November 2012 - mkgmap-dev - The mkgmap lists

type=building relations
by Thorsten Kukuk 23 Nov '12

23 Nov '12

Hi, I see now more and more type=building relations. A very prominent example is relation 4599, the scottish parliament. My problem: the buildings itself have only building=yes, all other informations are part of the relation. Is it possible to write one rule which copys all missing attributes from the relation to the buildings itself? I can do that with some dozens rules for all tags I think are important, but that would become very error prune. Thanks, Thorsten -- Thorsten Kukuk, Project Manager/Release Manager SLES SUSE LINUX Products GmbH, Maxfeldstr. 5, D-90409 Nuernberg GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg)

1 0

splitter r241 and mkgmap r2378
by GerdP 22 Nov '12

22 Nov '12

Hi all, I've committed spltter r241: - added parameter --problem-report. If keep-complete is true this parm can be used to specify the output file for the generated problem list. The planet-xxx.kml files are no longer written. With r240 I started to implement a version info, but this doesn't work yet. @Steve: please can you review the build script? In the jar file I see two files: mkgmap-version.properties contains the correct svn.version splitter-version.properties has the correct name, but svn.version is "unknown" mkgmap r2378 is in the o5m_support branch. It allows to read o5m files. If you want to try o5m format with splitter and mkgmap, the tool chain should look similar to this: (download xxx.osm.pbf) osmconvert --drop-version xxx.osm.pbf -o=xxx.o5m java [JVM_opts] -jar splitter-problem-list-r241.jar [options] --output=o5m xxx.o5m java [JVM_opts] -jar mkgmap-o5m_support-r2378.jar -c templates.args or java [JVM_opts] -jar mkgmap-o5m_support-r2378.jar *.o5m A few numbers: I have a niedersachsen.osm.pbf with ca. 138Mb osmconvert requires ~ 10 seconds to write niedersachsen.o5m with ca. 173Mb splitter-problem-list-r241 a) input pbf, output pbf, keep-complete=false, overlap=2000 : 50 seconds, ~ 121 MB output b)input pbf, output pbf, keep-complete=true, overlap=0 : 117 seconds, ~ 118 MB output c) input o5m, output o5m, keep-complete=false, overlap=2000 : 22 seconds, ~ 181 MB output d) input o5m, output o5m, keep-complete=true, overlap=0 : 49 seconds, ~ 175 MB output mkgmap runs more or less with the same speed, no matter if o5m or pbf is used. Ciao, Gerd -- View this message in context: http://gis.19327.n5.nabble.com/splitter-r241-and-mkgmap-r2378-tp5736852.html Sent from the Mkgmap Development mailing list archive at Nabble.com.

2 5

Commit: r2379: Make build scripts take account of the product name
by svn commit 21 Nov '12

21 Nov '12

Version 2379 was commited by steve on 2012-11-21 16:24:50 +0000 (Wed, 21 Nov 2012) Make build scripts take account of the product name when writing svn.version. Also add build.timestamp.

1 0

splitter r239
by GerdP 21 Nov '12

21 Nov '12

Hi all, I've committed r239 now. - r237 and r238 fixed o5m read and write issues r239: - don't calculate bbox for all ways of a multipolygon rel, instead try to find closed polygons and calculate the bbox for each polygon. - don't treat type=boundariy relation like multipolygon Relation - add --output=simulate for testing, this allows to run the split process without writing the OSM output files. - make output better readable I think the --keep-complete is now working good enough, please let me know if anything is wrong or missing. Next on my todo list: - add/complete javadoc - update the doc file intro.txt - add o5m support to mkgmap - make sea generator compatible with new osmwriter interface - try to use thread(s) for the OSM reader - try to calculate reasonable value for max-areas based on available heap, number of tiles and file size -- View this message in context: http://gis.19327.n5.nabble.com/splitter-r239-tp5736433.html Sent from the Mkgmap Development mailing list archive at Nabble.com.

7 26

Flooded norway tile with latest splitter, OSM data correct?
by Thorsten Kukuk 21 Nov '12

21 Nov '12

Hi, with the current splitter/mkgmap, I have a flooded tile in norway. The reason seems to be multipolygon 96565, "Randsforden". The mulitpolygon is complete, but it is tagged as "natural=coastline". But it is inside of Norway, so no real coastline in my opinion. I think this shouldn't be "natural=coastline" but "natural=water". My questions are only: - should this be changed? I think yes. - why does this flood the tile? Thorsten -- Thorsten Kukuk, Project Manager/Release Manager SLES SUSE LINUX Products GmbH, Maxfeldstr. 5, D-90409 Nuernberg GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg)

3 3

Commit: r2375: revert wrongly added branch
by svn commit 21 Nov '12

21 Nov '12

Version 2375 was commited by gerd on 2012-11-21 09:54:51 +0000 (Wed, 21 Nov 2012) revert wrongly added branch

1 0

o5m_support branch for mkgmap
by GerdP 21 Nov '12

21 Nov '12

Sorry, I wanted to add a new branch using trunk, but I added the whole mkgmap project. Can this be reverted? Gerd -- View this message in context: http://gis.19327.n5.nabble.com/o5m-support-branch-for-mkgmap-tp5736781.html Sent from the Mkgmap Development mailing list archive at Nabble.com.

1 0

Commit: r2374: branch for the o5m format
by svn commit 21 Nov '12

21 Nov '12

Version 2374 was commited by gerd on 2012-11-21 09:40:42 +0000 (Wed, 21 Nov 2012) branch for the o5m format

1 0

Inconsistent search function
by Carlos Dávila 20 Nov '12

20 Nov '12

An user of my Central America & Carib map [1] reported some problems searching for POI's. I've been testing it and found some inconsistency in the way search function works between MapSource/BaseCamp, Legend HCx and nuvi 300. MapSource: if I search for a POI called "La Bodeguita" (under all POI or Food & Drink categories) it finds La Bodeguita [2] and La Bodeguita del Medio [3] (no information about city/country is displayed). If I search for a POI called "Bar2", it finds "Bar2 del Malecon" [4] (no information about city/country is displayed). Under the City tab, if I introduce "C" in country box, the following countries are displayed: Cayman Islands (CYM), COL (Colombia), Colombia (COL), Costa Rica (CRI), Country, CRI (Costa Rica), CUB (Cuba), Cuba (CUB), CYM (Cayman Islands). BaseCamp: searching for "La Bodeguita" returns: La Bodeguita del Medio, Ciudad de la Habana, CUB and La Bodeguita, DOM. Bar2 returns Bar2 del Malecon, Ciudad de la Habana, CUB. HCx: If I search for "La Bodeguita", only La Bodeguita is found (no city/country). Searching for Bar2, it returns "Bar2 del Malecon, Ciudad de la Habana". nuvi 300: it finds: La Bodeguita del Medio, Ciudad de la Habana; La Bodeguita (no city/country); Bar2 del Malecon, Ciudad de la Habana. If I try to find an address, and type "C" at the "Spell country" prompt, only Costa Rica and Country can be found. Results are the same with MapSource or mkgmap generated gmapsupp. None of the objects found has any is_in:* or addr:* information, so city/country must be gathered from admin boundaries. I use default mkgmap:place rules in my style. Map is generated with the following options (among others): --bounds=bounds --latin1 --code-page=1252 --area-name=$MAPA --family-name="OpenStreetMap $MAPA" --series-name="OSM-$MAPA" --index --road-name-pois=0x640a --add-pois-to-areas --link-pois-to-ways --location-autofill=is_in ($MAPA=Central America) Any idea why these differences happen and how to fix them? Why some POI/Countries are not found? [1] http://mapas.alternativaslibres.es/OpenStreetMap_Central-America.exe [2] http://www.openstreetmap.org/browse/node/596635668 [3] http://www.openstreetmap.org/browse/node/355338535 [4] http://www.openstreetmap.org/browse/node/1849671117

3 9

motorway_link lacks oneway check
by Chris66 18 Nov '12

18 Nov '12

Hi, the "motorway_link lacks oneway" check in default style is very useful in finding routing errors like this one: <http://osrm.at/1Kq> oneway=no was removed some days ago -> oneway=yes is implied. But the check is incomplete because some values are missing. Here the corrected lines (I suggest the enable by default): # Validation-like checks (comment to disable) highway=motorway_link & oneway!=yes & oneway!=no & oneway!=1 & oneway!=-1 & oneway!=true & oneway!=false { echo "motorway_link lacks oneway" } Chris

3 2