This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 10/12] Add another test for setcontext
* Carlos O'Donell:
> On 07/25/2018 05:21 PM, Florian Weimer wrote:
>> * Carlos O'Donell:
>>
>>> On 07/25/2018 04:42 PM, Florian Weimer wrote:
>>>> * Carlos O'Donell:
>>>>
>>>>> /* The uc_sigmask on IA64 has the wrong type and this needs fixing,
>>>>> but until that change is evaluated, we fix this here with a cast.
>>>>> See bug XXX. */
>>>>
>>>> Is this really a bug? After this long, I would consider it just a
>>>> quirk of the ia64 API/ABI and move on …
>>>
>>> You can't use sigismember(); I would consider that a bug?
>>
>> Well, if the ABI says the member has the type it has, and that's not
>> compatible with sigismember, then that's how things are. Surely this
>> isn't the only such example. What about the return type of the
>> tsearch function?
>
> What about tsearch? It returns a void* and so you do have to cast that
> to something. Are you saying this is a similar issue? The void* creates
> an expectation of a cast, which I think we don't have here.
The return type is a pointer to a const void *. You are not supposed
to cast the returned pointer, but that const void * (after checking
that the return value is not NULL).