Re: Time stamp resolution

[Jason Pearce]

Johan Bezem wrote:

> A few ideas:
> - Are all machines in a network, sync'ed by NTP or comparable? 
> Especially when using remote storage like mounted systems from 
> different hosts? Or is everything local on a single machine?

Originally the project was runnning over NFS, and this was my first port 
of call. So I relocated it to /tmp, which is definitely on a local disk. 
There was no change. That machine was running NTPD and was reporting 
itself as within 20ms of the NFS server anyway, and I was not getting 
any clock skew warnings from make. I also tried another  Linux machine 
which is not running NTPD, again I get identical behaviour.

Can someone please try the following for me

[~]$ touch foo1 ; touch foo2 ; ls -lrt foo*
-rw-r--r--    1 jasonp   None            0 Jan 24 16:01 foo1
-rw-r--r--    1 jasonp   None            0 Jan 24 16:01 foo2
[~]$ touch foo2 ; touch foo1 ; ls -lrt foo*
-rw-r--r--    1 jasonp   None            0 Jan 24 16:01 foo2
-rw-r--r--    1 jasonp   None            0 Jan 24 16:01 foo1

Now thats from CYGWIN and it works as expected. From the other machines 
I get this

england4% touch foo2 ; touch foo1 ; ls -lrt foo*
-rw-r--r--    1 jasonp       games           0 Jan 24 17:04 foo2
-rw-r--r--    1 jasonp       games           0 Jan 24 17:04 foo1
england4% touch foo1 ; touch foo2 ; ls -lrt foo*
-rw-r--r--    1 jasonp       None           0 Jan 24 17:04 foo2
-rw-r--r--    1 jasonp       None           0 Jan 24 17:04 foo1

NOT what I'd expect.

> - IIRC there was an issue with NT FAT using 2-second increments, in 
> stead of 1-sec. Don't know if that applies. I had some similar issues 
> years back.
The Cygwin machine its actually NT 5.0, but that IS the machine that 
works, unexpectedly it the real UNIX machines giving me problems!

> HTH,
>
> Johan
>
> Jason Pearce wrote:
>
> > I am having problems with a build system based on Make. It is quite a 
> > complex issue but I have traced it to what I believe to be a time 
> > stamp issue. I would really appreciate some help.
> > ...