[mkgmap-dev] routing bug after r3602

2015-07-27 Thread Minko
Hi, On the Dutch osm forum a routing bug was reported, http://forum.openstreetmap.org/viewtopic.php?id=31714 It is this route in Amsterdam: http://osrm.at/dR7 I have traced it down and found out that this bug is introduced between mkgmap-r3602 and r3616, the car routing now makes a detour via

Re: [mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Carlos Dávila
El 27/07/15 a las 11:26, Gerd Petermann escribió: Hi Carlos, these ways are removed because WrongAngleFixer finds that they are too short, means, all points are rounded to the same Garmin Map point. For the ids listed below this looks wrong. I am not able to reproduce the messages with the

Re: [mkgmap-dev] routing bug after r3602

2015-07-27 Thread Gerd Petermann
Hi Minko, thanks for reporting this. I can reproduce the wrong routing with r3623, but only with option --housenumbers active. Seems that the new housenumber code has an unwanted impact on routing. I'll try to find out what happens. Gerd Date: Mon, 27 Jul 2015 13:18:41 +0200 From:

Re: [mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Gerd Petermann
Hi Carlos, the server can not be reached: http://mapas.alternativaslibres.es/55114007.o5m Please try to upload to http://files.mkgmap.org.uk/ Gerd Date: Mon, 27 Jul 2015 14:16:23 +0200 From: cdavi...@orangecorreo.es To: mkgmap-dev@lists.mkgmap.org.uk Subject: Re: [mkgmap-dev] Question

Re: [mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Carlos Dávila
El 27/07/15 a las 18:12, Gerd Petermann escribió: Hi Carlos, the server can not be reached: http://mapas.alternativaslibres.es/55114007.o5m Server was down this afternoon for a few hours, sorry Please try to upload to http://files.mkgmap.org.uk/ Done:

Re: [mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Gerd Petermann
Hi Carlos, thanks, I was now able to reproduce the problem. It was introduced with a change in r3180. Before this change ways which were clipped at tile boundaries so that only one point was left were rejected much earlier. With r3624 I've now restored that old behaviour, the messages from

[mkgmap-dev] Commit: r3624: reject ways without tags and ways with less than two points again as before r3180

2015-07-27 Thread svn commit
Version mkgmap-r3624 was committed by gerd on Tue, 28 Jul 2015 reject ways without tags and ways with less than two points again as before r3180 which accidentally removed the return statement. ___ mkgmap-dev mailing list

[mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Carlos Dávila
There are several warnings in my log of the type below. I have check some of them and they all seem to be waterway=* ways. Why are these waterways (and no others) removed? Could it be avoided? (WrongAngleFixer): non-routable way 183447340 was removed (WrongAngleFixer): non-routable way

[mkgmap-dev] Commit: r3623: Only produce a warning about a missing motorway ref when the node is a highway=motorway_junction.

2015-07-27 Thread svn commit
Version mkgmap-r3623 was committed by gerd on Mon, 27 Jul 2015 Only produce a warning about a missing motorway ref when the node is a highway=motorway_junction. Carlos Dávila: Warnings about motorway exits in mkgmap.log include cases were motorway_junction is placed on motorway_link ways.

Re: [mkgmap-dev] Question about non-routable ways removed by WrongAngleFixer

2015-07-27 Thread Gerd Petermann
Hi Carlos, these ways are removed because WrongAngleFixer finds that they are too short, means, all points are rounded to the same Garmin Map point. For the ids listed below this looks wrong. I am not able to reproduce the messages with the default style. Please let me know more details.