[Patch] Coastline validator doesn't report some problems and is slow

[GerdP]

What steps will reproduce the problem?

1. open file coastlines-samples.osm.zip
2. press Shift+V to run validator

What is the expected result?

All but the two ways in the top right box should be flagged

What happens instead?

Some ways are not flagged.

Please provide any additional information below. Attach a screenshot if possible.

The unpatched code is also rather complex and slow when the data contains lots of coastline ways
as the example for #13385 (y.osm.bz2)
The attached patch fixes the problems and makes the test much faster.

URL:http://josm.openstreetmap.de/svn/trunk
Repository:UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
Last:Changed Date: 2016-09-06 00:16:07 +0200 (Tue, 06 Sep 2016)
Build-Date:2016-09-05 22:21:00
Revision:10966
Relative:URL: ^/trunk

Identification: JOSM/1.5 (10966 en) Windows 10 64-Bit
Memory Usage: 2301 MB / 5461 MB (2032 MB allocated, but free)
Java version: 1.8.0_102-b14, Oracle Corporation, Java HotSpot(TM) 64-Bit Server VM
Screen: \Display0 1920x1080
Maximum Screen Size: 1920x1080
VM arguments: [-XX:StartFlightRecording=name=MyRecording2,settings=d:\dbg\gerd.jfc, -XX:FlightRecorderOptions=defaultrecording=true,dumponexit=true,dumponexitpath=e:\ld\perf_20160907_163056.jfr]
Dataset consistency test: No problems found

Plugins:
+ FastDraw (32930)
+ apache-commons (32699)
+ buildings_tools (32796)
+ ejml (32680)
+ geotools (32813)
+ jts (32699)
+ o5m (32857)
+ opendata (32898)
+ pbf (32865)
+ poly (32699)
+ reverter (32796)
+ utilsplugin2 (32815)

Map paint styles:
+ https://josm.openstreetmap.de/josmfile?page=Styles/Lane_and_Road_Attributes&zip=1

[Don-vip]

Thanks for the patch! It would be great to add a unit test on the same model as MultipolygonTestTest.testMultipolygonFile with your sample file.

[GerdP]

OK, I also thought about that.

[GerdP]

Please review. I am not sure about the unit test setup. The coastline test uses Main.getLayerManager().getEditLayer() so
it needs an edit layer, but I have no idea why I see the messages below (similar problem in MultipolygonTestTest):

Sep 10, 2016 4:30:22 PM org.openstreetmap.josm.Main info
INFO: GET ​http://api06.dev.openstreetmap.org/api/capabilities -> 200 (391 B)
Sep 10, 2016 4:30:22 PM org.openstreetmap.josm.Main info
INFO: OK
Sep 10, 2016 4:30:23 PM org.openstreetmap.josm.Main info
INFO: GET https://josm.openstreetmap.de/maps -> 200
Sep 10, 2016 4:30:23 PM org.openstreetmap.josm.Main info
INFO: GET https://josm.openstreetmap.de/wiki/StartupPage -> 200
Sep 10, 2016 4:30:26 PM org.openstreetmap.josm.Main info
INFO: Keystroke shift pressed ESCAPE is already assigned to org.openstreetmap.josm.actions.relation.RecentRelationsAction$1@c3d81ef, will be overridden by org.openstreetmap.josm.actions.relation.RecentRelationsAction$1@6a47a37a

[Don-vip]

In 10991/josm:

fix #13570 - Coastline validator doesn't report some problems and is slow (patch by Gerd Petermann, modified)

[Don-vip]

thank you! I removed the need for a layer, and refactored the code using Java 8 features to avoid copy&paste

[GerdP]

Fine!
I am not happy with the solution reg. downloadedArea, see also #10638.
My understanding is that it can be done once in visit(), see new patch. The effect on performance is near zero, it just looks clearer to me.

Besides that I still see the delay when I start the unit test in Eclipse. Do you see a chance to change that? On my machine the delay is ~ 5 seconds, and the time is added to the first executed unit test, so it somehow messes up the reported duration times.

[simon04]

Milestone renamed