https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119141
--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> --- And the note you quoted is about "implicit truncation" to a lower resolution duration i.e. preventing milliseconds being implicitly converted to seconds, which would lose precision. It has nothing to do with preventing milliseconds(LLONG_MAX) from being truncated when converted to nanoseconds **because that's impossible**. You cannot fit LLONG_MAX * 1000000 in a long long without truncation. See the example: [Example 2: duration<int, milli> ms(3); duration<int, micro> us = ms; // OK duration<int, milli> ms2 = us; // error — end example]
