Thread: Is there a way make the lex program match multiple line?

Hi community,

I am making a config program, use the lex program to analyse input.

/*
    config.l
        The lex rule file for toysql
    Wen Yi
*/
%option noyywrap
%{
#include <string.h>
#include <stdlib.h>   
%}
%%
-[0-9]+.[0-9]+|[0-9]+.[0-9]+|-[0-9]+|[0-9]+ { printf("Number = %lf\n", atof(yytext)); }
[a-zA-Z_0-9]+  { printf("Token = %s\n", yytext); }
['].+['] {
    yytext[yyleng - 1] = '\0';
    memmove(yytext, yytext + 1, yyleng - 1);
    printf("String = %s\n", yytext);
}

; {}
. { printf("Anything: '%s'\n", yytext);  }
%%
int main()
{
    yylex();
    return 0;
}

But when I try to run it:

[beginnerc@bogon config]$ ./a.out
'This is a single line string'
String = This is a single line string

'This is a multiple line string
Anything: '''
Token = This
Anything: ' '
Token = is
Anything: ' '
Token = a
Anything: ' '
Token = multiple
Anything: ' '
Token = line
Anything: ' '
Token = string

Can someone give me some advice to make the ['].+['] match multiple string?

Thanks in advance!

Yours,

Wen Yi

---

I rewrite, and I think the [']([^']|\n)['] should be [']([^']|\n)+[']

Thanks very much!

Yours,

Wen Yi.

 

"Wen Yi" <896634148@qq.com> writes:
> Can someone give me some advice to make the ['].+['] match multiple string?

You should check the flex manual, but it's likely that "." doesn't
match newline.  Another problem with this pattern is that "."
*does* match "'", so it's ambiguous what will happen with quotes.
You probably need something closer to [']([^']|\n)[']

regards, tom lane