logo separator

[mkgmap-dev] highway=track with no road_class and road_speed set breaks routing ## was ## "No roads near target" bug in Schwabmünchen

From GerdP gpetermann_muenchen at hotmail.com on Thu Apr 11 20:56:36 BST 2013

Hi Franco,

okay, I've committed the patch. I assume that the remaining problem is
related to the housenumber 
code, and Bernds idea that it maybe be related to the roundabout sounds
reasonable.

Gerd


franco_bez wrote
> 
> GerdP wrote
>> Hi Franco,
>> 
>> I did a few more tests now. I am able to reproduce the problem with a
>> small input file:
>> franco1.osm.pbf
>> <http://gis.19327.n5.nabble.com/file/n5756490/franco1.osm.pbf>  
>> 
>> My result with an Oregon 450t , software version 5.50:
>> version 2447: routing to POI "Schönwetter Automobile" works, but 
>> address "Robert-Bosch-Straße 14" doesn't work.
>> 
>> in all later versions both did not work. 
>> 
>> The change in r2448 was the remove-short-arc patch.
>> One effect of this patch was that the order of roads and other ways did
>> change,
>> esp. the other ways were written first, roads later.
>> In r2447, they were written in order of appearance in the osm data
>> (because 
>> preserve-element-order=true is used)
>> I've coded a small patch that changes the logic so that roads are always
>> placed before other ways,
>> and that seems to solve the problem for the POI search.
>> roads_first_v1.patch
>> <http://gis.19327.n5.nabble.com/file/n5756490/roads_first_v1.patch>  
>> 
>> The compiled binary is here:
>>  http://files.mkgmap.org.uk/download/103/mkgmap.jar
>> 
>> With --x-housenumbers this version supports searching for addresses "a
>> little bit":
>> If you type housenumber  1 instead of 14, it shows you a list with two
>> numbers (3 and 5),
>> and routing to these works. It works also if you directly type 3 or 5 in
>> the search dialog.
>> 
>> I did not yet look at the overlay problem, maybe tomorrow.
>> 
>> Ciao,
>> Gerd
> Hi Gerd,
> I can confirm that the routing to the POI works perfectly with your
> patched Version.
> 
> I can also confirm that the routing to the address is still broken.
> 
> Ciao,
>   Franco





--
View this message in context: http://gis.19327.n5.nabble.com/No-roads-near-target-bug-in-Schwabmunchen-tp5753364p5756716.html
Sent from the Mkgmap Development mailing list archive at Nabble.com.


More information about the mkgmap-dev mailing list