mariadb/storage/innobase/pars/pars0lex.l

/******************************************************
SQL parser lexical analyzer: input file for the GNU Flex lexer generator

(c) 1997 Innobase Oy

Created 12/14/1997 Heikki Tuuri
Published under the GPL version 2

The InnoDB parser is frozen because MySQL takes care of SQL parsing.
Therefore we normally keep the InnoDB parser C files as they are, and do
not automatically generate them from pars0grm.y and pars0lex.l.

How to make the InnoDB parser and lexer C files:

1. Run ./make_flex.sh to generate lexer files.

2. Run ./make_bison.sh to generate parser files.

These instructions seem to work at least with bison-1.875d and flex-2.5.31 on
Linux.
*******************************************************/

%option nostdinit
%option 8bit
%option warn
%option pointer
%option never-interactive
%option nodefault
%option noinput
%option nounput
%option noyywrap
%option noyy_scan_buffer
%option noyy_scan_bytes
%option noyy_scan_string
%option nounistd

%{
#define YYSTYPE que_node_t*

#include "univ.i"
#include "pars0pars.h"
#include "pars0grm.h"
#include "pars0sym.h"
#include "mem0mem.h"
#include "os0proc.h"

#define malloc(A)	ut_malloc(A)
#define free(A)		ut_free(A)
#define realloc(P, A)	ut_realloc(P, A)
#define exit(A) 	ut_error

#define YY_INPUT(buf, result, max_size) pars_get_lex_chars(buf, &result, max_size)

/* String buffer for removing quotes */
static ulint	stringbuf_len_alloc = 0; /* Allocated length */
static ulint	stringbuf_len = 0; /* Current length */
static char*	stringbuf; /* Start of buffer */
/* Appends a string to the buffer. */
static
void
string_append(
/*==========*/
	const char*	str,	/* in: string to be appended */
	ulint		len)	/* in: length of the string */
{
	if (stringbuf == NULL) {
		stringbuf = malloc(1);
		stringbuf_len_alloc = 1;
	}

	if (stringbuf_len + len > stringbuf_len_alloc) {
		while (stringbuf_len + len > stringbuf_len_alloc) {
			stringbuf_len_alloc <<= 1;
		}
		stringbuf = realloc(stringbuf, stringbuf_len_alloc);
	}

	memcpy(stringbuf + stringbuf_len, str, len);
	stringbuf_len += len;
}

%}

DIGIT	[0-9]
ID	[a-z_A-Z][a-z_A-Z0-9]*
BOUND_LIT	\:[a-z_A-Z0-9]+
BOUND_ID	\$[a-z_A-Z0-9]+

%x comment
%x quoted
%x id
%%

{DIGIT}+	{
			yylval = sym_tab_add_int_lit(pars_sym_tab_global,
								atoi(yytext));
			return(PARS_INT_LIT);
}

{DIGIT}+"."{DIGIT}* {
			ut_error;	/* not implemented */

			return(PARS_FLOAT_LIT);
}

{BOUND_LIT}	{
			ulint	type;

			yylval = sym_tab_add_bound_lit(pars_sym_tab_global,
				yytext + 1, &type);

			return(type);
}

{BOUND_ID}	{
			yylval = sym_tab_add_bound_id(pars_sym_tab_global,
				yytext + 1);

			return(PARS_ID_TOKEN);
}

"'"		{
/* Quoted character string literals are handled in an explicit
start state 'quoted'.  This state is entered and the buffer for
the scanned string is emptied upon encountering a starting quote.

In the state 'quoted', only two actions are possible (defined below). */
			BEGIN(quoted);
			stringbuf_len = 0;
}
<quoted>[^\']+	{
			/* Got a sequence of characters other than "'":
			append to string buffer */
			string_append(yytext, yyleng);
}
<quoted>"'"+	{
			/* Got a sequence of "'" characters:
			append half of them to string buffer,
			as "''" represents a single "'".
			We apply truncating division,
			so that "'''" will result in "'". */

			string_append(yytext, yyleng / 2);

			/* If we got an odd number of quotes, then the
			last quote we got is the terminating quote.
			At the end of the string, we return to the
			initial start state and report the scanned
			string literal. */

			if (yyleng % 2) {
				BEGIN(INITIAL);
				yylval = sym_tab_add_str_lit(
					pars_sym_tab_global,
					(byte*) stringbuf, stringbuf_len);
				return(PARS_STR_LIT);
			}
}

\"		{
/* Quoted identifiers are handled in an explicit start state 'id'.
This state is entered and the buffer for the scanned string is emptied
upon encountering a starting quote.

In the state 'id', only two actions are possible (defined below). */
			BEGIN(id);
			stringbuf_len = 0;
}
<id>[^\"]+	{
			/* Got a sequence of characters other than '"':
			append to string buffer */
			string_append(yytext, yyleng);
}
<id>\"+	{
			/* Got a sequence of '"' characters:
			append half of them to string buffer,
			as '""' represents a single '"'.
			We apply truncating division,
			so that '"""' will result in '"'. */

			string_append(yytext, yyleng / 2);

			/* If we got an odd number of quotes, then the
			last quote we got is the terminating quote.
			At the end of the string, we return to the
			initial start state and report the scanned
			identifier. */

			if (yyleng % 2) {
				BEGIN(INITIAL);
				yylval = sym_tab_add_id(
					pars_sym_tab_global,
					(byte*) stringbuf, stringbuf_len);

				return(PARS_ID_TOKEN);
			}
}

"NULL"		{
			yylval = sym_tab_add_null_lit(pars_sym_tab_global);

			return(PARS_NULL_LIT);
}

"SQL"		{
			/* Implicit cursor name */
			yylval = sym_tab_add_str_lit(pars_sym_tab_global,
							(byte*) yytext, yyleng);
			return(PARS_SQL_TOKEN);
}

"AND"		{
			return(PARS_AND_TOKEN);
}

"OR"		{
			return(PARS_OR_TOKEN);
}

"NOT"		{
			return(PARS_NOT_TOKEN);
}

"PROCEDURE"	{
			return(PARS_PROCEDURE_TOKEN);
}

"IN"		{
			return(PARS_IN_TOKEN);
}

"OUT"		{
			return(PARS_OUT_TOKEN);
}

"BINARY"	{
	 		return(PARS_BINARY_TOKEN);
}

"BLOB"		{
	 		return(PARS_BLOB_TOKEN);
}

"INT"		{
	 		return(PARS_INT_TOKEN);
}

"INTEGER"	{
	 		return(PARS_INT_TOKEN);
}

"FLOAT"		{
	 		return(PARS_FLOAT_TOKEN);
}

"CHAR"		{
	 		return(PARS_CHAR_TOKEN);
}

"IS"		{
			return(PARS_IS_TOKEN);
}

"BEGIN"		{
			return(PARS_BEGIN_TOKEN);
}

"END"		{
			return(PARS_END_TOKEN);
}

"IF"		{
			return(PARS_IF_TOKEN);
}

"THEN"		{
			return(PARS_THEN_TOKEN);
}

"ELSE"		{
			return(PARS_ELSE_TOKEN);
}

"ELSIF"		{
			return(PARS_ELSIF_TOKEN);
}

"LOOP"		{
			return(PARS_LOOP_TOKEN);
}

"WHILE"		{
			return(PARS_WHILE_TOKEN);
}

"RETURN"	{
			return(PARS_RETURN_TOKEN);
}

"SELECT"	{
			return(PARS_SELECT_TOKEN);
}

"SUM"		{
			return(PARS_SUM_TOKEN);
}

"COUNT"		{
			return(PARS_COUNT_TOKEN);
}

"DISTINCT"	{
			return(PARS_DISTINCT_TOKEN);
}

"FROM"		{
			return(PARS_FROM_TOKEN);
}

"WHERE"		{
			return(PARS_WHERE_TOKEN);
}

"FOR"		{
			return(PARS_FOR_TOKEN);
}

"CONSISTENT"	{
			return(PARS_CONSISTENT_TOKEN);
}

"READ"		{
			return(PARS_READ_TOKEN);
}

"ORDER"		{
			return(PARS_ORDER_TOKEN);
}

"BY"		{
			return(PARS_BY_TOKEN);
}

"ASC"		{
			return(PARS_ASC_TOKEN);
}

"DESC"		{
			return(PARS_DESC_TOKEN);
}

"INSERT"	{
			return(PARS_INSERT_TOKEN);
}

"INTO"		{
			return(PARS_INTO_TOKEN);
}

"VALUES"	{
			return(PARS_VALUES_TOKEN);
}

"UPDATE"	{
			return(PARS_UPDATE_TOKEN);
}

"SET"		{
			return(PARS_SET_TOKEN);
}

"DELETE"	{
			return(PARS_DELETE_TOKEN);
}

"CURRENT"	{
			return(PARS_CURRENT_TOKEN);
}

"OF"		{
			return(PARS_OF_TOKEN);
}

"CREATE"	{
			return(PARS_CREATE_TOKEN);
}

"TABLE"		{
			return(PARS_TABLE_TOKEN);
}

"INDEX"		{
	 		return(PARS_INDEX_TOKEN);
}

"UNIQUE"	{
	 		return(PARS_UNIQUE_TOKEN);
}

"CLUSTERED"	{
	 		return(PARS_CLUSTERED_TOKEN);
}

"DOES_NOT_FIT_IN_MEMORY"	{
			return(PARS_DOES_NOT_FIT_IN_MEM_TOKEN);
}

"ON"		{
	 		return(PARS_ON_TOKEN);
}

"DECLARE"	{
			return(PARS_DECLARE_TOKEN);
}

"CURSOR"	{
			return(PARS_CURSOR_TOKEN);
}

"OPEN"	{
			return(PARS_OPEN_TOKEN);
}

"FETCH"	{
			return(PARS_FETCH_TOKEN);
}

"CLOSE"	{
			return(PARS_CLOSE_TOKEN);
}

"NOTFOUND"	{
			return(PARS_NOTFOUND_TOKEN);
}

"TO_CHAR"	{
			return(PARS_TO_CHAR_TOKEN);
}

"TO_NUMBER"	{
			return(PARS_TO_NUMBER_TOKEN);
}

"TO_BINARY"	{
			return(PARS_TO_BINARY_TOKEN);
}

"BINARY_TO_NUMBER" {
			return(PARS_BINARY_TO_NUMBER_TOKEN);
}

"SUBSTR"	{
			return(PARS_SUBSTR_TOKEN);
}

"REPLSTR"	{
			return(PARS_REPLSTR_TOKEN);
}

"CONCAT"	{
			return(PARS_CONCAT_TOKEN);
}

"INSTR"		{
			return(PARS_INSTR_TOKEN);
}

"LENGTH"	{
			return(PARS_LENGTH_TOKEN);
}

"SYSDATE"	{
			return(PARS_SYSDATE_TOKEN);
}

"PRINTF"	{
			return(PARS_PRINTF_TOKEN);
}

"ASSERT"	{
			return(PARS_ASSERT_TOKEN);
}

"RND"		{
			return(PARS_RND_TOKEN);
}

"RND_STR"	{
			return(PARS_RND_STR_TOKEN);
}

"ROW_PRINTF"	{
			return(PARS_ROW_PRINTF_TOKEN);
}

"COMMIT"	{
			return(PARS_COMMIT_TOKEN);
}

"ROLLBACK"	{
			return(PARS_ROLLBACK_TOKEN);
}

"WORK"		{
			return(PARS_WORK_TOKEN);
}

"UNSIGNED"	{
			return(PARS_UNSIGNED_TOKEN);
}

"EXIT"		{
			return(PARS_EXIT_TOKEN);
}

"FUNCTION"	{
			return(PARS_FUNCTION_TOKEN);
}

{ID}		{
			yylval = sym_tab_add_id(pars_sym_tab_global,
							(byte*)yytext,
							ut_strlen(yytext));
			return(PARS_ID_TOKEN);
}

".."		{
			return(PARS_DDOT_TOKEN);
}

":="		{
			return(PARS_ASSIGN_TOKEN);
}

"<="		{
			return(PARS_LE_TOKEN);
}

">="		{
			return(PARS_GE_TOKEN);
}

"<>"		{
			return(PARS_NE_TOKEN);
}

"("		{

			return((int)(*yytext));
}

"="		{

			return((int)(*yytext));
}

">"		{

			return((int)(*yytext));
}

"<"		{

			return((int)(*yytext));
}

","		{

			return((int)(*yytext));
}

";"		{

			return((int)(*yytext));
}

")"		{

			return((int)(*yytext));
}

"+" 		{

			return((int)(*yytext));
}

"-"		{

			return((int)(*yytext));
}

"*"		{

			return((int)(*yytext));
}

"/"		{

			return((int)(*yytext));
}

"%"		{

			return((int)(*yytext));
}

"{"		{

			return((int)(*yytext));
}

"}"		{

			return((int)(*yytext));
}

"?"		{

			return((int)(*yytext));
}

"/*"			BEGIN(comment); /* eat up comment */

<comment>[^*]*
<comment>"*"+[^*/]*
<comment>"*"+"/"        BEGIN(INITIAL);

[ \t\n]+		/* eat up whitespace */


.		{
			fprintf(stderr,"Unrecognized character: %02x\n",
				*yytext);

			ut_error;

			return(0);
}

%%