Hallo Ulrich,

Ulrich Weigand wrote:
> Markus Deuling wrote:
> 
>> this patch removes following macros from gdbarch.sh:
>>
>> * TARGET_SHORT_BIT
>> * TARGET_INT_BIT
>> * TARGET_LONG_BIT
>> * TARGET_LONG_LONG_BIT 
> 
> Just some minor formatting issues:
> 
>  
> -      high_bit = ((ULONGEST)1) << (TARGET_LONG_BIT-1);
> +      high_bit = ((ULONGEST)1) << (gdbarch_long_bit (current_gdbarch)  - 1);
> This adds one space too many.
> 
> -      if (sizeof (ULONGEST) * HOST_CHAR_BIT < TARGET_LONG_LONG_BIT)
> +      if (sizeof (ULONGEST) * HOST_CHAR_BIT < gdbarch_long_long_bit
> +						(current_gdbarch))
> I'd rather write this as
>       if (sizeof (ULONGEST) * HOST_CHAR_BIT
>           < gdbarch_long_long_bit (current_gdbarch))
> 
> -	  && aux->x_sym.x_misc.x_lnsz.x_size > TARGET_LONG_BIT)
> +	  && aux->x_sym.x_misc.x_lnsz.x_size > gdbarch_long_bit
> +						 (current_gdbarch))
> Likewise.
> 
> -	  && aux->x_sym.x_misc.x_lnsz.x_size > TARGET_LONG_BIT)
> +	  && aux->x_sym.x_misc.x_lnsz.x_size > gdbarch_long_bit 
> +						 (current_gdbarch))
> Likewise.
> 
> -  if ((TARGET_INT_BIT != TARGET_LONG_BIT 
> -       && ((n >> 2) >> (TARGET_INT_BIT-2)))   /* Avoid shift warning */
> +  if ((gdbarch_int_bit (current_gdbarch) != gdbarch_long_bit (current_gdbarch)
> +       && ((n >> 2) >> (gdbarch_int_bit (current_gdbarch)-2)))   /* Avoid shift warning */
> Line too long.
> 
> -      if (sizeof (ULONGEST) * HOST_CHAR_BIT < TARGET_LONG_LONG_BIT)
> +      if (sizeof (ULONGEST) * HOST_CHAR_BIT < gdbarch_long_long_bit
> +						(current_gdbarch))
> See above.
> 
> -	       && FIELD_BITSIZE (fip->list->field) == TARGET_INT_BIT)
> +	       && FIELD_BITSIZE 
> +		(fip->list->field) == gdbarch_int_bit (current_gdbarch))
> I'd rather write this as:
>                   FIELD_BITSIZE (fip->list->field)
>                   == gdbarch_int_bit (current_gdbarch)
> 
> -#define TARGET_LONG_BYTES (TARGET_LONG_BIT / TARGET_CHAR_BIT)
> +#define TARGET_LONG_BYTES (gdbarch_long_bit (current_gdbarch) \
> +  / TARGET_CHAR_BIT)
> Indentation.
> 
>  
> Otherwise this looks fine to me.  (It does conflict heavily with my
> builtin_type cleanup patch series, unfortunately ...   But I guess
> your patch still should go first, and I'll update mine accordingly.)
> 
> Bye,
> Ulrich
> 

thank you for your review. I attached the new version of the patch.


-- 
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com