Re: [lto] Re-implement pass_ipa_free_lang_specifics [3/3]

[Diego Novillo <dnovillo at google dot com>]

On Tue, Jan 20, 2009 at 11:31, Richard Guenther
<richard.guenther@gmail.com> wrote:

>> +         if (TREE_CODE (f1) != TREE_CODE (f2)
>> +             || DECL_NAME (f1) != DECL_NAME (f2))
>
> Uh, using DECL_NAME looks ... funny?  Any reason to not use
>
>         if (gimple_types_compatible_p (TREE_TYPE (f1), TREE_TYPE (f2))
>
> here?
>
>> +           break;
>> +       }
>> +
>> +      return f1 && f2 ? 0 : 1;
>> +    }
>> +
>> +  return TYPE_MAIN_VARIANT (type1) == TYPE_MAIN_VARIANT (type2);
>
> shouldn't this honour TYPE_CANONICAL as well?

Thanks.  Yes, I agree with both comments.

This is ancient code that I didn't want to change just yet, but will
likely need changes now that we are going to start using it
exclusively after pass_ipa_free_lang_data.


Diego.