From: Roland Schwarz (roland.schwarz_at_[hidden])
Date: 2006-10-30 05:47:52
Anthony Williams wrote:
> No. This is required. The memory for an object with static-storage-duration
> MUST be zero initialized prior to the constructor running.
>
>> 1) "Might": the standard does not require it.
>
> Wrong.
You are obviously misunderstanding me.
Of course are static objects zero initialized. I had the impression that
the original poster meant that this would be a general case for
dynamic objects as well on some compilers. And this is waht I meant by
"might". Did you get me now?
>> 2) E.g.: MSVC initializes memory to "CDCDCDCDCDCD...." in debugging
>> builds, so there is at least one prominent case where the
>> assumption is false.
>
> It does this for automatic and heap-allocated objects (which is allowed). If
> it did it for static objects, it would be non-conforming.
Same than before.
But the problem is a different one:
Even a static zero initialized ähm memory area, is not a object until
its ctor has been run! And since my mutex proposal tries to implement
a mutex that is valid before ctor runs have been completed, only POD's
can be used.
Does this make sense?
Roland