From: John Maddock (jz.maddock_at_[hidden])
Date: 2022-08-13 08:12:37
On 10/08/2022 21:17, Gero Peterhoff via Boost wrote:
> Am 10.08.22 um 20:23 schrieb John Maddock via Boost:
>> On 10/08/2022 18:09, Gero Peterhoff via Boost wrote:
>>> I would like to add several missing math functions in boost or
>>> specialize/overload existing ones for more types.
>> Which functions?
> 1) scalar types
> - isdenormal
> - trigonometric cot/sec/csc, acot/asec/acsc, acoth/asech/acsch,
> acot2(like atan2) -> the problem is not the implementation as such,
> but a standardized error handling
Yep, I have wondered about those in the past.
> - abs/div for unsigned types -> useful/need in templates
Yep.Â I had to add an "unsigned_abs" function to multiprecision to
handle generic abs calls.
> - inv
> - sgn (support nan/inf, Â±0)
boost::math::sign should do what you want already.
> - rounding
Almost a library in it's own right ;)Â But OK.
> 2) std::complex
> - complete all std-functions for complex
> - complete all special-functions for complex
That's a heck of a lot of work - certainly to get them reliable.
Do you want to make a start with some PR's for the low-hanging fruit?
> for the future
> 3) complex types
> - dual numbers https://en.wikipedia.org/wiki/Dual_number
> - split-complex numbers
> 4) logic
> - a modern constexpr implementation for tribool -> bool3
> - bool4 https://en.wikipedia.org/wiki/Four-valued_logic
> Best, Gero
>>> These should be constexpr if possible.
>>> My problems with the concrete implementations are:
>>> - Which C++ standard to use C++11/14/17/20/23 ?
>>> - Should these be implemented in ccmath and math ?
>>> Â Â a) ccmath requires C++17 (eg "if constexpr"), so it is not
>>> possible to simply call the ccmath version in math, since math
>>> requires C++11.
>> Maybe, but we have announced a move to C++14, and in any case new
>> additions can use whatever std version they require, as long as that
>> requirement is documented and actually needed (ie not just using cool
>> new features "just because").
>> I do note however, that the constexpr function may not be as
>> efficient *at runtime* as they could be due to the need not to use
>> compiler intrinsics / std lib functions and such (so they should
>> probably be marked as consteval in C++20).Â So yes, possibly two
>> versions.Â But it depends what the functions are?
>> Best, John.
>>> Â Â b) On the other hand, the math functions are already included in
>>> the standard (since C++11), so it wouldn't be necessary to continue
>>> to provide them (especially since they are not constexpr).
>>> Where and how exactly and with which C++ standard can I add the
>>> functions ?
>>> Unsubscribe & other changes:
>> Unsubscribe & other changes:
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk