It looks like I was overcomplicating this. The constructor for a std::chrono::time_point does take in a duration from the clock's epoch, which is ultimately what we want to do.
PR up on github for review. -Robert Middleton On Sat, Nov 26, 2022 at 11:10 PM Stephen Webb <swebb2...@gmail.com> wrote: > > Fmt seems to have a formatter for std::chrono::time_point templated by > std::chrono::system_clock and (optionally) std::chrono::utc_clock if my > interpretation of the fmt/chrono.h code in the https://github.com/fmtlib > master branch is correct. > > Quoting from https://en.cppreference.com/w/cpp/chrono/high_resolution_clock > > > > It may be an alias of std::chrono::system_clock > > <https://en.cppreference.com/w/cpp/chrono/system_clock> or > > std::chrono::steady_clock > > <https://en.cppreference.com/w/cpp/chrono/steady_clock>, or a third, > > independent clock. > > > On Sun, Nov 27, 2022 at 2:14 AM Robert Middleton <rmiddle...@apache.org> > wrote: > > > Odd. How would we format the timestamp in that case? According to > > the documentation, fmt has formatters for std::chrono::time_point > > already. > > > > Is this some sort of difference between MSVC and gcc? It should be > > trying to format > > std::chrono::time_point<std::chrono::high_resolution_clock>, which I > > thought was std::chrono::system_clock instead of steady_clock. > > > > -Robert Middleton > > > > > > On Fri, Nov 25, 2022 at 11:03 PM Stephen Webb <swebb2...@gmail.com> wrote: > > > > > > In Visual studio 2019 the line > > > > > > fmt::arg("d", event->getChronoTimeStamp()), > > > > > > > > > causes the compile time error: > > > > > > error C2338: Cannot format an argument. To make type T formattable > > provide > > > a formatter<T> specialization: https://fmt.dev/latest/api.html#udt > > > > > > > > > It seems to need a formatter specialization for > > > std::chrono::steady_clock::time_point. > > > > > > The branch build OK with the above line commented out. > > > > > > On Sat, Nov 26, 2022 at 2:13 AM Robert Middleton <rmiddle...@apache.org> > > > wrote: > > > > > > > I've been working on LOGCXX-514(using fmt as an alternative layout to > > > > the PatternLayout) and I've got a working implementation on Linux. > > > > However, the Windows build currently fails with some sort of symbol or > > > > linking error. Since I'm not all that familiar with Windows, would > > > > somebody with more experience on the Windows side be able to take a > > > > look at it? > > > > > > > > I suspect that this could also be something to do with the wchar on > > > > Windows, but it is not clear to me if that is the case or not. > > > > > > > > -Robert Middleton > > > > > >
