Use flex to uppercase C comments

I want to uppercase C comments in a text using flex.

Here is my flex code：

%{
#include <ctype.h>
%}

%%

\/\/.* {
    for(int i = 0; i < strlen(yytext); i++)
        printf("%c", toupper(yytext[i]));   
}

\/\*[^\*]*\*(\*|[^\*\/][^\*]*)*\/ {
    for(int i = 0; i < strlen(yytext); i++)
        printf("%c", toupper(yytext[i]));
}

%%

int main(int argc ,char* argv[]) {
    yylex();
    return 0;
}

int yywrap() {
    return 1;
}

Here is the testing text:

/*aBc*aBc/aBc*/
/** /aBc*/
/*aBc*/aBc*/
aBc
aBc/*aBC
aBc/aBc*aBc
aBc**/
/*aBc/*aBc
//aBc
//aBc
aBc

The result:

/*ABC*ABC/ABC*/
/** /ABC*/
/*ABC*/aBc*/
aBc
aBc/*ABC
ABC/ABC*ABC
ABC**/
/*ABC/*ABC
//ABC
//aBc
aBc

The second and fourth line from the end in the result are weird.

What's wrong with my program?

Solution

The problem is simply that the regular expression you are using is incorrect.

Here's a correct flex regular expression for C-style comments:

"/*"[^*]*"*"+([^*/][^*]*"*"+)*"/"

That uses double-quotes (a flex feature) to quote the regex metacharacters. Note that it is not necessary to escape regex operators inside character classes.

The alternatives aren't very pretty either:

Forest of leaning timber:
```
\/\*[^*]*\*+([^*/][^*]*\*+)*\/
```
Clutter of character classes:
```
[/][*][^*]*[*]+([^*/][^*]*[*]+)*[/]
```