[lfs-support] e2fsprogs test failures on version SVN-20120806

Lewis Pike phaselocker at gmail.com
Thu Aug 16 18:45:54 PDT 2012

Bruce Dubbs wrote:
> Yes.  That's due to a change in timezone data and adding leap
> seconds in glibc.  The tests have not caught up with the changes.
> Look at the output of the tests:
> -File /lost+found (inode #11, mod time Mon Sep 20 03:26:36 1993)
> +File /lost+found (inode #11, mod time Mon Sep 20 03:26:18 1993)
> Notice the time difference of 18 seconds?  I think that's the number
> of leap seconds between Jan 1 1970 and September 1993.

Digging a bit deeper into the test suits logs does indeed reveal a
difference of 18 seconds on the time stamps.  I checked and 18 leap
seconds did occur between the aforementioned dates.

I tried duplicating the test errors by building and testing
e2fsprogs-1.42.5 on a separate system with glibc-2.16.0 and
tzdata2012c installed.  In this case however, all the tests passed.
Why might these tests fail on one system but pass on another if the
versions of e2fsprogs, glibc, and tzdata are the same?

More information about the lfs-support mailing list