check-0.10.0

[Fernando de Oliveira]

​http://downloads.sourceforge.net/check/check-0.9.14.tar.gz

​http://downloads.sourceforge.net/check/NEWS-README-0.10.0

or

​http://sourceforge.net/p/check/code/HEAD/tree/trunk/NEWS

or

​http://sourceforge.net/p/check/mailman/check-announce/?viewmonth=201508

[Check-announce] check-0.10.0 released
From: Branden Archer <b.archer4@gm...> - 2015-08-02 19:59:37

Greetings,

Today, check-0.10.0 is released. In addition to a few bug fixes and
improvements, the handing of Check when compiled without fork() has changed
slightly. Several API calls in the past would intentionally result in
errors when they required fork() to make sense. However this has been
changed to instead ignore the call. This should help improve unit test
interoperability between *nix and Windows.

Thanks,
Branden

Sun Aug 2, 2015: Released Check 0.10.0
  based on r1217 (2015-08-02 19:21:14 +0000)

* CMake on MinGW and MSVC was unable to find time related types because
  time.h was not included. This header is now included for the checks.
  Patch #53.

* If the test runner process catches a SIGTERM or SIGINT signal the running
  tests are now also killed.
  Patch #52.

* If Check is compiled without support for fork(), the behavior of
  functions which require fork() to be useful have been changed.
  Functions that attempt to set CK_FORK mode are no-ops,
  check_fork() returns in failure, and check_waitpid_and_exit()
  exits in failure.

* Add space around operators in assert messages for readability.
  Bug #102.

* Use mkstemp() if available instead of tmpfile() or tempnam().
  Patch #51.

* Fix issue with string formatting in ck_assert(), where using
  the % operator would be interpreted as a string formatter. Bug #96.

* In nofork mode, the location of a failed assertion within a test
  case was lost if that test case has a checked teardown fixture
  (even if that fixture function is empty). This is now fixed.
  Bug #99

[Fernando de Oliveira]

Bruce, you are going to do this ticket for LFS. Perhaps it would be better that you did it also for BLFS.

I'm having problems with the Check checks. And I can't understand very well, because it seems some suites are run several times.

I will attach my check checks log (compressed, 8K, because it is over 100k, uncompressed)

Please, if you don't mind, I would prefer you taking this ticket.

[Fernando de Oliveira]

make -k check log

[bdubbs@…]

OK, I'll update this when I update LFS.

[bdubbs@…]

I'll probably update this later tonight after I update LFS.

I built a fresh LFS with this package (and a few other new packages), entered chroot, and rebuilt check using the BLFS procedures.

I had no failures.

PACKAGE=check-0.10.0
DEST=/tmp/check

rm -rf $DEST
rm -rf $PACKAGE

tar -xf $PACKAGE.tar.?z*
cd $PACKAGE

./configure --prefix=/usr --disable-static
make || exit 1
make check

[Fernando de Oliveira]

I have:

$ xzgrep -EB1 "Failures: 5|Failures: 1" check-0.10.0-make-k-check-2015.08.03-15h07m41s.log.xz
 Check Servant2
27%: Checks: 147, Failures: 59, Errors: 48
--
Running suite(s): Fix Sub
0%: Checks: 1, Failures: 1, Errors: 0
--
 Check Servant2
27%: Checks: 147, Failures: 59, Errors: 48
--
Running suite(s): Setup Fail
0%: Checks: 1, Failures: 1, Errors: 0
--
Running suite(s): Setup Fail Nofork
0%: Checks: 1, Failures: 1, Errors: 0
--
Running suite(s): Setup Fail Nofork 2
0%: Checks: 1, Failures: 1, Errors: 0
--
Running suite(s): Teardown Fail
0%: Checks: 1, Failures: 1, Errors: 0
--
Running suite(s): Teardown Fail No Fork
0%: Checks: 1, Failures: 1, Errors: 0

[bdubbs@…]

Indeed, I have six entries like:

Running suite(s): Fix Sub
0%: Checks: 1, Failures: 1, Errors: 0

but my log indicates they are supposed to fail. If it didn't, we would get a message: "Test run counts not correct for checked setup failure"

[bdubbs@…]

Fixed at revision 16324.

[Fernando de Oliveira]

I've attached a log because I didn't understand Still don't understand:

First, do you have any Failures and Errors of the order of 50?

As I mentioned above and below:

27%: Checks: 147, Failures: 59, Errors: 48

The results below can be considered good?

$ xzgrep -EC2 "Failures: 5|Failures: 1" check-0.10.0-make-k-check-2015.08.03-15h07m41s.log.xz
check.c:506: Bad status in set_fork_status
 Check Servant2
27%: Checks: 147, Failures: 59, Errors: 48
check_check_sub.c:12:F:Simple Tests:test_lno:0: Failure expected
check_check_sub.c:20:E:Simple Tests:test_mark_lno:0: (after this point) Early exit with return value 1
--
check_check_fork.c:34:P:Core:test_nofork_pid:0: Passed
Running suite(s): Fix Sub
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:16:S:Fix Sub:unchecked_setup:0: Test failure in fixture
Running suite(s): Check Servant
--
check.c:506: Bad status in set_fork_status
 Check Servant2
27%: Checks: 147, Failures: 59, Errors: 48
check_check_sub.c:12:F:Simple Tests:test_lno:0: Failure expected
check_check_sub.c:20:E:Simple Tests:test_mark_lno:0: (after this point) Early exit with return value 1
--
check_check_fixture.c:96:P:Fixture Norm Sub:test_sub_ch_setup_norm:0: Passed
Running suite(s): Setup Fail
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:130:S:Setup Fail:test_sub_fail:0: Failed setup
Running suite(s): Setup Fail Nofork
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:130:S:Setup Fail Nofork:checked_setup:0: Failed setup
Running suite(s): Setup Fail Nofork 2
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:130:S:Setup Fail Nofork 2:checked_setup:0: Failed setup
Running suite(s): Setup Pass Multiple fixtures
--
check_check_fixture.c:348:P:Fixture Two setups:test_sub_two_setups:0: Passed
Running suite(s): Teardown Fail
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:135:S:Teardown Fail:test_sub_pass:0: Failed teardown
Running suite(s): Teardown Fail No Fork
0%: Checks: 1, Failures: 1, Errors: 0
check_check_fixture.c:135:S:Teardown Fail No Fork:test_sub_pass:0: Failed teardown
Running suite(s): Teardown Sig