[
https://issues.apache.org/jira/browse/LUCENE-8227?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16418092#comment-16418092
]
Karl Wright edited comment on LUCENE-8227 at 3/28/18 9:15 PM:
--------------------------------------------------------------
So, the following is true:
(1) All polygon points are in the hemisphere y>=0
(2) The bounds point we think is right is on the edge plane, precision 1.0e-17
(3) The bounds point we think is right is in the hemisphere y>=0
(4) But, the membership sided plane that rejects the bounds point we think is
right accepts the antipodes point, which is in the y < 0 part of the world
(5) All edges are properly constructed so that each endpoint is within the
other endpoint's membership sided plane
The conclusion is that the sector of the edge plane that is not contributing
properly towards the bounds is at least partly in the y<0 space. But since the
two endpoints are both in the y>=0 space, this makes no sense, because any arc
that connects the two would have to remain in the y>=0 space as well.
Adding the following code to the GeoComplexPolygon constructor demonstrates
that nothing odd is happening with edge construction:
{code}
if (startPoint.y < 0.0 || endPoint.y < 0.0) {
System.out.println("Found start/end in y < 0 space");
throw new IllegalStateException("blah");
}
{code}
No exception is thrown, so there are no unexpected points being included
somehow. So the min x point is being excluded because it really is in the
wrong sector. The membership bounds sided planes divide the space into four
sections, and only one is legal; the excluded bounds point therefore cannot be
on the right part of the sector in question. So we're looking in the wrong
place; this polygon edge cannot contribute to the min/max.
was (Author: [email protected]):
So, the following is true:
(1) All polygon points are in the hemisphere y>=0
(2) The bounds point we think is right is on the edge plane, precision 1.0e-17
(3) The bounds point we think is right is in the hemisphere y>=0
(4) But, the membership sided plane that rejects the bounds point we think is
right accepts the antipodes point, which is in the y < 0 part of the world
(5) All edges are properly constructed so that each endpoint is within the
other endpoint's membership sided plane
The conclusion is that the sector of the edge plane that is not contributing
properly towards the bounds is at least partly in the y<0 space. But since the
two endpoints are both in the y>=0 space, this makes no sense, because any arc
that connects the two would have to remain in the y>=0 space as well.
Adding the following code to the GeoComplexPolygon constructor demonstrates
that nothing odd is happening with edge construction:
{code}
if (startPoint.y < 0.0 || endPoint.y < 0.0) {
System.out.println("Found start/end in y < 0 space");
throw new IllegalStateException("blah");
}
{code}
No exception is thrown, so there are no unexpected points being included
somehow.
I'm stumped.
> TestGeo3DPoint.testGeo3DRelations() reproducing failures
> --------------------------------------------------------
>
> Key: LUCENE-8227
> URL: https://issues.apache.org/jira/browse/LUCENE-8227
> Project: Lucene - Core
> Issue Type: Bug
> Components: general/test, modules/spatial3d
> Reporter: Steve Rowe
> Assignee: Karl Wright
> Priority: Major
>
> Three failures: two NPEs and one assert "assess edge that ends in a crossing
> can't both up and down":
> 1.a. (NPE) From
> [https://builds.apache.org/job/Lucene-Solr-NightlyTests-master/1512/]:
> {noformat}
> [junit4] 2> NOTE: reproduce with: ant test -Dtestcase=TestGeo3DPoint
> -Dtests.method=testGeo3DRelations -Dtests.seed=C1F88333EC85EAE0
> -Dtests.multiplier=2 -Dtests.nightly=true -Dtests.slow=true
> -Dtests.linedocsfile=/home/jenkins/jenkins-slave/workspace/Lucene-Solr-NightlyTests-master/test-data/enwiki.random.lines.txt
> -Dtests.locale=ga -Dtests.timezone=America/Ojinaga -Dtests.asserts=true
> -Dtests.file.encoding=UTF-8
> [junit4] ERROR 10.4s J1 | TestGeo3DPoint.testGeo3DRelations <<<
> [junit4] > Throwable #1: java.lang.NullPointerException
> [junit4] > at
> __randomizedtesting.SeedInfo.seed([C1F88333EC85EAE0:7187FEA763C8447C]:0)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.countCrossingPoint(GeoComplexPolygon.java:1382)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.matches(GeoComplexPolygon.java:1283)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:564)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:572)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:569)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:660)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:646)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.isWithin(GeoComplexPolygon.java:370)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseMembershipShape.isWithin(GeoBaseMembershipShape.java:36)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseShape.getBounds(GeoBaseShape.java:35)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.getBounds(GeoComplexPolygon.java:440)
> [junit4] > at
> org.apache.lucene.spatial3d.TestGeo3DPoint.testGeo3DRelations(TestGeo3DPoint.java:225)
> [junit4] > at java.lang.Thread.run(Thread.java:748)
> {noformat}
> 1.b. (NPE) From
> [https://builds.apache.org/job/Lucene-Solr-SmokeRelease-7.x/184/]:
> {noformat}
> [smoker] [junit4] 2> NOTE: reproduce with: ant test
> -Dtestcase=TestGeo3DPoint -Dtests.method=testGeo3DRelations
> -Dtests.seed=F2A368AB96A2FD75 -Dtests.multiplier=2 -Dtests.locale=fr-ML
> -Dtests.timezone=America/Godthab -Dtests.asserts=true
> -Dtests.file.encoding=UTF-8
> [smoker] [junit4] ERROR 0.99s J0 | TestGeo3DPoint.testGeo3DRelations
> <<<
> [smoker] [junit4] > Throwable #1: java.lang.NullPointerException
> [smoker] [junit4] > at
> __randomizedtesting.SeedInfo.seed([F2A368AB96A2FD75:42DC153F19EF53E9]:0)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.countCrossingPoint(GeoComplexPolygon.java:1382)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.matches(GeoComplexPolygon.java:1283)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:564)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:572)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:572)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:660)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:646)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.isWithin(GeoComplexPolygon.java:370)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseMembershipShape.isWithin(GeoBaseMembershipShape.java:36)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseShape.getBounds(GeoBaseShape.java:43)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.getBounds(GeoComplexPolygon.java:440)
> [smoker] [junit4] > at
> org.apache.lucene.spatial3d.TestGeo3DPoint.testGeo3DRelations(TestGeo3DPoint.java:224)
> [smoker] [junit4] > at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> [smoker] [junit4] > at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> [smoker] [junit4] > at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> [smoker] [junit4] > at
> java.base/java.lang.reflect.Method.invoke(Method.java:564)
> [smoker] [junit4] > at
> java.base/java.lang.Thread.run(Thread.java:844)
> {noformat}
> 2. (both up&down) From
> [https://jenkins.thetaphi.de/job/Lucene-Solr-7.x-Linux/1598/]:
> {noformat}
> [junit4] 2> NOTE: reproduce with: ant test -Dtestcase=TestGeo3DPoint
> -Dtests.method=testGeo3DRelations -Dtests.seed=BACC479CC2D38CCA
> -Dtests.multiplier=3 -Dtests.slow=true -Dtests.badapples=true
> -Dtests.locale=lv -Dtests.timezone=SystemV/AST4ADT -Dtests.asserts=true
> -Dtests.file.encoding=UTF-8
> [junit4] FAILURE 3.01s J2 | TestGeo3DPoint.testGeo3DRelations <<<
> [junit4] > Throwable #1: java.lang.AssertionError: assess edge that
> ends in a crossing can't both up and down
> [junit4] > at
> __randomizedtesting.SeedInfo.seed([BACC479CC2D38CCA:AB33A084D9E2256]:0)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.countCrossingPoint(GeoComplexPolygon.java:1438)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$DualCrossingEdgeIterator.matches(GeoComplexPolygon.java:1283)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Node.traverse(GeoComplexPolygon.java:564)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:660)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon$Tree.traverse(GeoComplexPolygon.java:646)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.isWithin(GeoComplexPolygon.java:370)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseMembershipShape.isWithin(GeoBaseMembershipShape.java:36)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoBaseShape.getBounds(GeoBaseShape.java:43)
> [junit4] > at
> org.apache.lucene.spatial3d.geom.GeoComplexPolygon.getBounds(GeoComplexPolygon.java:440)
> [junit4] > at
> org.apache.lucene.spatial3d.TestGeo3DPoint.testGeo3DRelations(TestGeo3DPoint.java:224)
> [junit4] > at java.lang.Thread.run(Thread.java:748)
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
