Re: [gfortran] add SIGNAL and ALARM intrinsics

[Bernhard Fischer <rep dot nop at aon dot at>]

On Sun, Nov 20, 2005 at 10:07:19AM +0100, Andreas Jaeger wrote:
>FX Coudert <fxcoudert@gmail.com> writes:
>
>>>> Then you should at least use *status = (long) signal (*number, handler);
>>>> to shut up the warnings about the cast if that's intentional.
>>
>> Doesn't that only works on platforms where a pointer has the same size
>> as long?
>
>Which is most platforms except MS Windows on x86-64.  You should use
>intptr_t if it's defined...
>
>>> Plus a comment in the source would help too I guess.
>>
>> OK. Will do it when I understand the first point.
>

I'm attaching a patch to use intptr_t for the signals. I do not know if
the #if 0'd test is the proper thing to do and if that test will work as
expected, though.

Please let me know if the test is appropriate and commit or let me know
if it is ok so i can commit it.

TIA,
Bernhard

Attachment: gcc.gcc-4_1-branch.intptr_t__for_signal_intrinsic.0.diff
Description: Text document