On Fri, Mar 13, 2015 at 06:02:51PM +0100, Olivier Goffart wrote: > On Friday 13 March 2015 09:08:35 Thiago Macieira wrote: > > It's a minor SIC and I still call it justified. Anyone who was using #ifndef > > was making assumptions about Q_ASSERT's internals and how it is declared. > > We should all agree that making assumptions is a bad idea. > > It was documented: > > http://doc.qt.io/qt-5/qtglobal.html#Q_ASSERT > "It does nothing if QT_NO_DEBUG was defined during compilation." > "doing nothing" doesn't necessarily imply "ignoring its arguments".
> > As for adding a new macro... I don't want to replace every one of them in > > Qt. > > > > $ git submodule --quiet foreach "git grep Q_ASSERT -- '*.cpp' '*.mm' '*.h' > > || true" | wc -l > > 9390 > > Which can still be done in one line. > it's still insane. lots of integrations, history noise, and uglier code. > But we don't even need to do it. We can leave the old one as it, and use > Q_ASSERT_UNUSED for the new asserts > still ugly. i suggest to introduce a #define which forces the old behavior. _______________________________________________ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
