[ 
https://issues.apache.org/jira/browse/LUCENE-8239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426789#comment-16426789
 ] 

Karl Wright edited comment on LUCENE-8239 at 4/5/18 11:53 AM:
--------------------------------------------------------------

[~ivera] The "boundary plane is off the world" case is a legit case we need to 
code for.  I'm not sure how we wind up using the Dual iterator for that case 
though, since by definition the test point or check point is *on* the pole, and 
therefore the travel in that dimension is distance zero.  We should have 
selected one of the Linear iterators instead already before we get that far.

I would like to first try and understand why we don't already handle these 
cases before applying the proposed fix.  If you have details of why the 
expected things are not happening and how the patch addresses it, I'm happy to 
go ahead without a careful review.  I've got a tight deadline today but 
tomorrow should be a little bit easier.




was (Author: [email protected]):
[~ivera] The "boundary plane is off the world" case is a legit case we need to 
code for.  I'm not sure how we wind up using the Dual iterator for that case 
though, since by definition the test point or check point is *on* the pole, and 
therefore the travel in that dimension is distance zero.  We should have 
selected one of the Linear iterators instead already before we get that far.


> GeoComplexPolygon fails when test or/and check point are near a pole
> --------------------------------------------------------------------
>
>                 Key: LUCENE-8239
>                 URL: https://issues.apache.org/jira/browse/LUCENE-8239
>             Project: Lucene - Core
>          Issue Type: Bug
>          Components: modules/spatial3d
>            Reporter: Ignacio Vera
>            Priority: Major
>         Attachments: LUCENE-8239.patch
>
>
> When calling {{within}} method in GeoComplexPolygon you can get errors if the 
> test point of the polygon or the given point is near a pole.
> The reason is that one of the planes defined by these points is tangent to 
> the world therefore intersection with the above plane fails. We should 
> prevent navigating those planes ( we should not even construct them).



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to