bug#7858: coreutils-8.9 test failed on rh5 amd

[Jim Meyering]

Jim Meyering wrote:

> sam sirlin wrote:
> ...
>> This is, I believe the relevant section:
>
> Thanks for the report.
> That is just what I needed.
>
>> FAIL: cp/cp-parents (exit: 1)
>> =============================
> ...
>> ++ ls -ld e/d/a
>> ++ cut -b-10
>> + p=drwxr-xr-x
>> + case $p in
>> + fail=1
>> ++ ls -ld g/sym
>> ++ cut -b-10
>> + p=drwxr-xr-x
>> + case $p in
>> + fail=1
>> ++ ls -ld e/d/a/b/c
>> ++ cut -b-10
>> + p=drwxr-xr-x
>> + case $p in
>> ++ ls -ld g/sym/b/c
> ...
>
> However, that test is rather old, and this is the first
> report of its failure I've seen in a long time, so I expect
> there must be something unusual about your environment.
>
> I build on AMD/rhel-6 pretty regularly and have never seen that,
> but I'm usually using local ext4 or tmpfs partitions.
>
> Can you give more detail?
> On what type of file system is your build dir?  (i.e., run df -T .)
> If it's NFS, what can you tell me about the server?

Hi Sam,

Sorry about the delay.  You replied,

  > It is indeed NFS, fairly big (2.3T), I think NAS, linux (striped down),
  > xfs filesystem.

in http://debbugs.gnu.org/cgi/bugreport.cgi?bug=7858#11
but I never received that message.

I saw it just now, only while going through old pending bug
reports via http://debbugs.gnu.org/coreutils.

Can you retest using the latest release, coreutils-8.13?

  https://lists.gnu.org/archive/html/coreutils/2011-09/msg00130.html