Re: utimens and non-standardized futimesat

[Jim Meyering]

Eric Blake <address@hidden> wrote:
> According to Eric Blake on 5/22/2008 6:51 AM:
> | According to Jim Meyering on 5/22/2008 6:42 AM:
> | |> |> No need to refer the dir by name:
> | |> |>
> | |> |>   futimens (dirfd. timespec);
> | |> |
> | |> | Btw., even if you don't consider the Posix 200x functions
> | |> | futimens/utimensat, you don't need futimsat(dirfd, NULL, timeval):
> | |> |
> | |> |     futimes (dirfd, timeval);
> | |> |
> | |> | So, actually I think that using futimesat this way is a gratuitous
> | |> | utilization of a glibc extension.
> | |>
> | |
> | | Thanks for the write-up.
> | | That sounds like the right way to go.
> | | Do you feel like doing it?
> |
> | Yes, I'll be tackling this.
> |
>
> Tested on cygwin 1.7.0, where futimens and utimensat exist, and on cygwin
> 1.5.25, where those and futimesat are all missing.  OK to apply?  This
> means that coreutils can now support nanosecond resolution on new enough
> kernels for things like touch and cp -p.

Looks fine to me (modulo Bruno's comments, with which I agree).
I confirmed it makes touch set/preserve full nanosecond
precision on a linux/tmpfs file system.

Thanks!