@tygre

Quote:

I worked for a while with the JCK, the Java compliance tests... every Java compilers must pass them to be called "Java"...

SO... if a Java compiler fails JCK, does that mean it can't do "jack"?

Quote:

Of course, these tests do not prevent the use of null references and such

Neither do Ada's FWIW, because Ada still has the "nil" pointer. However, Ada is designed in such a way that the ordinary programmer rarely has to use pointers. -- Well, in my very limited experience. Also, that's more true of Ada 12 than of Ada 83 I think. When s/he does use pointers, there are many safeguards to prevent dangling pointers, and even then you can create them. (I've done it.)

As I understand it, only Eiffel and Kotlin make a big deal about their null safety. (Well, also Rust and SPARK but we've already talked about that.) Eiffel calls it Void safety and their version is supposedly more robust than Kotlin's. I don't remember the details; it was a conversation on comp.lang.eiffel a couple of years ago. It's too bad Bertrand Meyer decided to invent Yet Another Programming Language and trying to wage war on Ada instead of going to, say, the Ada Rapporteur Group and suggesting that they implement his idea of Design by Contract (which they eventually did anyway, and in the SPARK dialect it's A Really Big Deal).

Quote:

And, actually, they (used to?) include plenty of "special cases" because the "baseline implementation" is wrong (either the Sun/Oracle Java compiler or JVM) but it becomes the "standard" to be compatible...

Rust aficionados take note: that's what happens when a company controls the standard.

_________________
I've decided to follow an awful lot of people I respect and leave AmigaWorld. If for some reason you want to talk to me, it shouldn't take much effort to find me.