Re: [PATCH, Fortran] ABSTRACT INTERFACE

[Uros Bizjak <ubizjak at gmail dot com>]

Hello!

> +/* Tests whether name is the name of an intrinsic type. If it is
> +   the passed error_message is printed and TRUE is returned.
> +   The error message must to contain a single %s for the name.  */
> +
> +bool
> +gfc_is_intrinsic_typename (const char *name, const char *error_message)
> +{
> +  if (strcmp (name, "integer") == 0
> +      || strcmp (name, "real") == 0
> +      || strcmp (name, "character") == 0
> +      || strcmp (name, "logical") == 0
> +      || strcmp (name, "complex") == 0
> +      || strcmp (name, "doubleprecision") == 0
> +      || strcmp (name, "doublecomplex") == 0)
> +    {
> +      gfc_error (error_message, name);
> +      return true;
> +    }
> +
> +  return false;
> +}
>   

IMO, it would better if this function only returns true or false, 
without emitting error message.

>        gfc_new_block->attr.abstract = 1;
> -      if (!strcmp(gfc_new_block->name,"integer")
> -	  || !strcmp(gfc_new_block->name,"real")
> -	  || !strcmp(gfc_new_block->name,"complex")
> -	  || !strcmp(gfc_new_block->name,"character")
> -	  || !strcmp(gfc_new_block->name,"logical"))
> -	gfc_error ("Name of ABSTRACT INTERFACE at %C cannot be the same as "
> -		   "an intrinsic type: %s",gfc_new_block->name);
> +
> +      /* Print an error if it has the name of an intrinsic type.  */
> +      gfc_is_intrinsic_typename (gfc_new_block->name,
> +				 "Name '%s' of ABSTRACT INTERFACE at %C "
> +				 "cannot be the same as an intrinsic type");
>   

if (gfc_is_intrinsic_typename (gfc_new_block->name))
   gfc_error ("... %s", gfc_new_block->name);

For me, it is easier to debug gcc/gfortran this way. The call to 
gfc_error() is the place of an error, and IMO, the code is more readable.

Uros.