Re: [boost] new library (space partitioning)

Subject: Re: [boost] new library (space partitioning)
From: Barend Gehrels (barend_at_[hidden])
Date: 2010-08-16 04:04:21

hi Adam,

>> Le 15/08/2010 11:07, Adam Wulkiewicz a écrit :
>>
>>> It strongly depends on used algorithm if you need it or not. If you use
>>> some kind of internal, temporary objects you need it.
>>> See e.g.
>>> http://www.cgg.cvut.cz/members/havran/ARTICLES/ingo06rtKdtree.pdf or
>>> http://kesen.realtimerendering.com/Intel-EG07.pdf=
>>
>> I may be missing something here, but I don't see why the temporary
>> objects in question cannot be stored on the execution stack.
>
> Implementing dynamic memory allocation on the stack could be quite 
> tricky.

Thanks for the links.

In Boost.Geometry we do not allocate dynamic memory ourselves; all 
polygons, rings, lines, multi-polygons etc are stored as an std:: 
container, which can be selected at compile-time by the library user (as 
a template-template-parameter). Besides that it is not part of the 
concept so library users might select another solution.

Don't know if this is possible in this case, but if it is possible, I 
would prefer that approach.

Regards, Barend