Assaf Gordon wrote: > Jim Meyering wrote, On 12/17/2010 05:07 AM: >> Here's a preview of what should soon appear as coreutils-8.8. [...] >> Any testing you can perform over the weekend would be most welcome. >> > On CentOS 5.4, du/bigtime fails (in a reproducible manner). > > $ uname -a > Linux XXXXXX 2.6.18-164.11.1.el5 #1 SMP Wed Jan 20 07:32:21 EST 2010 > x86_64 GNU/Linux > > ====================================================== > GNU coreutils 8.7.66-561f8: tests/test-suite.log > ====================================================== > > 1 of 1 test failed. > > .. contents:: :depth: 2 > > > FAIL: du/bigtime (exit: 1) > ========================== > > --- out 2010-12-17 15:29:06.000000000 +0000 > +++ exp 2010-12-17 15:29:06.000000000 +0000 > @@ -1 +1 @@ > -4 9223372036854775807 future > +0 9223372036854775807 future
Thanks for the speedy testing and report. This patch should fix it: >From bad74f92a9024e01d6b2bd4aadb765231e4db565 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyer...@redhat.com> Date: Fri, 17 Dec 2010 16:55:39 +0100 Subject: [PATCH] tests: avoid spurious du/bigtime failure * tests/du/bigtime: Map nonzero block count to 0. Reported by Assaf Gordon. * THANKS: Update. --- THANKS | 1 + tests/du/bigtime | 5 +++++ 2 files changed, 6 insertions(+), 0 deletions(-) diff --git a/THANKS b/THANKS index 9bd78c8..b17f4bf 100644 --- a/THANKS +++ b/THANKS @@ -62,6 +62,7 @@ Arnold Robbins arn...@skeeve.com Arthur Pool p...@commerce.uq.edu.au Arun Sharma arun.sha...@intel.com Arvind Autar autar...@planet.nl +Assaf Gordon assafgor...@gmail.com Augey Mikus mi...@dqc.org Aurelien Jarno aure...@debian.org Austin Donnelly austin.donne...@cl.cam.ac.uk diff --git a/tests/du/bigtime b/tests/du/bigtime index 0a7f32b..d515947 100755 --- a/tests/du/bigtime +++ b/tests/du/bigtime @@ -39,6 +39,11 @@ printf "0\t$bignum\tfuture\n" > exp || framework_failure_ printf "du: time $bignum is out of range\n" > err_ok || framework_failure_ du --time future >out 2>err || fail=1 + +# On some systems an empty file occupies 4 blocks. +# Map the number of blocks to 0. +sed 's/^[0-9][0-9]*/0/' out > k && mv k out + compare out exp || fail=1 compare err err_ok || fail=1 -- 1.7.3.4
