slides #110 to #113 are cursed: we have v, val and value, should this all be the same variable? Also, if postconditions require parameters to be const, how would you ever be able to check a non-const member function (i.e. specify that size() after push_back is the same as size()+1 before mutation?)
Not sure if this is an improvement over the straightforward assert or _ASSERT macro's. It seems to have its own set of quirks which has to be learnt separately. Adding const (1:00:28) to built in argument types was deprecated by Scott Meyers book so code has to be changed for this. Let this contract be completely removed in release builds so that one can add as much as one want without compromising performance in release ever. Visual Studio 2008 had checked iterators in release builds which suppressed inlining and compiler optimizations.
slides #110 to #113 are cursed: we have v, val and value, should this all be the same variable? Also, if postconditions require parameters to be const, how would you ever be able to check a non-const member function (i.e. specify that size() after push_back is the same as size()+1 before mutation?)
Not sure if this is an improvement over the straightforward assert or _ASSERT macro's. It seems to have its own set of quirks which has to be learnt separately. Adding const (1:00:28) to built in argument types was deprecated by Scott Meyers book so code has to be changed for this. Let this contract be completely removed in release builds so that one can add as much as one want without compromising performance in release ever. Visual Studio 2008 had checked iterators in release builds which suppressed inlining and compiler optimizations.
1:13:23 John Lakos’ talk from the same conference: th-cam.com/video/VCwC1cvP8i0/w-d-xo.html
1:13:50 Timurs’ CppNorth keynote: th-cam.com/video/jTuIzSUxYDc/w-d-xo.html