From: Peter Dimov (pdimov_at_[hidden])
Date: 2001-05-07 07:54:12
From: "Jesse Jones" <jesjones_at_[hidden]>
> The problem with this is that if boost::function doesn't have
> operator== users are hosed. Their only recourse is to code some sort
> of work around or use a different callback class with a huge amount
> of overlap with boost::function.
I believe that users can get around the restriction, if necessary, although
it won't be as convenient.
The problem with providing operator== for a variant class (it can be done -
I've done it) is that it starts breaking even more compilers than usual [
:-) ] and (the killer argument that convinced me to drop
variant<T>::operator== and operator<) breaks MSVC compatibility due to lack
of partial ordering.
Anyway, in order for function::operator== to work it would need an
operator== from the function object, which is rarely, if ever, defined,
-- Peter Dimov Multi Media Ltd.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk