
 parse_target.c	handle the result list of the query
 parse_relation.c support routines for tables and column handling
 parse_type.c	support routines for type handling
+parse_utilcmd.c	parse analysis for utility commands (done at execution time)

src/backend/parser/gram.y

@@ -11,7 +11,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/parser/gram.y,v 2.595 2007/06/18 21:40:57 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/parser/gram.y,v 2.596 2007/06/23 22:12:51 tgl Exp $
  *
  * HISTORY
  *	  AUTHOR			DATE			MAJOR EVENT
@@ -33,7 +33,7 @@
  *			SET SQL_inheritance TO off; SELECT * FROM foo;
  *	  because the entire string is parsed by gram.y before the SET gets
  *	  executed.  Anything that depends on the database or changeable state
- *	  should be handled inside parse_analyze() so that it happens at the
+ *	  should be handled during parse analysis so that it happens at the
  *	  right time not the wrong time.  The handling of SQL_inheritance is
  *	  a good example.
  *
@@ -2093,9 +2093,10 @@ ColConstraintElem:
  * ConstraintAttr represents constraint attributes, which we parse as if
  * they were independent constraint clauses, in order to avoid shift/reduce
  * conflicts (since NOT might start either an independent NOT NULL clause
- * or an attribute).  analyze.c is responsible for attaching the attribute
- * information to the preceding "real" constraint node, and for complaining
- * if attribute clauses appear in the wrong place or wrong combinations.
+ * or an attribute).  parse_utilcmd.c is responsible for attaching the
+ * attribute information to the preceding "real" constraint node, and for
+ * complaining if attribute clauses appear in the wrong place or wrong
+ * combinations.
  *
  * See also ConstraintAttributeSpec, which can be used in places where
  * there is no parsing conflict.

src/backend/parser/parse_clause.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/parser/parse_clause.c,v 1.165 2007/04/27 22:05:48 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/parser/parse_clause.c,v 1.166 2007/06/23 22:12:51 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -152,8 +152,8 @@ setTargetTable(ParseState *pstate, RangeVar *relation,
 	 * Open target rel and grab suitable lock (which we will hold till end of
 	 * transaction).
 	 *
-	 * analyze.c will eventually do the corresponding heap_close(), but *not*
-	 * release the lock.
+	 * free_parsestate() will eventually do the corresponding
+	 * heap_close(), but *not* release the lock.
 	 */
 	pstate->p_target_relation = heap_openrv(relation, RowExclusiveLock);
 
@@ -193,7 +193,7 @@ setTargetTable(ParseState *pstate, RangeVar *relation,
  * Simplify InhOption (yes/no/default) into boolean yes/no.
  *
  * The reason we do things this way is that we don't want to examine the
- * SQL_inheritance option flag until parse_analyze is run.	Otherwise,
+ * SQL_inheritance option flag until parse_analyze() is run.	Otherwise,
  * we'd do the wrong thing with query strings that intermix SET commands
  * with queries.
  */
@@ -417,7 +417,6 @@ transformTableEntry(ParseState *pstate, RangeVar *r)
 static RangeTblEntry *
 transformRangeSubselect(ParseState *pstate, RangeSubselect *r)
 {
-	List	   *parsetrees;
 	Query	   *query;
 	RangeTblEntry *rte;
 
@@ -434,19 +433,12 @@ transformRangeSubselect(ParseState *pstate, RangeSubselect *r)
 	/*
 	 * Analyze and transform the subquery.
 	 */
-	parsetrees = parse_sub_analyze(r->subquery, pstate);
+	query = parse_sub_analyze(r->subquery, pstate);
 
 	/*
-	 * Check that we got something reasonable.	Most of these conditions are
-	 * probably impossible given restrictions of the grammar, but check 'em
-	 * anyway.
+	 * Check that we got something reasonable.	Many of these conditions are
+	 * impossible given restrictions of the grammar, but check 'em anyway.
 	 */
-	if (list_length(parsetrees) != 1)
-		elog(ERROR, "unexpected parse analysis result for subquery in FROM");
-	query = (Query *) linitial(parsetrees);
-	if (query == NULL || !IsA(query, Query))
-		elog(ERROR, "unexpected parse analysis result for subquery in FROM");
-
 	if (query->commandType != CMD_SELECT ||
 		query->utilityStmt != NULL)
 		elog(ERROR, "expected SELECT query from subquery in FROM");

src/backend/parser/parse_expr.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/parser/parse_expr.c,v 1.220 2007/06/11 22:22:42 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/parser/parse_expr.c,v 1.221 2007/06/23 22:12:51 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1120,19 +1120,15 @@ transformCaseExpr(ParseState *pstate, CaseExpr *c)
 static Node *
 transformSubLink(ParseState *pstate, SubLink *sublink)
 {
-	List	   *qtrees;
-	Query	   *qtree;
 	Node	   *result = (Node *) sublink;
+	Query	   *qtree;
 
 	/* If we already transformed this node, do nothing */
 	if (IsA(sublink->subselect, Query))
 		return result;
 
 	pstate->p_hasSubLinks = true;
-	qtrees = parse_sub_analyze(sublink->subselect, pstate);
-	if (list_length(qtrees) != 1)
-		elog(ERROR, "bad query in sub-select");
-	qtree = (Query *) linitial(qtrees);
+	qtree = parse_sub_analyze(sublink->subselect, pstate);
 	if (qtree->commandType != CMD_SELECT ||
 		qtree->utilityStmt != NULL ||
 		qtree->intoClause != NULL)

src/backend/parser/parse_node.c

@@ -8,12 +8,13 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/parser/parse_node.c,v 1.97 2007/03/17 00:11:04 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/parser/parse_node.c,v 1.98 2007/06/23 22:12:51 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
 #include "postgres.h"
 
+#include "access/heapam.h"
 #include "catalog/pg_type.h"
 #include "mb/pg_wchar.h"
 #include "nodes/makefuncs.h"
@@ -27,9 +28,11 @@
 #include "utils/varbit.h"
 
 
-/* make_parsestate()
- * Allocate and initialize a new ParseState.
- * The CALLER is responsible for freeing the ParseState* returned.
+/*
+ * make_parsestate
+ *		Allocate and initialize a new ParseState.
+ *
+ * Caller should eventually release the ParseState via free_parsestate().
  */
 ParseState *
 make_parsestate(ParseState *parentParseState)
@@ -52,6 +55,30 @@ make_parsestate(ParseState *parentParseState)
 	return pstate;
 }
 
+/*
+ * free_parsestate
+ *		Release a ParseState and any subsidiary resources.
+ */
+void
+free_parsestate(ParseState *pstate)
+{
+	/*
+	 * Check that we did not produce too many resnos; at the very least we
+	 * cannot allow more than 2^16, since that would exceed the range of a
+	 * AttrNumber. It seems safest to use MaxTupleAttributeNumber.
+	 */
+	if (pstate->p_next_resno - 1 > MaxTupleAttributeNumber)
+		ereport(ERROR,
+				(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
+				 errmsg("target lists can have at most %d entries",
+						MaxTupleAttributeNumber)));
+
+	if (pstate->p_target_relation != NULL)
+		heap_close(pstate->p_target_relation, NoLock);
+
+	pfree(pstate);
+}
+
 
 /*
  * parser_errposition