windows: birthtime and ctime

[ThomasWaldmann]

From: https://docs.python.org/3.12/library/os.html

"""
Changed in version 3.12: st_birthtime[_ns] is now available on Windows.
"""

"""
Changed in version 3.12: The st_ctime attribute of a stat result is deprecated on Windows. The file creation time is properly available as st_birthtime, and in the future st_ctime may be changed to return zero or the metadata change time, if available.
"""

"""
Changed in version 3.12: st_ctime_ns is deprecated on Windows. Use st_birthtime_ns for the file creation time. In the future, st_ctime will contain the time of the most recent metadata change, as for other platforms.
"""

"""
Changed in version 3.12: On Windows, st_ctime is deprecated. Eventually, it will contain the last metadata change time, for consistency with other platforms, but for now still contains creation time. Use st_birthtime for the creation time.
"""

These are nice consistency changes with the goal of making windows timestamp semantics closer to posix OSes like linux.

[songxiaocheng]

I couldn't find any Windows-specific handling for timestamp restoration, even though Borg always stores ctime and birthtime where available. Am I missing something here?

[ThomasWaldmann]

Looks like restore_attrs only deals with mtime/atime on windows right now.

Guess for now, we only want to add archiving birthtime on windows:

If st has birthtime_ns attribute, we use that (should work for py312+).

Otherwise we fall back to ctime_ns (which should have the birthtime for older pythons).

When the day comes and they release a python that has "change time" available in ctime_ns, we can also add archiving ctime.