<chrono>: Formatting large hh_mm_ss values should be permitted

[StephanTLavavej]

Reported by @josuttis to the LWG mailing list. @HowardHinnant confirmed that this is a bug. After checking the Standard, I agree that I don't see any direct justification for the exception that we're throwing. [tab:time.format.spec] simply says "%H The hour (24-hour clock) as a decimal number. If the result is a single digit, it is prefixed with 0." and the parenthetical "(24-hour clock)" doesn't seem to justify emitting an error. (If the type contains information, we should print it.)

C:\Temp>type meow.cpp

#include <chrono>
#include <format>
#include <iostream>
using namespace std;

int main() {
    try {
        const chrono::days d{3};
        const chrono::hh_mm_ss hms{d};
        cout << "d: " << d << "\n";
        cout << "hms: " << hms << "\n";
    } catch (const format_error& fe) {
        cout << fe.what() << "\n";
    }
}

C:\Temp>cl /EHsc /nologo /W4 /std:c++latest /MTd /Od meow.cpp && meow
meow.cpp
d: 3d
hms: Cannot localize hh_mm_ss with an absolute value of 24 hours or more.

Compare https://godbolt.org/z/WjsPGP771 where libstdc++ prints:

d: 3d
hms: 72:00:00

The fix may be as simple as removing this code:

STL/stl/inc/chrono

Lines 5683 to 5688 in 0a2d59e

	if constexpr (_Is_specialization_v<_Ty, hh_mm_ss>) {
	if (_Spec._Type == 'H' && _Val.hours() >= hours{24}) {
	_Throw_format_error("Cannot localize hh_mm_ss with an absolute value of 24 hours or more.");
	}
	return;
	}

[josuttis]

Thanks Stephan. Did you also look at the inconsistencies of using std::format()? Am 27. April 2023 22:10:14 MESZ schrieb "Stephan T. Lavavej" ***@***.***>:

Reported by @josuttis to the LWG mailing list. @HowardHinnant confirmed that this is a bug. After checking the Standard, I agree that I don't see any direct justification for the exception that we're throwing. \[tab:time.format.spec\] simply says "`%H` The hour (24-hour clock) as a decimal number. If the result is a single digit, it is prefixed with `0`." and the parenthetical "(24-hour clock)" doesn't seem to justify emitting an error. (If the type contains information, we should print it.) ``` C:\Temp>type meow.cpp ``` ```cpp #include <chrono> #include <format> #include <iostream> using namespace std; int main() { try { const chrono::days d{3}; const chrono::hh_mm_ss hms{d}; cout << "d: " << d << "\n"; cout << "hms: " << hms << "\n"; } catch (const format_error& fe) { cout << fe.what() << "\n"; } } ``` ``` C:\Temp>cl /EHsc /nologo /W4 /std:c++latest /MTd /Od meow.cpp && meow meow.cpp d: 3d hms: Cannot localize hh_mm_ss with an absolute value of 24 hours or more. ``` Compare https://godbolt.org/z/WjsPGP771 where libstdc++ prints: ``` d: 3d hms: 72:00:00 ``` The fix may be as simple as removing this code: https://github.com/microsoft/STL/blob/0a2d59ed7bfe099199237090998b758c9b0b7403/stl/inc/chrono#L5683-L5688 -- Reply to this email directly or view it on GitHub: #3676 You are receiving this because you were mentioned. Message ID: ***@***.***>

-- Nico Josuttis (sent from my mobile phone)

[cpplearner]

inconsistencies of using std::format()? Sounds like LWG-3856

[josuttis]

In the discussion I added:

std::chrono::days d{3};     std::cout << "3 days: "<< std::format("{:%T}\n", d); gcc:     3 days: 72:00:00 VC++:     3 days: 00:00:00

Or is this the same bug? Am 28.04.2023 um 10:44 schrieb S. B. Tam:

inconsistencies of using |std::format()|? Sounds like LWG-3856 <https://cplusplus.github.io/LWG/issue3856> — Reply to this email directly, view it on GitHub <#3676 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAHXEUYZYPRXOYFSYI5IIJTXDN7NTANCNFSM6AAAAAAXOKXIGU>. You are receiving this because you were mentioned.Message ID: ***@***.***>

-- --- Nicolai M. Josuttis www.josuttis.de +49 (0)531 / 129 88 86 +49 (0)700 / JOSUTTIS Books: C++:http://cppstd20.com,http://cppstd17.com,http://cppmove.com, http://cppstdlib.com,http://tmplbook.com

[StephanTLavavej]

It's likely the same bug but we can check when we're looking at a fix.