Warning references bad object ID
Hi folks, I periodically generate a map of Ireland from the Geofabrik ireland.osm extract. With warnings enabled, I get the following: 2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Multipolygon http://www.openstreetmap.org/browse/relation/4611686018427389788 contains errors. 2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Polygon 4611686018427393761(6P)(4611686018427392539[6P]) carries role inner but lies inside an inner polygon. Potentially its role should be outer. Clearly there is no such object ID. Can anybody identify where this is going wrong? Thanks, Dermot -- -------------------------------------- Iren sind menschlich
I dont know the internal of the multipolygon code, but this number is in hex 400000000000121B So I think the upper bit is used for somthing. If I mask it out I will get 121B and this will be 4635 in decimal. Could this be the faulty relation? Regards, Johann Dermot McNally schrieb:
Hi folks,
I periodically generate a map of Ireland from the Geofabrik ireland.osm extract. With warnings enabled, I get the following:
2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Multipolygon http://www.openstreetmap.org/browse/relation/4611686018427389788 contains errors. 2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Polygon 4611686018427393761(6P)(4611686018427392539[6P]) carries role inner but lies inside an inner polygon. Potentially its role should be outer.
Clearly there is no such object ID. Can anybody identify where this is going wrong?
Thanks, Dermot
On Sun, Jun 13, 2010 at 10:05:16PM +0200, Johann Gail wrote:
I dont know the internal of the multipolygon code, but this number is in hex 400000000000121B
So I think the upper bit is used for somthing. If I mask it out I will get 121B and this will be 4635 in decimal. Could this be the faulty relation?
The high-order bit is set on objects generated by the multipolygon processor. I guess that the low-order bits are just a running number (count of generated objects). Are there any MultiPolygon messages before that one? One possible improvement to the error messages would be to display the location of an arbitrary node of the erroneous way, if the original way ID is not available. Marko
On 14 June 2010 06:34, Marko Mäkelä <marko.makela@iki.fi> wrote:
The high-order bit is set on objects generated by the multipolygon processor. I guess that the low-order bits are just a running number (count of generated objects). Are there any MultiPolygon messages before that one?
No, unfortunately those are the only warnings related to multipolygons.
One possible improvement to the error messages would be to display the location of an arbitrary node of the erroneous way, if the original way ID is not available.
That would certainly help. Clearly, it would be good if mkgmap could point to the right object, but my immediate concern is to find and fix the bad data. Dermot -- -------------------------------------- Iren sind menschlich
On 13 June 2010 21:05, Johann Gail <johann.gail@gmx.de> wrote:
So I think the upper bit is used for somthing. If I mask it out I will get 121B and this will be 4635 in decimal. Could this be the faulty relation?
Hi, There's no sign of this number being the relation in question. In addition, I tried the same trick on all three of the overlong numbers, to no effect. Could this be a signed-versus-unsigned problem? Dermot -- -------------------------------------- Iren sind menschlich
Hi folks,
I periodically generate a map of Ireland from the Geofabrik ireland.osm extract. With warnings enabled, I get the following:
2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Multipolygon http://www.openstreetmap.org/browse/relation/4611686018427389788 contains errors. 2010/06/13 20:39:27 WARNING (MultiPolygonRelation): ireland.osm: Polygon 4611686018427393761(6P)(4611686018427392539[6P]) carries role inner but lies inside an inner polygon. Potentially its role should be outer.
Clearly there is no such object ID. Can anybody identify where this is going wrong?
Thanks, Dermot
Hi, the multipolygon with id 4611686018427393761 is an artificial polygon created by the --generate-sea=multipolygon option. There is definitely no chance for the multipolygon code to get the original object ID. This must be fixed in generate-sea code. Have fun! WanMil
On 14 June 2010 18:50, WanMil <wmgcnfg@web.de> wrote:
the multipolygon with id 4611686018427393761 is an artificial polygon created by the --generate-sea=multipolygon option. There is definitely no chance for the multipolygon code to get the original object ID. This must be fixed in generate-sea code.
Ahhhh! That would explain it. In that case, I may need to find some bad coastline. Thanks for the tip! Dermot -- -------------------------------------- Iren sind menschlich
On 14 June 2010 18:50, WanMil<wmgcnfg@web.de> wrote:
the multipolygon with id 4611686018427393761 is an artificial polygon created by the --generate-sea=multipolygon option. There is definitely no chance for the multipolygon code to get the original object ID. This must be fixed in generate-sea code.
Ahhhh! That would explain it. In that case, I may need to find some bad coastline. Thanks for the tip!
Dermot
There is an easy rule how to detect that. The mp-code always displays the polygon-id followed by a list of way ids surrounded by brackets (). If the way-ids are a long number starting with 461168.... the problem is caused by the generate-sea multipolygon. Example: 4611686018427393761(6P)(4611686018427392539[6P]): Generate-sea 4611686018427393769(18P)(1049929[6P], 1049945[12P]): Mp created a polygon consisting of the way-ids 1049929 and 1049945 WanMil
participants (4)
-
Dermot McNally -
Johann Gail -
Marko Mäkelä -
WanMil