On Sun, Jan 17, 2010 at 09:59:46AM +0100, Andreas Schwab wrote:
Michael Stone <address@hidden> writes:
It seems that touch -a does update ctime on btrfs, invalidating one of the
assumptions behind this test and causing it to fail.
Did you mean "does not update ctime"?
Yes. My eyes are starting to cross. I thought I had duplicated something
manually, but now I can't. So scratch what I said in the original
message. I keep getting test failures like this:
failed ls ctime test -- this failure is expected at least for SunOS4.1.4
and for tmpfs file systems on Solaris 5.5.1.
In the output below, `c' should have had a ctime more recent than
that of `a', but does not.
-rw-rw-r-- 1 mstone mstone 0 2010-01-17 14:22:47.702311166 +0000 a
-rw-rw-r-- 2 mstone mstone 0 2010-01-17 14:22:45.698811694 +0000 c
So I guess the problem is that ln isn't updating ctime. (On linux
2.6.31 + btrfs.)
+ touch -m -d '1998-01-15 23:00' a
+ touch -m -d '1998-01-15 22:00' b
+ touch -m -d '1998-01-15 21:00' c
+ touch -a -d '1998-01-14 13:00' c
+ touch -a -d '1998-01-14 12:00' b
+ sleep 2
+ touch -a -d '1998-01-14 11:00' a
++ ls -c a c
+ set c a
+ test 'c a' = 'a c'
+ fail=1
Mike Stone