|
Boost : |
From: Lie-Quan Lee (llee1_at_[hidden])
Date: 2001-03-26 09:41:53
At Mon, 26 Mar 2001 09:16:31 -0500,
David Abrahams wrote:
> I think you miss my point ;-)
>
> Which of the following 2 invocations is clearer?
>
> mutex_generator<checking_policy<checked>::recursion_policy<nonrecursive>::bl
> ocking_policy<blocking>::sheduling_policy<priority> >::type
>
> mutex_generator<checked, nonrecursive, blocking, priority>::type
>
> Asking people to remember the parameter names (e.g. "recursion_policy") in
> addition to the possible parameters (e.g. "nonrecursive") is no service to
That is a very good point.
> users. It should be possible to arrange for arbitrary parameter ordering and
> sensible defaults without using named template parameters here. I think that
> it's actually easier to implement as well.
As long as arbitrary parameter ordering and default parameters could
be achieved, the latter is much better :-) These two are my main
points about the coming mutex interface.
-- ************************************************************ Rich Lee (Lie-Quan) Lab for Scientific Computing Department of Computer Science & Engineering University of Notre Dame Email : llee1_at_[hidden] Tel : 219-631-3906 (Office) HomePage: http://www.cse.nd.edu/~llee1/ ************************************************************
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk