Victor Zverovich
Victor Zverovich
Thanks for reporting. This is a known limitation of the current implementation relying on conversion to `tm` which has a seconds resolution. The solution would be to parse format specs...
Thanks to @matrackif subsecond formatting now works for durations (#2623). We still need to do the same for time points so keeping the issue open.
Merged, thanks!
CI is failing: https://github.com/fmtlib/fmt/runs/7734651761?check_suite_focus=true
`FMT_CONSTEXPR20` should definitely not be changed. I suggest making the test conditional on whether `__cpp_lib_bit_cast` is defined, removing the `bit_cast` changes and new CI configs (we already have way too...
Looks good but there are some merge conflicts, please rebase.
Thanks!
The behavior in the first example is caused by range items becoming convertible to `string_view` in C++23: https://godbolt.org/z/eYvdszbof. So I guess this is expected although a bit weird and we...
Range elements convertible to `string_view` are now escaped: https://github.com/fmtlib/fmt/commit/796662a612023090a421f9a397fe58b506e4829d. ```c++ std::string line = "a,b-c,d-e,f"; fmt::print("{}\n", line | std::views::split(',')); ``` now produces the following output in C++23 (https://godbolt.org/z/f465jEq4s): ``` ["a", "b-c",...
The problem in the second example is caused by only partial support for non-const-iterable ranges. In particular, `write_range_entry` taking a const reference: https://github.com/fmtlib/fmt/blob/796662a612023090a421f9a397fe58b506e4829d/include/fmt/ranges.h#L545 and a few other places. A PR...