Bug report
9 tests in test_os.UtimeTests fail on DragonFly BSD:
FAIL: test_utime (test.test_os.test_os.UtimeTests.test_utime)
self.assertAlmostEqual(st.st_atime, atime_ns * 1e-9, delta=1e-6)
AssertionError: 4.005006 != 1.002003 within 1e-06 delta
4.005006 is the modification time which the test sets. HAMMER2 does not store the access time, and os.stat() returns the modification time instead. On tmpfs on the same system both are correct:
| file system |
st_atime |
st_mtime |
| HAMMER2 |
4.005006 |
4.005006 |
| tmpfs |
1.002003 |
4.005006 |
Affected are test_utime, test_utime_by_indexed, test_utime_by_times, test_utime_decimal, test_utime_dir_fd, test_utime_directory, test_utime_fd, test_utime_fraction and test_utime_nofollow_symlinks -- all of them use the same helper.
The tests already probe the file system for the timestamp resolution (support_subsecond()), so they can probe it for the access time too, and only check the modification time if it is not stored.
Linked PRs
Bug report
9 tests in
test_os.UtimeTestsfail on DragonFly BSD:4.005006 is the modification time which the test sets. HAMMER2 does not store the access time, and
os.stat()returns the modification time instead. On tmpfs on the same system both are correct:st_atimest_mtimeAffected are
test_utime,test_utime_by_indexed,test_utime_by_times,test_utime_decimal,test_utime_dir_fd,test_utime_directory,test_utime_fd,test_utime_fractionandtest_utime_nofollow_symlinks-- all of them use the same helper.The tests already probe the file system for the timestamp resolution (
support_subsecond()), so they can probe it for the access time too, and only check the modification time if it is not stored.Linked PRs