https://bugs.kde.org/show_bug.cgi?id=400632
--- Comment #7 from Christoph Cullmann <cullm...@kde.org> --- >> As we don't use them nor do we use any libraries that emit any exceptions > Do you use the C++ standard library occasionally? We do, but as said, any exceptions that will be emitted there will always lead to program abort for us, as we can't handle them anyways. See below > I suggest to reconsider such a view once more. > Would you like to be prepared for “std::bad_alloc” (as a well-known example)? That is a good example for 1) we can't do anything then anyways, there is really no way to handle that in a sane way for a GUI application beside just die, which happens without any special handling 2) you will get that close to never, as e.g. Linux over-commits, you will just later silently segfault (and yes, that is the cruel reality, in my daily work I must always cope with that) Same holds for any of the other exceptions e.g. std::vector might throw for out-of-bounds accesses. It's perfectly fine to abort, we can't handle that anyways, if we could, we shall have already done an out-of-bound check before. (like we do in many places) > Interruptions can occur. I guess that it can become more important if > reasonable program behaviour is provided also in exceptional software > situations. => As told above: Interruptions can only for occur situations in which we don't care, as we will die anyways. We don't use exceptions for any recoverable error handling. > Does this feedback qualify to reopen the bug report? No, but you can provide patches for improvements, but please not for exceptions safety issues that just not exist given the lack of them. Bugs are not there to discuss about theoretical software engineering improvements but for concrete issues ;=) If you can point out that we have a leak, crash, ..., please open a bug for that. -- You are receiving this mail because: You are watching all bug changes.
