Re: Decimal: Formal review begins.

From: Matt Borland (matt_at_[hidden])
Date: 2025-01-16 17:01:01

On Thursday, January 16th, 2025 at 11:45 AM, Peter Dimov via Boost <boost_at_[hidden]> wrote:

> 

> 

> Ruben Perez wrote:
> 

> > 4. From sprintf's documentation: "In the interest of safety sprintf simply calls
> > snprintf with buf_size equal to sizeof(buffer). ". This doesn't look right. This is
> > what the implementation looks like:
> > 

> > template <typename... T>
> > inline auto sprintf(char* buffer, const char* format, T... values) noexcept
> > #ifndef BOOST_DECIMAL_HAS_CONCEPTS
> > ->
> > std::enable_if_t<detail::is_decimal_floating_point_v<std::common_type_t<T.
> > ..>>,
> > int>
> > #else
> > -> int requires
> > detail::is_decimal_floating_point_v<std::common_type_t<T...>>
> > #endif
> 

> ...
> 

> It would be better to make the constraint say what we actually mean, namely,
> "each type in T... should be integral, floating point, or decimal floating point".
> 

> Using something entirely different just because we think it's equivalent to the
> above isn't superior in any way to just saying what we actually mean, even if
> it were actually equivalent. (It isn't because common_type can be specialized
> by the user.)
> 

I guess I don't know a "superior" way to write all types in T should be decimal floating point types. I thought that was pretty succinct and communicated requirements fine.

> (Providing sprintf in 2025 is a bit debatable, std::format is the way to go here.)
> 

I found that to get <format> to actually work correctly you need GCC >= 13, Clang >= 18, and MSVC >= 1940 which is a relatively high bar.

Matt