Propagating failing unit tests to Qt Creator issues-pane with QTestLib

Borkr

I recently switched from CppUnit to QTestLib for my projects, and I find making tests is faster and easier. However, I have two concerns. The first being as stated in the title. Using CppUnit I managed to get failing unit tests listed in issues-pane of Qt Creator, which also allowed me to click on it to get to the failing line. How do I set this up with Qt Creator? It is a little hideous to read the Application output/Compile output to find the failing unit tests (even though said output contains the exact same link I wish to have in Issues pane). Test runner is shown below.

The other concern is more of a nicety, which I guess is what made test development with CppUnit so laborious: auto registering test classes (and registering them.. So many macros).
Is there a way to do this? As of now I have to manually include their header in my my test runner, as shown below.
This solution is not scalable....

#include "unitmain.h"

#include "firstunittest.h"
#include "secondunittest.h"
#include "thirdunittest.h"

int UnitMain::unitMain(int argc, char *args[])
{
    int status = 0;
    auto runTest = [&](QObject * testObject)
    {
        status |= QTest::qExec(testObject, argc, args);
        delete testObject;
    };

    runTest(new FirstUnitTest);
    runTest(new SecondUnitTest);
    runTest(new ThirdUnitTest);
    // repeat for all unit tests in the project...
    return status;
}

int main(int argc, char *args[])
{
    return UnitMain().unitMain(argc, args);
}

SGaist

Hi and welcome to devnet,

AFAIK, it's a feature missing from Qt Creator. You can check that on the bug report system. If you don't find anything, you can open a new feature request for it.

The general Qt way of unit testing is to have each test as an independent application. I don't see a really easy way to bring them all in only one executable code wise.

Borkr

Hi, and thank you! :-)

I'll see what I can find in the issue tracker. As for the test, do you mean that every single instance of a unit test is run in a separate executable? If I have 100+ unit tests, I end up with 100+ executables?

EDIT:
It seems this was proposed a long time ago
https://bugreports.qt.io/browse/QTCREATORBUG-86

SGaist

Yes, each unit test gets its own executable. You can check Qt's sources, it's working quite well.

Thanks for pointing to the request.