[mkgmap-dev] Patch v1: check if road is connected

Sat Mar 16 09:30:09 GMT 2013

Hi,

I finally found a case which creates such a message:
The way 
http://www.openstreetmap.org/browse/way/126623972
is part of mp-rel 
http://www.openstreetmap.org/browse/relation/1715809

So, WanMil guessed right, it is probably the mp code that creates these
ways.

The mp code creates a new way 4611686018427396419 for 
126623972 and thus the strange message appears.

The mp-created way has the tags
mkgmap:stylefilter=polyline
mkgmap:mp_created=true
area=yes
highway=service

The original way is not connected to any road, so I think it is correct to
make this way 
"not rouable". 
One question remains: Would it be possible to print a better message?
I think about adding a tag with the id(s) of the OSM object(s) that were
used to create the
mp-way. Would that make sense?

Gerd

WanMil wrote
>> Hi WanMil,
>>
>>
>> WanMil wrote
>>> process-exits does not create non connected ways. Otherwise it's a bug.
>>>
>>> I am quite sure the ways are created by the mp algorithm.
>>
>> maybe. That would mean that the mp algo creates a way and the style
>> makes it a road.
>> I think that's not very likely.
>> If the way is created by the process-exit method, I assume that the exit
>> is
>> on or near to a
>> boundary.
>>
>> Gerd
>>
> 
> b = boundary point
>  >> = way
> . = point
> 
>  >>.>>b
> 
> Is one of the ways mentioned as unconnected?
> 
> I think ways connecting or crossing the boundary should not be 
> identified as unconnected.
> In case the original way is connected a way created by process-exit is 
> *always* connected to another way or touches/crosses the boundary. It's 
> a bug if that's not the case.
> 
> WanMil
> 
> _______________________________________________
> mkgmap-dev mailing list

> mkgmap-dev at .org

> http://lists.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

--
View this message in context: http://gis.19327.n5.nabble.com/Patch-v1-check-if-road-is-connected-tp5751822p5753373.html
Sent from the Mkgmap Development mailing list archive at Nabble.com.