Date: Wed, 26 Nov 2025 15:05:31 -0800
> On Nov 26, 2025, at 9:55 AM, Thiago Macieira via Std-Proposals <std-proposals_at_[hidden]> wrote:
>
> On Wednesday, 26 November 2025 08:27:39 Pacific Standard Time David Brown
> wrote:
>>> Would that make the 512-bit integer the "ludicrously long" and 1024 go
>>> "plaid"? :-)
>>
>> We are onto "Space Balls The Movie" now, aren't we? Can we look up the
>> VHS version of C++29 to see how it all ends? :-)
>
> Oh, why didn't we think of just fast-forwarding!
>
>> I remember reading an April Fool's proposal for adding new integer sizes
>> to C, using combinations of "long" and "short" to get different sizes -
>> like a "short long int" would be 24 bits. Unfortunately I am unable to
>> find it again.
>
> https://wg21.link/p0989 - Standardizing Extended Integers
>
> short long int j; // (1 + 1/0.5)(32) = 48 bits
> short int -> 16 bits
> long short int -> 24 bits
> long long short int -> 28 bits
> long long long short int -> 30 bits
Has this been taken to EWG yet?
I would like to suggest an additional type, this focuses on requesting specific precisions but it lacks a mechanism for simply getting the maximum supported size. I think this paper should be extended with a type to get that maximum supported precision, I propose
longest long -> widest supported integer type
I think this will be useful for many use cases as people can write generic code that automatically works on the widest supported precision rather than having to use a pile of template misery to find the widest type. Including being able to easily write safe multiplication by using `(short longest long) * (short longest long)` and similar.
—Oliver
(Credit to Ryan M for the spelling - not including his complete name because he may not actually want anyone to ever find his name next to this :D :D :D)
>
> --
> Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
> Principal Engineer - Intel Data Center - Platform & Sys. Eng.
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals
>
> On Wednesday, 26 November 2025 08:27:39 Pacific Standard Time David Brown
> wrote:
>>> Would that make the 512-bit integer the "ludicrously long" and 1024 go
>>> "plaid"? :-)
>>
>> We are onto "Space Balls The Movie" now, aren't we? Can we look up the
>> VHS version of C++29 to see how it all ends? :-)
>
> Oh, why didn't we think of just fast-forwarding!
>
>> I remember reading an April Fool's proposal for adding new integer sizes
>> to C, using combinations of "long" and "short" to get different sizes -
>> like a "short long int" would be 24 bits. Unfortunately I am unable to
>> find it again.
>
> https://wg21.link/p0989 - Standardizing Extended Integers
>
> short long int j; // (1 + 1/0.5)(32) = 48 bits
> short int -> 16 bits
> long short int -> 24 bits
> long long short int -> 28 bits
> long long long short int -> 30 bits
Has this been taken to EWG yet?
I would like to suggest an additional type, this focuses on requesting specific precisions but it lacks a mechanism for simply getting the maximum supported size. I think this paper should be extended with a type to get that maximum supported precision, I propose
longest long -> widest supported integer type
I think this will be useful for many use cases as people can write generic code that automatically works on the widest supported precision rather than having to use a pile of template misery to find the widest type. Including being able to easily write safe multiplication by using `(short longest long) * (short longest long)` and similar.
—Oliver
(Credit to Ryan M for the spelling - not including his complete name because he may not actually want anyone to ever find his name next to this :D :D :D)
>
> --
> Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
> Principal Engineer - Intel Data Center - Platform & Sys. Eng.
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals
Received on 2025-11-26 23:05:38