Re: PATCH: md5.c invalid use of void expression

[Bruce Korb]

>> 
>> There's no particular reason for having a nested call there anyway.
> 
> But there's also no particular reason why we should avoid it, unless we
> know more about which environment is broken if we don't avoid it, since
> as written, it is standards-compliant code.

P.S.  Just because code is standards compliant doesn't make it easy to
follow.  Disentangling function calls makes code easier to follow.
The only place where I'd put a function call in an argument list is
where you are using an accessor function.  "memcpy" is not doing that.
Instead, it is returning its first argument "as a convenience".
That has two consequences:  1) the code is somewhat more cluttered, and
2) GCC complains.  Rather forcefully.  Hence, the patch.