On 12/18/24 11:57, Nina Dinka Ranns via Liaison wrote:
> Hi all,
>
> We have two SG22 issues that are targeted for C++26.
>
> P3477R0 <https://wg21.link/p3477r0> There are exactly 8 bits in a byte
> (JF Bastien)
> P2746R0 <https://wg21.link/p2746r0> Deprecate and Replace Fenv
> Rounding Modes (Hans Boehm)
I agree C std pragmas are not implemented widely. That is unfortunate. I
would also prefer ISO/IEC 60559:2020 in the main part and the other
representation of floating-point numbers in the Annex.
ISO is presently voting on ISO/IEC DTS 18661-4 (Ed 2) and ISO/IEC DTS
18661-5 (Ed 2). Austria has voted in favor on 2024-12-10.

The pragmas we are talking about are in ISO/IEC DTS 18661-1 and in C99
... C23 already.

So deprecating C-Standard macros before most compilers haven even
implemented them correctly (That would be good.) and - much more
importantly - without an viable alternative for C seems premature at
least for C.

Having a viable alternate first and then maybe deprecate, what does not
work all that well in practice, would by far be the better sequencing.

I am no numerical expert but dare to express: Going back and forth with
our decisions without clear direction we are willing to follow for a
significant amount of time is not good.

Sorry, if I addressed to many and if the concerns I voiced for C are not
so relevant for C++.

Regards, Aaron Peter Bachmann

>
> Let's find a time to discuss them and give feedback from SG22
> perspective. Some possible slots are in the following doodle poll.
> https://doodle.com/meeting/participate/id/bD60G9qe
>
> The doodle poll is in Munich time because I was unable to figure out
> how to get it to be in UTC time. Munich time zone chosen at random. :)
> Advice welcome.
>
> Thank you,
> Nina
>
> _______________________________________________
> Liaison mailing list
> Liaison_at_[hidden]
> Subscription:https://lists.isocpp.org/mailman/listinfo.cgi/liaison
> Searchable archives:http://lists.isocpp.org/liaison/2024/12/index.php