From: Andy Little (andy_at_[hidden])
Date: 2006-01-14 11:38:50
"Arkadiy Vertleyb" wrote in message
> Why is it heavyweight? Registering a type is just creating a couple of
> template specializations -- that's it.
> What do you mean by the "space in the typeof vector"? How is it related to
> the number of registered types?
Sorry I got confused about the using the typeof vector . Looking in to it again
registering a new type is just a specialistion on an int operation. I guess
that isnt how I remembered it.
> What do you mean by everything? I would recommend registering types and
> templates that can be present in the types used by your users.
If the time for a large amount of registrations isnt significant then I
apologise fo being misleading about it. I always understood typeof useage to be
quite heavyweight, so I was quite worried about the large number of types
automatically registered in last version of boost::mpl::math library. However
maybe I'm blaming the long compile times I'm getting on typeof, when in fact its
just mpl::vector thats to blame. Currently I am in process of removing
mpl::vector from my code to test this out.
> I would register your own types/templates leaving mpl types/templates to
> either mpl authors or users. Then again, you could wear an mpl hat and
> register its types, but I don't think this should become a part of your
Bearing in mind that I have been getting a compiler out of keys error message
recently in even small tests, which as you will understand basically renders my
library a useless resource hog, I am now trying to be extremely careful to add
as little unnecessary clutter in headers as possible. That includes not wishing
to register unneccesary types with Typeof if at all possible.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk