[boost] Re: Swapability of named objects?

From: Daniel Spangenberg (dsp_at_[hidden])
Date: 2003-11-04 02:49:28

"Hurd, Matthew" schrieb:

> Hi Daniel,
>
> I think the non-copyable semantics is appropriate as the mutex may be a
> truly non-copyable OS related thing.
>
> There is no reason why you can't have a reference or pointer to a mutex and
> that can be swapped amongst your objects.
>
> Does that solve your issue?

Hello Matthew,

no, it doesn't. The problem is not, that I could not find any workaround for that
and I am also not preaching to introduce a copy semantic for those objects.
Please be aware that Swapable does not necessary imply Copyable.

My concerns are, that I think, that a Swapability property of named_objects
seems natural, because they provide **additional** state **despite** its
synchronization state, comparable to the difference between usual named and
unnamed namespaces.
This additional state can be viewed as part of the observable state of its
(hosting)
synchronee (or whatever the correct naming is...), so I think that a Swapability
property seems reasonable.
I don't think, that Swapability should be provided for unnamed synch. objects,
although I believe that it would be a realisable issue - Can someone
conform/disprove
this?

Greetings,

Daniel