Hi Mike, You dropped the list. That's fine if it were deliberate; just checking.
> > > >> const char *foo; > > > >> char const *foo; > > > > No, those two have identical meaning. ... > These examples are found there. > (( reformatted for emphasis )) > > void Foo( int * ptr, > int const * ptrToConst, The above `ptrToConst' is the second one I gave. > int * const constPtr, > int const * const constPtrToConst ) None of the four are the first one I gave. > const int *ptrToConst; //identical to: int const *ptrToConst, The above is the first one I gave. The comment claims it is identical to my second. I think you're having an off-day of word blindness here. :-) Just to make certain we're arguing about the same thing, I'm talking C here, not C++. I've no idea whether C++ is the same or different in this respect, and don't wish to learn. :-) > Please note that I am not trying to pick a fight here. No, me neither. But there's no harm in first working out why we disagree, even if we then agree to disagree. > Is there a path by which the confusions such as these can be > avoided in the design and implementation of C-like language? Yes, don't allow two alternatives, only one. Unfortunately, standardisation arrived too late, but going forward `const after' is consistent and logical, as I explain in one of my list emails with the sed and awk showing six varieties. If you like C, you'll like Go. Its `const' is for constants and they can be typed, `const foo int64 = 42', or untyped, `const pi = 3.141592653589793238462643383279502884197169399375105820974944592307'. (I do hope I told bc(1) to do that right.) The untyped ones are rounded, etc., when actually used for a type, e.g. `var i int = pi // 3'. This allows them to have higher precision than any native type. https://golang.org/ref/spec#Constant_declarations Cheers, Ralph.
