Subject: Re: [boost] GGL issues - Access Violation
From: Barend Gehrels (barend_at_[hidden])
Date: 2009-11-18 09:37:09
>>> I wouldn't call this test stressing the library to its limits. This
>>> operation is on two fairly ordinary small scale isovists (visibility
>>> polygons). As for reversed inputs being invalid, this doesn't make
>>> sense in the context that the negation operation creates a reverse
>>> of a polygon. This is how the boolean difference is defined (at
>>> least one way). How do you do the difference operation in GGL?
>> This is explained on that web-page I referred to yesterday, and
>> touched in previous discussions with Luke. Summary: it is not yet
>> available as a separate function but the operation itself is
>> supported by the algorithm, we have to walk through the input with
>> iterators, sometimes forward, sometimes reverse.
> My question must have confused you. I wasn't asking how the algorithm
> works. I'm asking how you have tested the difference operation, i.e.
> what methods do you call and what are the forms of the inputs? The
> documentation says that difference is supported.
I answered: "it is not yet available as a separate function". OK, new
more concrete answer: I tested it the way you did, so with reverse
polygons, I copied them in my testprogram.
Where says the documentation actually that difference is implemented? I
think it states that it is supported by the algorithm, but not
implemented as a separate function... Like I said above.