RFA: handling of mode attribute

[Kai Tietz]

2009/9/27 Joseph S. Myers <joseph@codesourcery.com>:
> On Sun, 27 Sep 2009, Kai Tietz wrote:
>
>> You are right that the initial point for this patch was to have new
>> types, which are combine-able with type specifiers, like
>> unsigned/signed/_Complex. To have the ability to define such types in
>> target scope would be of course the best, but as far as I can see the
>> current type-system in C/C++ FE doesn't provide this pretty easily. Eg
>
> That is a misconception.  Allowing new type specifier keywords (a limited
> set, where some targets may not support some types) is easy, and much
> simpler than inventing new attributes that do not bind in the normal way
> documented for attributes, as well as a more natural match for the actual
> problem to be solved.  These keywords are pretty similar to char, int,
> float, double, void or _Bool; they can be used with certain other type
> specifier keywords under easy to specify rules.

Ok, an new type concept would be good thing, so that targets can
define new type specifier keywords for modes they are additional want
to support.
But well, the approach of none-standard binding (see as example the
attribute vector_size) isn't that uncommon AFIACS.

Kai

-- 
|  (\_/) This is Bunny. Copy and paste
| (='.'=) Bunny into your signature to help
| (")_(") him gain world domination