Re: [Tinycc-devel] 0.9.22: GNU nano compile error

[charlesrandall4-tinycc]

I worked through the error in the nano code below. 

However, I have isolated the bug which resulted in the
line number being off by one. Here's the sample code.

     1  #define _(s) blah(s)
     2  extern void f();
     3  void version(void)
     4  {
     5      f(_
     6        ("x"));
     7  `
     8  }

Note the backtick at line seven.

When compiled with tcc it points to line six,

# tcc lineno.c
lineno.c:6: unrecognized character \x60

And when compiled with gcc, I get line seven as
expected,

# gcc lineno.c
lineno.c: In function `version':
lineno.c:7: stray '`' in program

-Charles


--- address@hidden wrote:

> --- Laurent Charmet <address@hidden> 
> > NULL is for pointer, and *read_buff_ptr is a char.
> > Ok there is a 
> > casting. Try to modify like this :
> > 
> > *read_buff_ptr = 0;
> 
> Ah, of course. It compiles with the changes below.
> 
> Now, why was the line number in the error message
> off
> by one and is it possible to include a variable name
> or something in the error message to give the user
> more information if this were a more complex
> statement?
> 
> -Charles
> 
> --- nano.c~     2003-12-27 09:35:21.000000000 -0700
> +++ nano.c      2004-11-17 02:41:04.000000000 -0700
> @@ -1803,7 +1803,7 @@
> 
>      }
> 
> -    *read_buff_ptr = (char)NULL;
> +    *read_buff_ptr = 0;
>      close(uniq_fd[0]);
> 
>      /* Process the spelling errors */
> @@ -1812,7 +1812,7 @@
>      while (*read_buff_ptr != '\0') {
> 
>         if ((*read_buff_ptr == '\n') ||
> (*read_buff_ptr == '\r')) {
> -           *read_buff_ptr = (char)NULL;
> +           *read_buff_ptr = 0;
>             if (read_buff_word != read_buff_ptr) {
>                 if
> (!do_int_spell_fix(read_buff_word))
> {
>                     read_buff_word = read_buff_ptr;
> 
> 
>