Transitive const sucks

[James Dennett]

Mike Capp wrote:
> James Dennett Wrote:
> 
>> Well, in C++ you'd be fine, but in D I thought the behavior
>> if you modified something after casting away const was
>> undefined, so you're out of luck.  (Not that you'd need
>> to cast away const for this in C++.)
> 
> You'll hit undefined behavior in C++ as well if the data
> being modified (via a const_cast) was originally declared
> as const.

Something of a mantra around here!  However, that's certainly
not the case we are talking about.  We're talking about situations
where an object (which is not changing) has a reference to another
(mutable) object and should be able to change that object.

> Compilers can (and some PPC compilers did) put
> such data in read-only memory, and even if they don't they
> tend to assume "really, truly const" and optimize accordingly.

Indeed.  Non-mutable parts of objects defined as const in C++
cannot be modified by any well-defined code, and optimizers can
and do use that fact to good effect.  It's just not what this
(sub-)thread was about.

-- James