Null references redux

[Michel Fortin]

On 2009-09-30 15:30:02 -0400, "Denis Koroskin" <2korden at gmail.com> said:

> Note that C stdlib (and other libraries/bindings) will need to be 
> updated  to reflect changes, e.g.
> 
> extern(C) void*? malloc(size_t size); // may return null!
> 
> which is great because it will provide additional safety. I've seen 
> quite  a lot of code that don't test returned value against null (which 
> is a  mistake, I believe).

Which makes me think of this: pointers being non-nullable by default 
will make it easy to make mistakes when writing C bindings. A 
programmer might see this C declaration:

	void* malloc(size_t size);

and naively translate it to D like this:

	extern(C) void* malloc(size_t size);

without noticing the change in semantics.

For pointer arguments it's not much of a problem: the worse that can 
happen is that it blocks you from passing a null value when you should 
(in which case you can update the bindings). For a return value it's 
more troublesome because you're implicitly adding a promise that the 
function will not return null, and you might not realize it's wrong 
until it does indeed return null and your program crashes with a 
segfault.

Not that I think it's worth bothering too much, but it's something to 
keep in mind.

-- 
Michel Fortin
michel.fortin at michelf.com
http://michelf.com/