Obtain correct first character in pattern for ESCAPE in LIKE (#649) (#658)

Whenever there is ESCAPE in LIKE operator and the pattern is either infix or postfix, Babelfish is unable to handle the wildcard properly.

This commit fixes that by ignoring the ESCAPE character in the pattern and finding the correct pattern.
Extension PR: babelfish-for-postgresql/babelfish_extensions#4251

Task: BABEL-6180

cherry picked from commit 3aa4957

Signed-off-by: Shameem Ahmed shmeeh@amazon.com

Author: ahmed-shameem

src/backend/utils/adt/like_match.c

@@ -77,6 +77,7 @@
 #endif
 
 #define BBF_ESC_CHAR_REPLC "\357\277\277"
+#define BBF_ESC_CHAR_REPLC_LEN (sizeof(BBF_ESC_CHAR_REPLC) - 1)
 
 static int
 MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
@@ -100,12 +101,12 @@ MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
 	{
 		/* Default escape '\\' is disabled in babelfish */
 		if ((*p == '\\' && sql_dialect == SQL_DIALECT_PG)||
-			(strncmp(p, BBF_ESC_CHAR_REPLC, strlen(BBF_ESC_CHAR_REPLC)) == 0 && sql_dialect == SQL_DIALECT_TSQL))
+			(sql_dialect == SQL_DIALECT_TSQL && strncmp(p, BBF_ESC_CHAR_REPLC, BBF_ESC_CHAR_REPLC_LEN) == 0))
 		{
 			/* Next pattern byte must match literally, whatever it is */
 			if (sql_dialect == SQL_DIALECT_TSQL)
 			{
-				for (int i = 0; i < strlen(BBF_ESC_CHAR_REPLC); i++)
+				for (int i = 0; i < BBF_ESC_CHAR_REPLC_LEN; i++)
 				{
 					NextByte(p, plen);
 				}
@@ -182,6 +183,15 @@ MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
 							 errmsg("LIKE pattern must not end with escape character")));
 				firstpat = GETCHAR(p[1], locale);
 			}
+			else if (sql_dialect == SQL_DIALECT_TSQL && strncmp(p, BBF_ESC_CHAR_REPLC, BBF_ESC_CHAR_REPLC_LEN) == 0)
+			{
+				int esc_len = BBF_ESC_CHAR_REPLC_LEN;
+				if (plen <= esc_len)
+					ereport(ERROR,
+							(errcode(ERRCODE_INVALID_ESCAPE_SEQUENCE),
+							 errmsg("LIKE pattern must not end with escape character")));
+				firstpat = GETCHAR(p[esc_len]);
+			}
 			else
 				firstpat = GETCHAR(*p, locale);
 
@@ -193,7 +203,7 @@ MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
 
 					if (matched != LIKE_FALSE)
 						return matched; /* TRUE or ABORT */
-				} else if (firstpat == '[' && sql_dialect == SQL_DIALECT_TSQL)
+				} else if (sql_dialect == SQL_DIALECT_TSQL && firstpat == '[')
 				{
 					int			matched = MatchText(t, tlen, p, plen,
 													locale);
@@ -217,7 +227,7 @@ MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
 			NextByte(p, plen);
 			continue;
 		}
-		else if (*p == '[' && sql_dialect == SQL_DIALECT_TSQL)
+		else if (sql_dialect == SQL_DIALECT_TSQL && *p == '[')
 		{
 			/* Tsql deal with [ and ] wild character */
 			Oid cid = InvalidOid;
@@ -443,7 +453,7 @@ MatchText(const char *t, int tlen, const char *p, int plen, pg_locale_t locale)
 		NextByte(p, plen);
 	}
 
-	if (tlen > 0 && sql_dialect == SQL_DIALECT_TSQL)
+	if (sql_dialect == SQL_DIALECT_TSQL && tlen > 0)
 	{
 		/* End of pattern, but not of text.
 		 *
@@ -511,7 +521,7 @@ do_like_escape(text *pat, text *esc)
 		result = (text *) palloc(plen * 2 + VARHDRSZ);
 	r = VARDATA(result);
 
-	if (elen==0 && sql_dialect == SQL_DIALECT_TSQL)
+	if (sql_dialect == SQL_DIALECT_TSQL && elen==0)
 	{
 		/*
 		 * Escape string is empty, just throw the error.
@@ -561,7 +571,7 @@ do_like_escape(text *pat, text *esc)
 				{
 					/* check if direct copy string (unicode char) is valid */
 
-					for (int i = 0; i < strlen(BBF_ESC_CHAR_REPLC); i++){
+					for (int i = 0; i < BBF_ESC_CHAR_REPLC_LEN; i++){
 						*r++ = BBF_ESC_CHAR_REPLC[i];
 					}
 					NextChar(p, plen);