Re: [PATCH v2 2/2] [gdb/tui] Handle unicode chars in prompt

[Tom de Vries <tdevries@suse.de>]

On 6/9/23 17:39, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Tom> +#ifdef HAVE_BTOWC
> Tom> +      {
> Tom> +	int mb_len;
> Tom> +	if (is_mb_char (string, mb_len) && mb_len != 1)
> Tom> +	  {
> Tom> +	    if (mb_len == 0)
> Tom> +	      {
> Tom> +		/* Multi-byte null char.  */
> Tom> +		break;
> Tom> +	      }
> Tom> +
> Tom> +	    waddnstr (w, string, mb_len);
> Tom> +	    string += mb_len;
> Tom> +	    handled = true;
> Tom> +	  }
> Tom> +      }
> Tom> +#endif
> 
> I wonder if this would be simplified by using wchar_iterator.
> 
> This iterator tries to convert just a single character, and has out
> parameters that reflect which input bytes were converted.
> 
> The main benefit would be less #ifdef and no need for is_mb_char in
> tui-io.c.
> 

The iterator constructor also needs a specification of encoding and 
width.  I suppose for encoding we could use host_charset (), but I don't 
know how to get the base width of that char set.

ISTM that's a problem that the multibyte functions take care of for us.

Thanks,
- Tom

> You may need to add a method to wchar_iterator to let the caller skip
> some bytes (you wouldn't want to create a new one on each iteration, as
> it calls iconv_open).  That way the escape handling could stay pretty
> much the same.
> 
> Tom