Change the parser to convert SQL "position" and "substring" syntax to

position() and substring() functions, so that it works transparently for
bit types as well.  Alias the text functions appropriately.

Add position() for bit types.

Add new constant node T_BitString that represents literals of the form
B'1001 and pass those to zpbit type.

Author: petere

src/backend/nodes/copyfuncs.c

@@ -15,7 +15,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/nodes/copyfuncs.c,v 1.127 2000/10/26 21:35:47 tgl Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/nodes/copyfuncs.c,v 1.128 2000/10/31 10:22:10 petere Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -2511,6 +2511,7 @@ _copyValue(Value *from)
 			break;
 		case T_Float:
 		case T_String:
+		case T_BitString:
 			newnode->val.str = pstrdup(from->val.str);
 			break;
 		default:
@@ -2703,6 +2704,7 @@ copyObject(void *from)
 		case T_Integer:
 		case T_Float:
 		case T_String:
+		case T_BitString:
 			retval = _copyValue(from);
 			break;
 		case T_List:

src/backend/nodes/equalfuncs.c

@@ -20,7 +20,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/nodes/equalfuncs.c,v 1.77 2000/10/18 16:16:04 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/nodes/equalfuncs.c,v 1.78 2000/10/31 10:22:10 petere Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1719,6 +1719,7 @@ _equalValue(Value *a, Value *b)
 			return a->val.ival == b->val.ival;
 		case T_Float:
 		case T_String:
+		case T_BitString:
 			return strcmp(a->val.str, b->val.str) == 0;
 		default:
 			break;
@@ -1874,6 +1875,7 @@ equal(void *a, void *b)
 		case T_Integer:
 		case T_Float:
 		case T_String:
+		case T_BitString:
 			retval = _equalValue(a, b);
 			break;
 

src/backend/nodes/list.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/nodes/list.c,v 1.35 2000/10/05 19:11:27 tgl Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/nodes/list.c,v 1.36 2000/10/31 10:22:10 petere Exp $
  *
  * NOTES
  *	  XXX a few of the following functions are duplicated to handle
@@ -70,6 +70,23 @@ makeString(char *str)
 	return v;
 }
 
+
+/*
+ *	makeBitString
+ *
+ * Caller is responsible for passing a palloc'd string.
+ */
+Value *
+makeBitString(char *str)
+{
+	Value	   *v = makeNode(Value);
+
+	v->type = T_BitString;
+	v->val.str = str;
+	return v;
+}
+
+
 /*
  *	lcons
  *

src/backend/nodes/outfuncs.c

@@ -6,7 +6,7 @@
  * Portions Copyright (c) 1996-2000, PostgreSQL, Inc
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- *	$Header: /cvsroot/pgsql/src/backend/nodes/outfuncs.c,v 1.129 2000/10/26 21:35:48 tgl Exp $
+ *	$Header: /cvsroot/pgsql/src/backend/nodes/outfuncs.c,v 1.130 2000/10/31 10:22:10 petere Exp $
  *
  * NOTES
  *	  Every (plan) node in POSTGRES has an associated "out" routine which
@@ -20,10 +20,10 @@
  *	  representation plus some other information (string length, etc.)
  *
  */
-#include <ctype.h>
-
 #include "postgres.h"
 
+#include <ctype.h>
+
 #include "access/heapam.h"
 #include "access/htup.h"
 #include "catalog/pg_type.h"
@@ -1352,6 +1352,9 @@ _outValue(StringInfo str, Value *value)
 			_outToken(str, value->val.str);
 			appendStringInfo(str, "\" ");
 			break;
+		case T_BitString:
+			appendStringInfo(str, " B%s ", value->val.str);
+			break;
 		default:
 			elog(NOTICE, "_outValue: don't know how to print type %d ",
 				 value->type);

src/backend/nodes/read.c

@@ -9,19 +9,19 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/nodes/read.c,v 1.23 2000/06/14 18:17:32 petere Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/nodes/read.c,v 1.24 2000/10/31 10:22:10 petere Exp $
  *
  * HISTORY
  *	  AUTHOR			DATE			MAJOR EVENT
  *	  Andrew Yu			Nov 2, 1994		file creation
  *
  *-------------------------------------------------------------------------
  */
+#include "postgres.h"
+
 #include <ctype.h>
 #include <errno.h>
 
-#include "postgres.h"
-
 #include "nodes/pg_list.h"
 #include "nodes/readfuncs.h"
 
@@ -184,7 +184,7 @@ debackslash(char *token, int length)
  * nodeTokenType -
  *	  returns the type of the node token contained in token.
  *	  It returns one of the following valid NodeTags:
- *		T_Integer, T_Float, T_String
+ *		T_Integer, T_Float, T_String, T_BitString
  *	  and some of its own:
  *		RIGHT_PAREN, LEFT_PAREN, PLAN_SYM, AT_SYMBOL, ATOM_TOKEN
  *
@@ -236,6 +236,8 @@ nodeTokenType(char *token, int length)
 		retval = AT_SYMBOL;
 	else if (*token == '\"' && length > 1 && token[length - 1] == '\"')
 		retval = T_String;
+	else if (*token == 'B')
+		retval = T_BitString;
 	else
 		retval = ATOM_TOKEN;
 	return retval;
@@ -346,6 +348,15 @@ nodeRead(bool read_car_only)
 			this_value = (Node *) makeString(debackslash(token + 1, tok_len - 2));
 			make_dotted_pair_cell = true;
 			break;
+		case T_BitString:
+		{
+			char * val = palloc(tok_len);
+			/* skip leading 'B'*/
+			strncpy(val, token + 1, tok_len - 1);
+			val[tok_len - 1] = '\0';
+			this_value = (Node *) makeBitString(val);
+			break;
+		}
 		default:
 			elog(ERROR, "nodeRead: Bad type %d", type);
 			this_value = NULL;	/* keep compiler happy */

src/backend/parser/gram.y

@@ -11,7 +11,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/parser/gram.y,v 2.201 2000/10/29 16:11:33 petere Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/parser/gram.y,v 2.202 2000/10/31 10:22:10 petere Exp $
  *
  * HISTORY
  *	  AUTHOR			DATE			MAJOR EVENT
@@ -353,7 +353,7 @@ static void doNegateFloat(Value *v);
 %token			UNIONJOIN
 
 /* Special keywords, not in the query language - see the "lex" file */
-%token <str>	IDENT, FCONST, SCONST, Op
+%token <str>	IDENT, FCONST, SCONST, BITCONST, Op
 %token <ival>	ICONST, PARAM
 
 /* these are not real. they are here so that they get generated as #define's*/
@@ -1798,6 +1798,10 @@ TriggerFuncArg:  ICONST
 				{
 					$$ = makeString($1);
 				}
+			| BITCONST
+				{
+					$$ = makeString($1);
+				}
 			| ColId
 				{
 					$$ = makeString($1);
@@ -4786,17 +4790,20 @@ c_expr:  attr
 				}
 		| POSITION '(' position_list ')'
 				{
+					/* position(A in B) is converted to position(B, A) */
 					FuncCall *n = makeNode(FuncCall);
-					n->funcname = "strpos";
+					n->funcname = "position";
 					n->args = $3;
 					n->agg_star = FALSE;
 					n->agg_distinct = FALSE;
 					$$ = (Node *)n;
 				}
 		| SUBSTRING '(' substr_list ')'
 				{
+					/* substring(A from B for C) is converted to
+					 * substring(A, B, C) */
 					FuncCall *n = makeNode(FuncCall);
-					n->funcname = "substr";
+					n->funcname = "substring";
 					n->args = $3;
 					n->agg_star = FALSE;
 					n->agg_distinct = FALSE;
@@ -5201,6 +5208,13 @@ AexprConst:  Iconst
 					n->val.val.str = $1;
 					$$ = (Node *)n;
 				}
+		| BITCONST
+				{
+					A_Const *n = makeNode(A_Const);
+					n->val.type = T_BitString;
+					n->val.val.str = $1;
+					$$ = (Node *)n;
+				}
 		/* This rule formerly used Typename,
 		 * but that causes reduce conflicts with subscripted column names.
 		 * Now, separate into ConstTypename and ConstInterval,

src/backend/parser/parse_node.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/parser/parse_node.c,v 1.47 2000/09/29 18:21:36 tgl Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/parser/parse_node.c,v 1.48 2000/10/31 10:22:11 petere Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -31,6 +31,7 @@
 #include "parser/parse_target.h"
 #include "parser/parse_type.h"
 #include "utils/builtins.h"
+#include "utils/varbit.h"
 #include "utils/lsyscache.h"
 #include "utils/syscache.h"
 
@@ -473,6 +474,16 @@ make_const(Value *value)
 			typebyval = false;
 			break;
 
+		case T_BitString:
+			val = DirectFunctionCall3(zpbit_in,
+									  CStringGetDatum(strVal(value)),
+									  ObjectIdGetDatum(InvalidOid),
+									  Int32GetDatum(-1));
+			typeid = ZPBITOID;
+			typelen = -1;
+			typebyval = false;
+			break;
+
 		default:
 			elog(NOTICE, "make_const: unknown type %d", nodeTag(value));
 			/* FALLTHROUGH */

src/backend/parser/scan.l

@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.79 2000/10/30 17:54:16 petere Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.80 2000/10/31 10:22:11 petere Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -93,25 +93,25 @@ static void addlit(char *ytext, int yleng);
  * We use exclusive states for quoted strings, extended comments,
  * and to eliminate parsing troubles for numeric strings.
  * Exclusive states:
- *  <xb> binary numeric string - thomas 1997-11-16
+ *  <xbit> bit string literal
  *  <xc> extended C-style comments - thomas 1997-07-12
  *  <xd> delimited identifiers (double-quoted identifiers) - thomas 1997-10-27
  *  <xh> hexadecimal numeric string - thomas 1997-11-16
  *  <xq> quoted strings - thomas 1997-07-30
  */
 
-%x xb
+%x xbit
 %x xc
 %x xd
 %x xh
 %x xq
 
-/* Binary number
+/* Bit string
  */
-xbstart			[bB]{quote}
-xbstop			{quote}
-xbinside		[^']+
-xbcat			{quote}{whitespace_with_newline}{quote}
+xbitstart		[bB]{quote}
+xbitstop		{quote}
+xbitinside		[^']*
+xbitcat			{quote}{whitespace_with_newline}{quote}
 
 /* Hexadecimal number
  */
@@ -279,30 +279,27 @@ other			.
 
 <xc><<EOF>>		{ elog(ERROR, "Unterminated /* comment"); }
 
-{xbstart}		{
-					BEGIN(xb);
+{xbitstart}		{
+					BEGIN(xbit);
 					startlit();
 				}
-<xb>{xbstop}	{
-					char* endptr;
-
+<xbit>{xbitstop}	{
 					BEGIN(INITIAL);
-					errno = 0;
-					yylval.ival = strtol(literalbuf, &endptr, 2);
-					if (*endptr != '\0' || errno == ERANGE)
-						elog(ERROR, "Bad binary integer input '%s'",
+					if (literalbuf[strspn(literalbuf, "01")] != '\0')
+						elog(ERROR, "invalid bit string input: '%s'",
 							 literalbuf);
-					return ICONST;
+					yylval.str = literalbuf;
+					return BITCONST;
 				}
 <xh>{xhinside}	|
-<xb>{xbinside}	{
+<xbit>{xbitinside}	{
 					addlit(yytext, yyleng);
 				}
 <xh>{xhcat}		|
-<xb>{xbcat}		{
+<xbit>{xbitcat}		{
 					/* ignore */
 				}
-<xb><<EOF>>		{ elog(ERROR, "Unterminated binary integer"); }
+<xbit><<EOF>>		{ elog(ERROR, "unterminated bit string literal"); }
 
 {xhstart}		{
 					BEGIN(xh);