As a user of QTestLib, I'm concerned about the runtime effects of the unconditional toString() calls in QCOMPARE

Currently, the toString() function is called for each QCOMPARE(), regardless of whether it succeeds or not, even though the result is only needed when the comparison fails. We have tens of thousands of test cases, and the string formatting eats a significant amount of the runtime compared to the (typically trivial) cost of the comparision.

Please change QCOMPARE to only format the expressions on failure.