https://bugs.kde.org/show_bug.cgi?id=408539
--- Comment #2 from RJVB <rjvber...@gmail.com> --- >And no, the assert is correct. >If we arrive here with a invalid iterator, all is broken, this shall no happen. Well, I don't agree with that. I have said it before and will continue to hammer it down: it's unprofessional to allow production builds to crash because of situations of which you have foreseen that they might happen. If the situation cannot be caught when it arrives an immediate abort is indeed the only solution rather than letting the software maybe crash at a later point. In all other situations thought should be given to users who might lose important data. Once there is sufficient proof that the situation can no longer arise the code can be cleaned up, but asserts have the nasty habit of remaining to clutter the code. I do agree that there should be a Q_ASSERT variant which makes it easy to execute a bit of graceful error handling. -- You are receiving this mail because: You are watching all bug changes.
