New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[fs.filesystem.syn] Fixes related to chrono::file_clock usage in <filesystem> #2285
Conversation
…nd chrono::file_clock, so it must include <chrono> for their declarations The removed paragraph refers to trivial-clock that has now been replaced by chrono::file_clock. The paragraph is now unnecessary since chrono::file_clock is fully described in [time.clock.file]. See also http://cplusplus.github.io/LWG/lwg-active.html#3145
Requiring |
Would the rest of the proposed change qualify as editorial if I remove the inclusion of |
I'm not sure if the rest is editorial even. Removing redundant wording that applies to an unspecified type that is no longer used is editorial, but I wonder if the wording about reflecting the OS-dependent resolution and range of file time values should be present on |
In [time.syn], there's a
which looks pretty close to |
It would be good to add "that is capable of representing and measuring file time values." to the definition of file_clock in [time.clock.file.overview] p1, though. @zygoloid, is that still editorial? |
We could suggest that change to [time.clock.file.overview] as part of the resolution of LWG issue 3145, which is currently under discussion on the reflector. |
The old type was required to be "capable of representing and measuring file time values". The new type "should [have sufficient] resolution and range [to] reflect the operating system dependent resolution and range of file time values", which seems like a strictly weaker requirement -- and appears to be an intentional weakening. (I would note that there are some Linux file systems that represent 128-bit file times, and we probably want to permit a 64-bit file time implementation regardless, so the existing "should" wording may well be the right thing. But whether it's right or not is up to LWG.) That is to say: I'm merging this to remove the remaining vestige of a deleted type; retaining the "capable of representing and measuring file time values" and applying it to the new type would in my view not be editorial. |
The synopsis of refers to the types chrono::time_point and chrono::file_clock, so it must include for their declarations
The removed paragraph refers to trivial-clock that has now been replaced by chrono::file_clock. The paragraph is now unnecessary since chrono::file_clock is fully described in [time.clock.file].
See also http://cplusplus.github.io/LWG/lwg-active.html#3145