Fix regex, LIKE, and some other second-rank text-manipulation functions

to not cause needless copying of text datums that have 1-byte headers.
Greg Stark, in response to performance gripe from Guillaume Smet and
ITAGAKI Takahiro.

src/backend/access/hash/hashfunc.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/access/hash/hashfunc.c,v 1.52 2007/06/01 15:33:18 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/access/hash/hashfunc.c,v 1.53 2007/09/21 22:52:52 tgl Exp $
  *
  * NOTES
  *	  These functions are stored in pg_amproc.	For each operator class
@@ -157,7 +157,7 @@ hashname(PG_FUNCTION_ARGS)
 Datum
 hashtext(PG_FUNCTION_ARGS)
 {
-	text	   *key = PG_GETARG_TEXT_P(0);
+	text	   *key = PG_GETARG_TEXT_PP(0);
 	Datum		result;
 
 	/*
@@ -165,8 +165,8 @@ hashtext(PG_FUNCTION_ARGS)
 	 * it as a separate function in case we someday want to do something
 	 * different in non-C locales.	(See also hashbpchar, if so.)
 	 */
-	result = hash_any((unsigned char *) VARDATA(key),
-					  VARSIZE(key) - VARHDRSZ);
+	result = hash_any((unsigned char *) VARDATA_ANY(key),
+					  VARSIZE_ANY_EXHDR(key));
 
 	/* Avoid leaking memory for toasted inputs */
 	PG_FREE_IF_COPY(key, 0);
@@ -181,11 +181,11 @@ hashtext(PG_FUNCTION_ARGS)
 Datum
 hashvarlena(PG_FUNCTION_ARGS)
 {
-	struct varlena *key = PG_GETARG_VARLENA_P(0);
+	struct varlena *key = PG_GETARG_VARLENA_PP(0);
 	Datum		result;
 
-	result = hash_any((unsigned char *) VARDATA(key),
-					  VARSIZE(key) - VARHDRSZ);
+	result = hash_any((unsigned char *) VARDATA_ANY(key),
+					  VARSIZE_ANY_EXHDR(key));
 
 	/* Avoid leaking memory for toasted inputs */
 	PG_FREE_IF_COPY(key, 0);

src/backend/utils/adt/like.c

@@ -11,7 +11,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	$PostgreSQL: pgsql/src/backend/utils/adt/like.c,v 1.69 2007/06/02 02:03:42 adunstan Exp $
+ *	$PostgreSQL: pgsql/src/backend/utils/adt/like.c,v 1.70 2007/09/21 22:52:52 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -135,6 +135,7 @@ Generic_Text_IC_like(text *str, text *pat)
 	/* Force inputs to lower case to achieve case insensitivity */
 	str = DatumGetTextP(DirectFunctionCall1(lower, PointerGetDatum(str)));
 	pat = DatumGetTextP(DirectFunctionCall1(lower, PointerGetDatum(pat)));
+	/* lower's result is never packed, so OK to use old macros here */
 	s = VARDATA(str);
 	slen = (VARSIZE(str) - VARHDRSZ);
 	p = VARDATA(pat);
@@ -151,7 +152,7 @@ Datum
 namelike(PG_FUNCTION_ARGS)
 {
 	Name		str = PG_GETARG_NAME(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
@@ -160,8 +161,8 @@ namelike(PG_FUNCTION_ARGS)
 
 	s = NameStr(*str);
 	slen = strlen(s);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (GenericMatchText(s, slen, p, plen) == LIKE_TRUE);
 
@@ -172,7 +173,7 @@ Datum
 namenlike(PG_FUNCTION_ARGS)
 {
 	Name		str = PG_GETARG_NAME(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
@@ -181,8 +182,8 @@ namenlike(PG_FUNCTION_ARGS)
 
 	s = NameStr(*str);
 	slen = strlen(s);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (GenericMatchText(s, slen, p, plen) != LIKE_TRUE);
 
@@ -192,18 +193,18 @@ namenlike(PG_FUNCTION_ARGS)
 Datum
 textlike(PG_FUNCTION_ARGS)
 {
-	text	   *str = PG_GETARG_TEXT_P(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *str = PG_GETARG_TEXT_PP(0);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
 	int			slen,
 				plen;
 
-	s = VARDATA(str);
-	slen = (VARSIZE(str) - VARHDRSZ);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	s = VARDATA_ANY(str);
+	slen = VARSIZE_ANY_EXHDR(str);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (GenericMatchText(s, slen, p, plen) == LIKE_TRUE);
 
@@ -213,18 +214,18 @@ textlike(PG_FUNCTION_ARGS)
 Datum
 textnlike(PG_FUNCTION_ARGS)
 {
-	text	   *str = PG_GETARG_TEXT_P(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *str = PG_GETARG_TEXT_PP(0);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
 	int			slen,
 				plen;
 
-	s = VARDATA(str);
-	slen = (VARSIZE(str) - VARHDRSZ);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	s = VARDATA_ANY(str);
+	slen = VARSIZE_ANY_EXHDR(str);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (GenericMatchText(s, slen, p, plen) != LIKE_TRUE);
 
@@ -234,18 +235,18 @@ textnlike(PG_FUNCTION_ARGS)
 Datum
 bytealike(PG_FUNCTION_ARGS)
 {
-	bytea	   *str = PG_GETARG_BYTEA_P(0);
-	bytea	   *pat = PG_GETARG_BYTEA_P(1);
+	bytea	   *str = PG_GETARG_BYTEA_PP(0);
+	bytea	   *pat = PG_GETARG_BYTEA_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
 	int			slen,
 				plen;
 
-	s = VARDATA(str);
-	slen = (VARSIZE(str) - VARHDRSZ);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	s = VARDATA_ANY(str);
+	slen = VARSIZE_ANY_EXHDR(str);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (SB_MatchText(s, slen, p, plen) == LIKE_TRUE);
 
@@ -255,18 +256,18 @@ bytealike(PG_FUNCTION_ARGS)
 Datum
 byteanlike(PG_FUNCTION_ARGS)
 {
-	bytea	   *str = PG_GETARG_BYTEA_P(0);
-	bytea	   *pat = PG_GETARG_BYTEA_P(1);
+	bytea	   *str = PG_GETARG_BYTEA_PP(0);
+	bytea	   *pat = PG_GETARG_BYTEA_PP(1);
 	bool		result;
 	char	   *s,
 			   *p;
 	int			slen,
 				plen;
 
-	s = VARDATA(str);
-	slen = (VARSIZE(str) - VARHDRSZ);
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
+	s = VARDATA_ANY(str);
+	slen = VARSIZE_ANY_EXHDR(str);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
 
 	result = (SB_MatchText(s, slen, p, plen) != LIKE_TRUE);
 
@@ -281,7 +282,7 @@ Datum
 nameiclike(PG_FUNCTION_ARGS)
 {
 	Name		str = PG_GETARG_NAME(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	text	   *strtext;
 
@@ -296,7 +297,7 @@ Datum
 nameicnlike(PG_FUNCTION_ARGS)
 {
 	Name		str = PG_GETARG_NAME(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 	text	   *strtext;
 
@@ -310,8 +311,8 @@ nameicnlike(PG_FUNCTION_ARGS)
 Datum
 texticlike(PG_FUNCTION_ARGS)
 {
-	text	   *str = PG_GETARG_TEXT_P(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *str = PG_GETARG_TEXT_PP(0);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 
 	result = (Generic_Text_IC_like(str, pat) == LIKE_TRUE);
@@ -322,8 +323,8 @@ texticlike(PG_FUNCTION_ARGS)
 Datum
 texticnlike(PG_FUNCTION_ARGS)
 {
-	text	   *str = PG_GETARG_TEXT_P(0);
-	text	   *pat = PG_GETARG_TEXT_P(1);
+	text	   *str = PG_GETARG_TEXT_PP(0);
+	text	   *pat = PG_GETARG_TEXT_PP(1);
 	bool		result;
 
 	result = (Generic_Text_IC_like(str, pat) != LIKE_TRUE);
@@ -338,8 +339,8 @@ texticnlike(PG_FUNCTION_ARGS)
 Datum
 like_escape(PG_FUNCTION_ARGS)
 {
-	text	   *pat = PG_GETARG_TEXT_P(0);
-	text	   *esc = PG_GETARG_TEXT_P(1);
+	text	   *pat = PG_GETARG_TEXT_PP(0);
+	text	   *esc = PG_GETARG_TEXT_PP(1);
 	text	   *result;
 
 	if (pg_database_encoding_max_length() == 1)
@@ -357,8 +358,8 @@ like_escape(PG_FUNCTION_ARGS)
 Datum
 like_escape_bytea(PG_FUNCTION_ARGS)
 {
-	bytea	   *pat = PG_GETARG_BYTEA_P(0);
-	bytea	   *esc = PG_GETARG_BYTEA_P(1);
+	bytea	   *pat = PG_GETARG_BYTEA_PP(0);
+	bytea	   *esc = PG_GETARG_BYTEA_PP(1);
 	bytea	   *result = SB_do_like_escape((text *)pat, (text *)esc);
 
 	PG_RETURN_BYTEA_P((bytea *)result);

src/backend/utils/adt/like_match.c

@@ -17,7 +17,7 @@
  * Copyright (c) 1996-2007, PostgreSQL Global Development Group
  *
  * IDENTIFICATION
- *	$PostgreSQL: pgsql/src/backend/utils/adt/like_match.c,v 1.16 2007/06/02 02:03:42 adunstan Exp $
+ *	$PostgreSQL: pgsql/src/backend/utils/adt/like_match.c,v 1.17 2007/09/21 22:52:52 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -242,10 +242,10 @@ do_like_escape(text *pat, text *esc)
 				elen;
 	bool		afterescape;
 
-	p = VARDATA(pat);
-	plen = (VARSIZE(pat) - VARHDRSZ);
-	e = VARDATA(esc);
-	elen = (VARSIZE(esc) - VARHDRSZ);
+	p = VARDATA_ANY(pat);
+	plen = VARSIZE_ANY_EXHDR(pat);
+	e = VARDATA_ANY(esc);
+	elen = VARSIZE_ANY_EXHDR(esc);
 
 	/*
 	 * Worst-case pattern growth is 2x --- unlikely, but it's hardly worth
@@ -279,14 +279,14 @@ do_like_escape(text *pat, text *esc)
 					 errmsg("invalid escape string"),
 				  errhint("Escape string must be empty or one character.")));
 
-		e = VARDATA(esc);
+		e = VARDATA_ANY(esc);
 
 		/*
 		 * If specified escape is '\', just copy the pattern as-is.
 		 */
 		if (*e == '\\')
 		{
-			memcpy(result, pat, VARSIZE(pat));
+			memcpy(result, pat, VARSIZE_ANY(pat));
 			return result;
 		}
 

src/include/fmgr.h

@@ -11,7 +11,7 @@
  * Portions Copyright (c) 1996-2007, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/fmgr.h,v 1.52 2007/07/25 12:22:52 mha Exp $
+ * $PostgreSQL: pgsql/src/include/fmgr.h,v 1.53 2007/09/21 22:52:52 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -220,6 +220,8 @@ extern struct varlena *pg_detoast_datum_packed(struct varlena * datum);
 #define PG_GETARG_RAW_VARLENA_P(n)	((struct varlena *) PG_GETARG_POINTER(n))
 /* use this if you want the input datum de-toasted: */
 #define PG_GETARG_VARLENA_P(n) PG_DETOAST_DATUM(PG_GETARG_DATUM(n))
+/* and this if you can handle 1-byte-header datums: */
+#define PG_GETARG_VARLENA_PP(n) PG_DETOAST_DATUM_PACKED(PG_GETARG_DATUM(n))
 /* DatumGetFoo macros for varlena types will typically look like this: */
 #define DatumGetByteaP(X)			((bytea *) PG_DETOAST_DATUM(X))
 #define DatumGetByteaPP(X)			((bytea *) PG_DETOAST_DATUM_PACKED(X))