Re: [ublas] bypass construction interest?

From: Neal Becker (ndbecker2_at_[hidden])
Date: 2007-02-02 06:43:40

Gunter Winkler wrote:

> On Thursday 01 February 2007 19:59, Neal Becker wrote:
>> I am (once again) looking into avoiding unnecessary construction of
>> vector/matrix.  In particular, this is interesting for std::complex,
>> since (unfortunately) it's constructor is non-trivial.
>>
>> I am looking into the following.  In storage.hpp, use
>>
>> boost::has_trivial_constructor
>>
>> to dispatch.
> 
> Can this be done via a special Allocator? I think the least intrusive
> method should be to have an allocator that initilizes the new storage and
> to have one that does not do this. (The change would be to call
> alloc.contruct(..) instead of explicite use value_type(). Right?)

I don't know.  Here is what I propose:

        unbounded_array (size_type size, const ALLOC &a = ALLOC()):
            alloc_(a), size_ (size) {
            if (size_) {
                data_ = alloc_.allocate (size_);
                if (not boost::has_trivial_constructor<T>::value) { << My Addition
                  // ISSUE some compilers may zero POD here
#ifdef BOOST_UBLAS_USEFUL_ARRAY_PLACEMENT_NEW
                  // array form fails on some compilers due to size cookie, is it standard conforming?
                  new (data_) value_type[size_];
#else
                  for (pointer d = data_; d != data_ + size_; ++d)
                    new (d) value_type;
#endif
                }

            }
 
How would a special allocator bypass the construction in the above?

>> vector (size, init_value)
>> matrix (size1, size2, init_value)
> 
> The same can be done by using the copy constructor:
> 
> vector<double> v = scalar_vector<double>(size, value);
> 
> IMO the default behaviour of 'no construction' should be enough.
> 

True, although I personally prefer the extra constructor, since it is only a small
addition and consistent with familiar stl container behavior.