QtTestDriverTest doesn't get built nor executed and leaks

[muggenhor]

Because of a typo in tests/CMakeLists.txt QtTestDriverTest doesn't get built and thus not executed either:

if(Qt5TEST_FOUND)

This should be (note the title case instead of full upper case):

if(Qt5Test_FOUND)

When fixing that however, I'm getting a memory leak: https://travis-ci.org/muggenhor/cucumber-cpp/jobs/381075041

@konserw Given that you originally introduced the QtTest driver in #165, can you have a look at this and try to fix it?

This is the output:

108/109 Test #108: QtTestDriverTest .................................................................***Failed    1.26 sec
==10192== Memcheck, a memory error detector
==10192== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==10192== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
==10192== Command: /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest
==10192== 
SUCCESS (The test body has not been run)
SUCCESS (The test body has been run)
= Step invocation =
SUCCESS (Succeeding step)
SUCCESS (Failing step)
SUCCESS (Failing step has a message)
SUCCESS (Pending step with no description - result)
SUCCESS (Pending step with no description - description)
SUCCESS (Pending step with description - result)
SUCCESS (Pending step with description - description)
SUCCESS (Inexistent step)
= Constructors and destructors =
== Succeeding step ==
SUCCESS (Contexts created after invoke)
SUCCESS (Contexts destroyed after invoke)
SUCCESS (Contexts created after end scenario)
SUCCESS (Contexts destroyed after end scenario)
== Failing step ==
SUCCESS (Contexts created after invoke)
SUCCESS (Contexts destroyed after invoke)
SUCCESS (Contexts created after end scenario)
SUCCESS (Contexts destroyed after end scenario)
= Step failure description is reset =
SUCCESS (Failing step description is set)
SUCCESS (Failing step description is the same)
==10192== 
==10192== HEAP SUMMARY:
==10192==     in use at exit: 116 bytes in 3 blocks
==10192==   total heap usage: 734 allocs, 731 frees, 68,470 bytes allocated
==10192== 
==10192== 4 bytes in 1 blocks are definitely lost in loss record 1 of 3
==10192==    at 0x4C2B0E0: operator new(unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==10192==    by 0x60EBAF1: qrand() (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x623974B: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x62146D1: QFile::open(QFlags<QIODevice::OpenModeFlag>) (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6239DC2: QTemporaryFile::open(QFlags<QIODevice::OpenModeFlag>) (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x45F463: QTemporaryFile::open() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x45EDE1: cucumber::internal::QtTestStep::invokeStepBody() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x445536: QtTestStepDouble::invokeStepBody() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x4455E7: QtTestDriverTest::stepInvocationRunsStepBody() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x445567: QtTestDriverTest::runAllTests() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x444570: cucumber::internal::DriverTest::run() (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192==    by 0x4434BD: main (in /home/travis/build/muggenhor/cucumber-cpp/build/tests/QtTestDriverTest)
==10192== 
==10192== 56 bytes in 1 blocks are still reachable in loss record 2 of 3
==10192==    at 0x4C2AB80: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==10192==    by 0x7350B58: UDataMemory_createNewInstance_52 (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x734ECCA: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x734F629: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x734F83D: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735015B: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735C59B: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735D2B9: ucnv_getStandardName_52 (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x63252CB: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6325DDC: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6321D66: QTextCodec::codecForLocale() (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6182C40: QString::toLocal8Bit() const (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192== 
==10192== 56 bytes in 1 blocks are still reachable in loss record 3 of 3
==10192==    at 0x4C2AB80: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==10192==    by 0x7350B58: UDataMemory_createNewInstance_52 (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x734EC9C: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x734F888: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735015B: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735C59B: ??? (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x735D2B9: ucnv_getStandardName_52 (in /usr/lib/x86_64-linux-gnu/libicuuc.so.52.1)
==10192==    by 0x63252CB: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6325DDC: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6321D66: QTextCodec::codecForLocale() (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x6182C40: QString::toLocal8Bit() const (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192==    by 0x626CC38: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.2.1)
==10192== 
==10192== LEAK SUMMARY:
==10192==    definitely lost: 4 bytes in 1 blocks
==10192==    indirectly lost: 0 bytes in 0 blocks
==10192==      possibly lost: 0 bytes in 0 blocks
==10192==    still reachable: 112 bytes in 2 blocks
==10192==         suppressed: 0 bytes in 0 blocks
==10192== 
==10192== For counts of detected and suppressed errors, rerun with: -v
==10192== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)

[konserw]

Yes I know about that for some time - one user send me email that Qt test driver doesn't work for him. I've quickly fixed that typo on my branch, but then noticed the leak. I was investigating it for some time but haven't been able to fix it.
I'm afraid we have been bitten by this bug: https://bugreports.qt.io/browse/QTBUG-21298

[konserw]

I've tried to fix leak by adding qcoreapp:
master...konserw:fix-qt
And it works on my machine (Arch linux) but still fails on travis:
https://travis-ci.org/konserw/cucumber-cpp/jobs/382327922
Suggestions? shall we just suppress it?

[muggenhor]

Yes, if we're not misusing any API then suppression is fine.

[ursfassler]

Mentioned bug is fixed in Qt version 5.10.0. I can no longer reproduce the problem with Qt 5.15.10.