Ah, my bad - the dependency on FunctionUtil should not be in the main jts
module.  I'll fix that now.  Thanks for finding this!

On Wed, May 27, 2015 at 1:16 PM, Jan Tosovsky <[email protected]> wrote:

> It seems to be a quite substantial rewrite, thanks for it!
>
> I’ve tried to build it from trunk (ant executed in the root folder), but I
> am getting following errors:
>
>
>
>     [javac]
> D:\Java\components\jts-trunk\jts\java\src\com\vividsolutions\jts\nod
>
> ing\SegmentStringUtil.java:96: error: cannot find symbol
>
>     [javac]       LineString line =
> FunctionsUtil.getFactoryOrDefault(null).crea
>
> teLineString(ss.getCoordinates());
>
>     [javac]                         ^
>
>     [javac]   symbol:   variable FunctionsUtil
>
>     [javac]   location: class SegmentStringUtil
>
>     [javac]
> D:\Java\components\jts-trunk\jts\java\src\com\vividsolutions\jts\nod
>
> ing\SegmentStringUtil.java:100: error: cannot find symbol
>
>     [javac]     return
> FunctionsUtil.getFactoryOrDefault(null).createMultiLineSt
>
> ring(lines);
>
>     [javac]            ^
>
>     [javac]   symbol:   variable FunctionsUtil
>
>     [javac]   location: class SegmentStringUtil
>
>
>
> Does it compile for you? Or have I overlooked something? This is my first
> attempt to build this project...
>
>
>
> Thanks, Jan
>
>
>
> From: Martin Davis [mailto:[email protected]]
> Sent: Wednesday, May 27, 2015 6:20 PM
> To: Jan Tosovsky
> Cc: [email protected]
> Subject: Re: [Jts-topo-suite-user] Polygonizer issue
>
>
>
> Here's the link to the rev:
> https://sourceforge.net/p/jts-topo-suite/code/974/tree/
>
>
>
> On Tue, May 26, 2015 at 4:32 PM, Martin Davis <[email protected]> wrote:
>
> Further to this issue: I"ve enhanced the Polygonizer to be able to extract
> only polygons which form a valid polygonal geometry.   The algorithm can
> handle nested polygons as described below.
>
>
>
> It's in SVN at r974
>
>
>
> On Wed, May 6, 2015 at 9:19 AM, Martin Davis <[email protected]> wrote:
>
> Actually I see now that Daniel Baston's algorithm is a bit smarter, since
> it
> can handle nested polygons as well.  I think the algorithm I posted can be
> extended to have this behaviour as well - it just needs to classify
> polygons
> as holes or non-holes in the same kind of way, iterating from the outside
> in.
>
>
>
> On Wed, May 6, 2015 at 9:11 AM, Martin Davis <[email protected]> wrote:
>
> Yes, it's not surprising that code that relies heavily on intersection()
> and
> intersects() is inefficient - because those methods are so general-purpose,
> they are expensive to execute.
>
>
>
> Here's one idea for a faster algorithm (caveat - I have not implemented
> this, and there may be corner cases):
>
>
>
> - Extract all the line segments from all holes in the set of polygons
> output
> from Polygonizer, normalize them, and add them to a Set holeSeg
>
> - For each polygon, test if any segment from its shell lies in the holeSeg
> Set.  If it does, discard it.  Otherwise, that polygon is an "outer"
> polygon.
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Jts-topo-suite-user mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/jts-topo-suite-user
>
>
------------------------------------------------------------------------------
_______________________________________________
Jts-topo-suite-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jts-topo-suite-user

Reply via email to