Bug (?) in performance branch
Hi, the performance branch does not find any bounds information for a few ways and nodes. It's about 10 ways per tile where this happens. Having on look on them some are well explainable because they are located in the North Sea. But others are located completely within a country (e.g. in the middle of France or Dutch). From my understanding they must be assigned at least with the country information beause these countries do not have holes. The following patch creates a GPX file for each node/way that is not assigned with any bounds information. This makes it quite easy to check where the errorneous ways/nodes are located. WanMil Index: src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java =================================================================== --- src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (revision 2234) +++ src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (working copy) @@ -12,6 +12,7 @@ */ package uk.me.parabola.mkgmap.reader.osm; +import java.util.Collections; import java.util.HashSet; import java.util.Iterator; import java.util.List; @@ -24,6 +25,7 @@ import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryQuadTree; import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryUtil; import uk.me.parabola.util.EnhancedProperties; +import uk.me.parabola.util.GpxCreator; public class LocationHook extends OsmReadingHooksAdaptor { private static final boolean PRINT_RESULT = false; @@ -157,6 +159,8 @@ if (elem instanceof Node){ Node node = (Node) elem; tags = search(node.getLocation()); + if (tags == null) + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"n"+elem.getId(), Collections.singletonList(((Node) elem).getLocation())); } else if (elem instanceof Way){ Way way = (Way) elem; @@ -181,8 +185,10 @@ break; } } - if (tags == null) + if (tags == null) { + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"w"+way.getId(), way.getPoints()); ++cntwayNotFnd; + } } if (tags == null){
Hi, with my test data, all of them were ways that had no points within the savers bounding box. Is it okay to ignore those? Gerd WanMil wrote
Hi,
the performance branch does not find any bounds information for a few ways and nodes. It's about 10 ways per tile where this happens.
Having on look on them some are well explainable because they are located in the North Sea. But others are located completely within a country (e.g. in the middle of France or Dutch). From my understanding they must be assigned at least with the country information beause these countries do not have holes.
The following patch creates a GPX file for each node/way that is not assigned with any bounds information. This makes it quite easy to check where the errorneous ways/nodes are located.
WanMil
Index: src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java =================================================================== --- src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (revision 2234) +++ src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (working copy) @@ -12,6 +12,7 @@ */ package uk.me.parabola.mkgmap.reader.osm;
+import java.util.Collections; import java.util.HashSet; import java.util.Iterator; import java.util.List; @@ -24,6 +25,7 @@ import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryQuadTree; import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryUtil; import uk.me.parabola.util.EnhancedProperties; +import uk.me.parabola.util.GpxCreator;
public class LocationHook extends OsmReadingHooksAdaptor { private static final boolean PRINT_RESULT = false; @@ -157,6 +159,8 @@ if (elem instanceof Node){ Node node = (Node) elem; tags = search(node.getLocation()); + if (tags == null) + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"n"+elem.getId(), Collections.singletonList(((Node) elem).getLocation())); } else if (elem instanceof Way){ Way way = (Way) elem; @@ -181,8 +185,10 @@ break; } } - if (tags == null) + if (tags == null) { + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"w"+way.getId(), way.getPoints()); ++cntwayNotFnd; + } }
if (tags == null){ _______________________________________________ mkgmap-dev mailing list mkgmap-dev@.org http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
-- View this message in context: http://gis.19327.n5.nabble.com/Bug-in-performance-branch-tp5535442p5535456.h... Sent from the Mkgmap Development mailing list archive at Nabble.com.
Ok, I've checked the bounding box too and got the same results like you. It's ok to ignore them because they will not appear in the resulting map. There is one little exceptions: A way having no point within the bounding box but intersecting the bounding box will appear in the map. But the cases are quite rare (~1 way in 10 tiles) so we can ignore that. WanMil
Hi,
with my test data, all of them were ways that had no points within the savers bounding box. Is it okay to ignore those?
Gerd
WanMil wrote
Hi,
the performance branch does not find any bounds information for a few ways and nodes. It's about 10 ways per tile where this happens.
Having on look on them some are well explainable because they are located in the North Sea. But others are located completely within a country (e.g. in the middle of France or Dutch). From my understanding they must be assigned at least with the country information beause these countries do not have holes.
The following patch creates a GPX file for each node/way that is not assigned with any bounds information. This makes it quite easy to check where the errorneous ways/nodes are located.
WanMil
Index: src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java =================================================================== --- src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (revision 2234) +++ src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (working copy) @@ -12,6 +12,7 @@ */ package uk.me.parabola.mkgmap.reader.osm;
+import java.util.Collections; import java.util.HashSet; import java.util.Iterator; import java.util.List; @@ -24,6 +25,7 @@ import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryQuadTree; import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryUtil; import uk.me.parabola.util.EnhancedProperties; +import uk.me.parabola.util.GpxCreator;
public class LocationHook extends OsmReadingHooksAdaptor { private static final boolean PRINT_RESULT = false; @@ -157,6 +159,8 @@ if (elem instanceof Node){ Node node = (Node) elem; tags = search(node.getLocation()); + if (tags == null) + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"n"+elem.getId(), Collections.singletonList(((Node) elem).getLocation())); } else if (elem instanceof Way){ Way way = (Way) elem; @@ -181,8 +185,10 @@ break; } } - if (tags == null) + if (tags == null) { + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"w"+way.getId(), way.getPoints()); ++cntwayNotFnd; + } }
if (tags == null){ _______________________________________________ mkgmap-dev mailing list mkgmap-dev@.org http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
-- View this message in context: http://gis.19327.n5.nabble.com/Bug-in-performance-branch-tp5535442p5535456.h... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi WanMil,
Date: Sun, 4 Mar 2012 19:40:31 +0100 From: wmgcnfg@web.de To: mkgmap-dev@lists.mkgmap.org.uk Subject: Re: [mkgmap-dev] Bug (?) in performance branch
Ok, I've checked the bounding box too and got the same results like you. It's ok to ignore them because they will not appear in the resulting map.
There is one little exceptions: A way having no point within the bounding box but intersecting the bounding box will appear in the map. But the cases are quite rare (~1 way in 10 tiles) so we can ignore that.
Okay. It should also be possible to calculate a "virtual" point, e.g. the middle of the intersection of the bounding box of that way and the saver bounding box. Gerd
WanMil
Hi,
with my test data, all of them were ways that had no points within the savers bounding box. Is it okay to ignore those?
Gerd
WanMil wrote
Hi,
the performance branch does not find any bounds information for a few ways and nodes. It's about 10 ways per tile where this happens.
Having on look on them some are well explainable because they are located in the North Sea. But others are located completely within a country (e.g. in the middle of France or Dutch). From my understanding they must be assigned at least with the country information beause these countries do not have holes.
The following patch creates a GPX file for each node/way that is not assigned with any bounds information. This makes it quite easy to check where the errorneous ways/nodes are located.
WanMil
Index: src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java =================================================================== --- src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (revision 2234) +++ src/uk/me/parabola/mkgmap/reader/osm/LocationHook.java (working copy) @@ -12,6 +12,7 @@ */ package uk.me.parabola.mkgmap.reader.osm;
+import java.util.Collections; import java.util.HashSet; import java.util.Iterator; import java.util.List; @@ -24,6 +25,7 @@ import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryQuadTree; import uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryUtil; import uk.me.parabola.util.EnhancedProperties; +import uk.me.parabola.util.GpxCreator;
public class LocationHook extends OsmReadingHooksAdaptor { private static final boolean PRINT_RESULT = false; @@ -157,6 +159,8 @@ if (elem instanceof Node){ Node node = (Node) elem; tags = search(node.getLocation()); + if (tags == null) + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"n"+elem.getId(), Collections.singletonList(((Node) elem).getLocation())); } else if (elem instanceof Way){ Way way = (Way) elem; @@ -181,8 +185,10 @@ break; } } - if (tags == null) + if (tags == null) { + GpxCreator.createGpx(GpxCreator.getGpxBaseName()+"w"+way.getId(), way.getPoints()); ++cntwayNotFnd; + } }
if (tags == null){ _______________________________________________ mkgmap-dev mailing list mkgmap-dev@.org http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
-- View this message in context: http://gis.19327.n5.nabble.com/Bug-in-performance-branch-tp5535442p5535456.h... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
participants (3)
-
Gerd Petermann -
GerdP
-
WanMil