Re: utimensat round 4

[Jim Meyering]

Eric Blake wrote:
> According to Eric Blake on 10/15/2009 12:22 PM:
>> Eric Blake <ebb9 <at> byu.net> writes:
>>
>>> The current state of the series is now pushed.  Subsequent improvements,
>>> like utimensat, will come later.
>>
>> Here's the current state; I'll probably commit up through patch 4 within the
>> next 24 hours, but spend more time hammering on cygwin with patch 5.
>
> And here's what I hope is the final round for gnulib for the *at family
> (utimensat was the last on my list).  I'd push now, except right now
> savannah is experiencing some problems and complaining that the file
> system is read-only.  Coreutils could probably use a followup to add a
> test for using touch with trailing slash.
>
> Eric Blake (2):
>       [1/2] fdutimensat: new module
> coreutils copy.c currently uses gl_utimens (earlier in the series, I
> swapped argument order and added the alternate name fdutimens, to make it
> clear that it is different than *at interfaces) and lutimens.  In order to
> convert copy.c to use fts, we need a directory-relative versions.  Hence:
> fdutimensat and lutimensat.
>
>       [2/2] utimensat: work around Solaris 9 bug
> utimes() (and hence 'touch -c file/') mistakenly succeeds on Solaris 9.
> And in the process of testing this, I discovered that cygwin 1.5.x stat()
> modifies directory atime (since fixed in cygwin 1.7).
> Don't work too hard, make some time for fun as well!
>
...
> Subject: [PATCH 1/2] fdutimensat: new module
>
> Needed for coreutils copy.c to be rewritten to use fts.
>
> * modules/fdutimensat: New file.
> * lib/fdutimensat.c (fdutimensat): Likewise.
> * lib/utimens.h (fdutimensat, lutimensat): Declare new functions.
> * MODULES.html.sh (File system functions): Mention module.
> * modules/fdutimensat-tests: New test.
> * tests/test-fdutimensat.c: Likewise.

Nice work.
Thanks yet again!