$include_dir="/home/hyper-archives/geometry/include"; include("$include_dir/msg-header.inc") ?>
Subject: Re: [geometry] Problems performing boolean operations on polygons with shared edges
From: Barend Gehrels (barend_at_[hidden])
Date: 2016-04-10 17:36:00
Hi Oleg,
Op 10-4-2016 om 22:48 schreef Oleg Evseev:
> > First, the upcoming 1.61 release will have changes in overlay. 
> Please check there if the problem still appears.
>
> Ok, I will try it, thank you! (Forgot to mention, but I think you had 
> understand that I'm using 1.60.0 release)
>
> > Second, the JPG's are illustrative, but hard to reproduce the 
> problem. Could you file a ticket with the WKT representations of two 
> polygons where the union fails?
>
> I create ticket https://svn.boost.org/trac/boost/ticket/12118
Thanks
>
> > WKT can easily be generated with the library using std::cout << 
> boost::geometry::wkt(input1), etc
>
> I used boost::geometry::dsv before, WKT return floats with more 
> decimals, but not enough, points that enclose those "almost zero 
> areas" are the same, so that area is totally empty, not "almost" (it 
> is just an edge).
> Is there any way to print floats in WKT representation with more 
> precision?
Sure, you can use the normal way, std::setprecision(20) for example, 
followed by the wkt.
>
> I wrote in the ticket, but want to repeat here:
> I forgot to mention that despite the fact that I use same initial data 
> (vehicle path) every program restart, boost geometry union_ makes 
> described problems not in same places.
> So WKT representation in the ticket different from pictures in 
> previous post.
>
> https://pp.vk.me/c631319/v631319687/1f012/2e1MycdOkB8.jpg now in last 
> program run is what it should to be - only one polygon: 
> https://pp.vk.me/c633617/v633617687/1cacb/mI45Se8H4OI.jpg
>
> Appended mpolygon is 
> https://pp.vk.me/c633617/v633617687/1cab7/r_7Ep30w_jc.jpg
> And union result now is 
> https://pp.vk.me/c633617/v633617687/1cac1/bc2qTKY-8Bs.jpg
OK, I will look at the ticket later
Regards, Barend