On May 11, 2009, at 11:49 AM, Dag Sverre Seljebotn wrote:
> Lisandro Dalcin wrote:
>> On Mon, May 11, 2009 at 11:52 AM, Dag Sverre Seljebotn
>>> Similarily one could have
>>>
>>> @gccattribute("...")
>>>
>>> which I'd prefer to hacking this into a callconvention mechanism.
>>>
>>> Dag Sverre
>>
>> OK, even better, though I would call it "@attribute(...)" or
>> something
>> like that, as the "gcc" bit seems not appropriate (because the the
>> syntax could target special features of other compilers, not just
>> GCC)
>>
>
> The reason I proposed to call it "gccattribute" is that AFAIK GCC
> is the
> compiler which introduces these and that is how they are referred to
> (although I think icc supports a lot of them as well).
>
> I think "attribute" is far to generic; "cattribute" is slightly better
> but still these are not standard C.
Yeah, gccattribute or cattribute are more clear.
> (In most cases we are better off implementing higher-level support
> anyway, like I did with packed structs, where differenct extensions
> from
> more than one compiler is used through #ifdefs. I'm +0 on adding
> gccattribute, I just responded to Lisandro's ideas on it.)
This opens the option for every obscure attribute one might want to
use, whether/before we support it at a higher level. I don't think
it's high priority though.
- Robert
_______________________________________________
Cython-dev mailing list
[email protected]
http://codespeak.net/mailman/listinfo/cython-dev
