|
Boost : |
From: Martin Wille (mw8329_at_[hidden])
Date: 2003-11-18 09:16:39
Joel de Guzman wrote:
> Martin Wille wrote:
>
>>>I'd welcome naming suggestions, "alternative_thread_specific_ptr"
>>>is a bit unwieldy.
>>>
>>>Is there interest to make alternative_thread_specific_ptr<>
>>>a part of Boost.Thread? If so, what would be the procedure
>>>(it is a small addition to an existing library, so I don't
>>>know wether a formal review is required)?
>
>
> 'fcourse you know I'm interested. As before, we can always use
> Spirit as the test-bed before the code migrates to its actual destination.
> We need the feature and we need it now ;)
Yes. However, I think this is of general use and not related
to Spirit. In fact, Spirit's code could become a lot simpler
when this is moved to Boost.Thread (the object_with_id stuff
would likely be easier to implement without the header-only
requirement. And the (ugly) code that is used for managing
the grammar-instances would become a lot simpler when we don't
have to hack around thread key limitations (grammar objects
would simply use an alternative_thread_specific_ptr instead
of all the object_with_id stuff, shared_ptr stuff, and ownership
hackery the current code uses (which in turn triggered the
need to workaround several broken compilers)))
Regards,
m
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk