logo separator

[mkgmap-dev] how to deal with "Attention: Tile contains both drive-on-left (303) and drive-on-right roads" ?

From Ticker Berkin rwb-mkgmap at jagit.co.uk on Wed Jul 10 19:19:34 BST 2019

Hi Alexis

On Wed, 2019-07-10 at 16:58 +0200, Alexis Huxley wrote:
> Hi, I'm using mkgmap r4287 and trying to process continent files from
> geofabrik, e.g.  http://download.geofabrik.de/asia-latest.osm.pbf.
> 
> For about the last six months (today included) I have 
> "drive-on-left-and-drive-on-right" errors for either Africa or Asia:
> 
> I call splitter like this:
> 
> 	java -Xmx380000M -jar splitter.jar --max-threads=40 -
> -mapid=38610001 --search-limit=1000000 /u/ahux/tmp-dir/as/asia
> -latest.osm.pbf

Wouldn't default value (auto) for max-threads be better
Maybe you need --max-nodes={some value lower than 1600000 to fix the
road problem later}

> 
> The splitter output looks okay, though I note it ends:
> 
> 	Number of MapFailedExceptions: 1

you should check out this error

> 	Number of ExitExceptions: 0
> 
> and then I call mkgmap like this:
> 
> 	java -Xmx380000M -jar mkgmap.jar --code-page=1252 --latin1 -
> -route --remove-short-arcs --road-name-pois --add-pois-to-areas -
> -index --location-autofill=bounds --tdbfile --max-jobs=40 -
> -mapname=38610001 --country-abbr=AS --country-name="Asia" --family
> -id=598 --family-name="OSM_AS" --overview-mapname=as --series
> -name="OSM Asia (2019 07)" --description="OSM_AS" --generate
> -sea=multipolygon --bounds=/u/ahux/tmp-dir/as/bounds-latest.zip -
> -precomp-sea=/u/ahux/tmp-dir/as/sea-latest.zip --style
> -file=/u/ahux/tmp-dir/as/mkgmap/distrib/mkgmap
> -r4287/examples/styles/default /u/ahux/tmp-dir/as/3861*.osm.pbf

--road-name-pois - probably not needed anymore.
--location-autofill=is_in,nearest better (bounds doesn't do anything).
--max-jobs is better without a value.
--country_name/abbr best removed as you are covering multiple
countries.
--mapname should be set per tile - see later.
--description ditto.
--generate-sea remove; you are using --precomp-sea

instead of .../3861*.osm.pbf, use the template file generated by the
splitter with -c ..../template.args


> and get a *lot* of these:
> 
> 	SEVERE (StyledConverter): /u/ahux/tmp-dir/as/38610086.osm.pbf:
> Attention: Tile contains both drive-on-left (71473) and drive-on
> -right roads (51)

I'd expect this from asia & africa; there are a mix of driving sides

> 
> and a few other assorted messages like:
> 
> 	SEVERE (MapFailedException): /u/ahux/tmp
> -dir/as/38610555.osm.pbf: (thrown in RoadDef.writeRgnOffsets())
> Overflow of the NET1. The tile (/u/ahux/tmp-dir/as/38610555.osm.pbf)
> must be split so that there are fewer roads in it

you need to experiment with splitter max-nodes to fix this

> 	SEVERE (RestrictionRelation): /u/ahux/tmp
> -dir/as/38610173.osm.pbf: Turn restriction (no_straight_on) 
> http://www.openstreetmap.org/relation/9335098 (at 
> http://www.openstreetmap.org/?mlat=-6.405884&mlon=106.829270&zoom=17)
> via node is not a routing node

Mostly can be ignored. Sometimes these are wrong in OSM, but often they
express a more complicated situation than can be expressed in the
garmin routing format. 
> 
> How should I respond to these messages? I mean: can I use mkgmap's
> output anyway? Or should I try to get upstream data fixed in OSM? (if
> yes: how can I turn these error messages into a location?) Or should
> I modify my calls to mkgmap or splitter?
> 
> Oh, also another small question: is there any convention/rule/list
> for '--mapname=NNNNNNNN' or '--family-id=NNN' numbers, or should just
> make them up? (which what I'm currently doing).

Generally make up the first 4 digits, then append 0001 as splitter -
-mapid and it will generate the --mapname in the template file.
I use the same 4 digits as the --family-id

> Can anybody offer any advice? Many thanks!
> 
> Alexis

Hope this helps
Ticker

> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev


More information about the mkgmap-dev mailing list