Re: [boost] Boost.Polygon issues

Subject: Re: [boost] Boost.Polygon issues
From: Barend Gehrels (barend_at_[hidden])
Date: 2011-03-07 14:40:05

Hi Luke,

Thanks for your explanations.

On 7-3-2011 2:19, Simonson, Lucanus J wrote:
> Your test case proved quite helpful in nailing the issue down, thank you for posting it.

It is now running correctly.

> [...]
>
> There are several things I could do to address the problem, but for now I have opted to disable the feature of using the 45-degree booleans algorithm from the arbitrary angle polygon set interface since it is merely an optimization.  I will update the documentation accordingly.

To me this sounds as a good solution. Why try to cast a polygon of which 
you don't know if it is 45 or 90 degree (sounds not probable to me in 
most cases)? If it is, people should use those specific concepts. If it 
is not, try to cast is waste of time (in the majority of the cases).

> I would encourage you to modify your test to use the polygon_45_set instead of polygon_set so that you can use that algorithm as a reference as you intended.

It is just by chance that this testset was 45 degrees, but there are 
also other possibilities. I took this test because it is the only test 
with integer coordinates that I have, it now has 90 or 45 degree but 
that does not mean that it cannot be used for other angles, any angle.

Regards, Barend

-- 
Barend Gehrels
http://about.me/barendgehrels