On Sun, Dec 18, 2011 at 12:52:25PM +0100, rupert THURNER wrote: > On Sun, Dec 18, 2011 at 12:17, Stefan Sperling <s...@elego.de> wrote: > > On Sun, Dec 18, 2011 at 02:35:39AM -0800, rupert.thurner wrote: > > > rupert @ unstable9s : ~/mgar-sav/pkg/subversion/trunk/work/solaris9- > > > sparc/build-isa-sparcv8/subversion-1.7.2/subversion/tests/cmdline > > > $ ./copy_tests.py 86 --verbose > > > > > PASS: copy_tests.py 86: changed dir after commit should match > > > checkout > > > > So this time, it passed. This is probably just a race condition in the > > test where it might not succeed if timestamps on disk aren't updated > > quickly enough. Can you get it to fail again by running it a couple > > of times? > > > > I think you can safely ignore this failure. > > > > every time i run it with the test suite it fails. from java i know that the > timestamps on solaris are 4 digits less accurate than on windows. can this > be an issue here as well?
Yes, certainly. Does this patch help? Index: copy_tests.py =================================================================== --- copy_tests.py (revision 1214870) +++ copy_tests.py (working copy) @@ -25,7 +25,7 @@ ###################################################################### # General modules -import stat, os, re, shutil +import stat, os, re, shutil, time # Our testing module import svntest @@ -4647,6 +4647,8 @@ def changed_dir_data_should_match_checkout(sbox): svntest.actions.run_and_verify_svn(None, None, [], 'co', sbox.repo_url, verify_dir) + time.sleep(1.1) # sleep for timestamps + was_cwd = os.getcwd() os.chdir(verify_dir)
