Hello Paul,

I ran your code (twice to make sure) and in both cases it returned :

last-modified times do not conform to:
http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_08

Cheers,

Bertrand

On Fri, Jan 23, 2015 at 5:27 PM, Paul Eggert <eggert@cs.ucla.edu> wrote:

> Thanks, the key problem is the fourth time the breakpoint is hit, where
> the output in your log looks like this:
>
> Breakpoint 3, Fverify_visited_file_modtime (buf=27021317) at fileio.c:5338
> 5338      if (timespec_cmp (mtime, b->modtime) == 0
> (gdb) p mtime
> $7 = {
>   tv_sec = 1422018451,
>   tv_nsec = 132061000
> }
> (gdb) p b->modtime
> $8 = {
>   tv_sec = 1422018451,
>   tv_nsec = 40059000
> }
>
> The two time stamps should be the same, but the nanoseconds component
> differ (the tv_sec components are the same, which is why we don't observe
> any bugs in the strace output).
>
> I suspect a bug in the file system.  Can you please try running the
> attached program, with the current directory being that file system?  It
> should take about 2 seconds.  You might try running it twice (it may depend
> on whether the file already exists).
>