Copying back in the mailing list as this email wasn't subscribed, and I've been having corporate email server issues so have had to switch emails.

Sam

> On 14 Mar 2023, at 10:57 am, Sam Elliott <sam_at_[hidden]> wrote:
>
>>
>> On 13 Mar 2023, at 4:28 pm, Jens Maurer <jens.maurer_at_[hidden]> wrote:
>>
>> On 13/03/2023 14.49, Sam Elliott via Std-Discussion wrote:
>>> I have noticed a discrepancy between the C++ standard's description of `<cfenv>` and the C standard's description of <fenv.h>:
>>>
>>> Specifically, in the current C++ working draft: https://eel.is/c++draft/cfenv <https://eel.is/c++draft/cfenv>, `std::fexcept_t` is denoted as being of "integer type", but the C standards I could find only say: "The type fexcept_t represents the floating-point status flags collectively, including any status the implementation associates with the flags", which by my reading doesn't require the type to be of integer type.
>>>
>>> I'm not sure where the discrepancy has come from in this case. Is there some history I'm missing for the C++ definition of `std::fexcept_t`, which might explain the discrepancy?
>>
>> This is the way it is since N2009 (the April, 2006 Working Draft).
>>
>> This is when <cfenv> was added; the prior working draft N1905 didn't
>> have that header.
>>
>> <cfenv> came to the standard via section 8.6 of
>> Technical Report on C++ Library Extensions
>>
>> https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1836.pdf
>>
>> Section 8 of that TR was voted into the C++ Working Draft in Berlin
>> (April 2006).
>>
>> <cfenv> went into the TR via N1568
>>
>> https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1568.htm
>> with some rationale in N1354
>> https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2002/n1354.htm
>>
>> but C99, as you correctly observe, doesn't mention "integer type".
>>
>> I've submitted an LWG issue and copied you on it.
>
> Thank you for going through the older standards on my behalf, and for filing the LWG issue. I have seen that email, and will follow up on a few more details.
>
>>
>>> the C library in the toolchain I work on has an `fexcept_t` that is not of integer type.
>>
>> Which one is that?
>
> The toolchain is Arm Compiler for Embedded (previously also called Arm Compiler). As far as I can see, the `fexcept_t` in our proprietary C library has been a struct since we added support for the `fenv.h` C header in the late 90's.
>
> Sam
>
>>
>> Jens