IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:
- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure
The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.
The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.
This patch should eliminate interemittent autobuild failures.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We can hence replace the assert after the loop by a success torture_comment.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This does not change the logic except for adding early
returns in failure cases.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:
- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure
The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.
The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.
This patch should eliminate interemittent autobuild failures.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We can hence replace the assert after the loop by a success torture_comment.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This does not change the logic except for adding early
returns in failure cases.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:
- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure
The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.
The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.
This patch should eliminate interemittent autobuild failures.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We can hence replace the assert after the loop by a success torture_comment.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This does not change the logic except for adding early
returns in failure cases.
But it makes the code more compact and obvious.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:
- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure
The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.
The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.
This patch should eliminate interemittent autobuild failures.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We can hence replace the assert after the loop by a success torture_comment.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This copes with cases where the server is very busy and
can't provide tortures more tight time scaling..
This is an attepmt to remove the flapping character of this test.
Signed-off-by: Michael Adam <obnox@samba.org>
Revieed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Sep 10 21:51:42 CEST 2014 on sn-devel-104
This is to be able to better tell what went wrong if something
went wrong. This is currently one of the main flapping tests.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This is where it belongs, and it prepares
subsequent patches.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
The fix missed one instance, as autobuild has just told me...
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Mon Aug 18 17:42:00 CEST 2014 on sn-devel-104
The write should never update the time, so the fraction of the write
time delay we use is not important.
Andrew Bartlett
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Jun 24 01:44:06 CEST 2014 on sn-devel-104
This removes the hardcoded TIMEDELAY_SECS that was then made variable
by the confusing "secs" variable
Andrew Bartlett
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The previous test was far, far too tight, it was in seconds 1/4 of the
fraction of the normal delay we had configured Samba to use so (1/4) *
(500 000 / 2000 000) = 1/16 (sec). This margin appears to just be too
tight for our loaded test server.
Andrew Bartlett
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
In particular, this avoids a comparison with
double diff = timeval_elapsed() being promoted to an integer.
Andrew Bartlett
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This only changes instances directly before a return false, ret =
false or goto fail statement.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Thu Jun 12 10:39:38 CEST 2014 on sn-devel-104
This only changes instances directly before a return false, ret =
false or goto fail statement.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Both offer the same functionality, sec_ace_equal() will be removed.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This causes flakey autobuilds from time to time.
The reduction in the time window to 30% was added in:
commit c2df97f57c
Author: Jeremy Allison <jra@samba.org>
Date: Thu Nov 5 15:37:26 2009 -0800
Fix up some of the timing constants for DELAYWRITE. Add some extra tests up test_delayed_write_update6
to investigate what happens to a sticky write handle after a second handle close.
Jeremy.
The original reduction to 75% was set in the new code
commit 0d0fddf8ae
Author: Jeremy Allison <jra@samba.org>
Date: Fri Sep 5 14:24:36 2008 -0700
Added tests that show that write time update is immediate
when changing file size using SMBwrite of size zero,
SET_END_OF_FILE, or SET_ALLOCATION_SIZE - no 2 second
delay in these cases.
Jeremy.
(This used to be commit 3aa7523d77)
and in:
commit edb3a83a06
Author: Stefan Metzmacher <metze@samba.org>
Date: Tue Apr 8 10:25:51 2008 +0200
BASE-DELAYWRITE: use timeval_* and make it possible to spefic the writetime update delay
metze
(This used to be commit 751ab2992a)
Change-Id: I8ff9fb8a8b66308f6d784351a0f43b5b234889d1
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Not fetching the latest modification time on a folder if we have read locks on it.
Prove we should just rely on the mtime value from the underlying
filesystem, even with an open handle.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=9870
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Dec 5 10:05:06 CET 2013 on sn-devel-104
Signed-off-by: Richard Sharpe <realrichardsharpe@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Feb 15 07:09:59 CET 2013 on sn-devel-104
