RE: GNU Make 3.80 problem on Solaris 8

[Martin Dorey]

> NFS filesystems (at least not NFSv2 or NFSv3) don't support sub-second

> timestamps 

That's definitely not true.  NFSv3 supports nanosecond timestamp
resolution.  This isn't just a theoretical capability.  I'm looking at a
file on a Solaris box exported with NFSv3 to a Linux client, showing a
microsecond precision timestamp all the way up the stack to the ls -l
--full-time output.  touch(1) on my Linux NFSv3 client won't send a
sub-microsecond precision timestamp over the wire to a server which
supports the full nanosecond resolution (a BlueArc Titan).  I'm not sure
where in the stack the nanoseconds have been rounded off, though
utimes(2) would be my guess.

The behavior of my Linux boxes over the last few months/year makes me
think that this
http://www.kerneltraffic.org/kernel-traffic/kt20040605_260.html#6 was
eventually resolved by rounding ext3 file times to a precision of one
second in cache as well as on disk.  A few minutes of googling failed to
confirm this, though, and I'm not quite motivated enough to dive into
the source.  Perhaps that's what Paul was thinking of.
-------------------------------------
Martin's Outlook, BlueArc Engineering


-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of Paul
D. Smith
Sent: Tuesday, December 13, 2005 6:27
To: address@hidden
Cc: address@hidden
Subject: Re: GNU Make 3.80 problem on Solaris 8

%% Warren L Dodge <address@hidden> writes:

  wld> I can't believe this is the first time we have noticed this in
  wld> the two years we have been using 3.80.

One thing to note is that NFS filesystems (at least not NFSv2 or NFSv3)
don't support sub-second timestamps.  So, building on an NFS filesystem
won't have this issue, only local filesystems.

-- 
------------------------------------------------------------------------
-------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad
Scientist


_______________________________________________
Bug-make mailing list
address@hidden
http://lists.gnu.org/mailman/listinfo/bug-make