- mkgmap-dev - The mkgmap lists

Routing Issue with Garmin Edge 830
by Stephan Warzecha 21 May '20

21 May '20

Hello, I am sorry but I have a question better called problem. my new Garmin Edge 830 ist crashing down, sometimes during navigation. I askes this question here because garmin told me that the map must be the problem. I am working with mkgmap since many years and I built "my own" map form my devices. At my old Garmin Vista Hcx the maps working without problem, also on MapSource and Basecamp. I checked the line file more than hunderet times again and again. Yesterday I produced the map only with "route" option and default style which is comming with mkgmap. The map that was built has the same problem. I used differnt typ-files to crosscheck if the problem comes from typ-file but no success. I also tried older versions of mkgmap, the same no success. I will try to explain what the problem is: Mostly ways, taged with highway=track and tracktype=grade4, are affected. They are part oft he map, and there is no problem while routing on MapSource, BaseCamp, Vista usw., but if I create the same route on the garmin edge 830, the route is built without any problems, but the violet line (route) has interruptions. Whenever I had such a part oft he way in a route the device is crashing down while driving the route. This ist the rule wich is used to create the line for tracktype=grade 4 highway=track & tracktype=grade4 {setaccess 'yes'; set mkgmap:unpaved=yes} [0x11 road_class=0 road_speed=1 resolution 23] I also tried it with different way types (0x01 - 0x12, 0x0e 0x0f etc) also I tried differnt spped and road classes. Next step was to try with all Tracks as the same line, like this... highway=track & tracktype=* {setaccess 'yes'; set mkgmap:unpaved=yes} [0x11 road_class=0 road_speed=1 resolution 23 Routing on all ways was possible, but the garmin routing line is broken at the ways, which was tracktype 4 before. Now I have no idea what is happening, the device can´t know that the line was a tracktype4, all tracks hast he same name and options. I will add some pictures to explan what im trying to explain, with my terrible english ;0)) An example route with MapSource: [cid:image005.png@01D62D1F.128DFE60] The same on Garmin Device Edge 830: [cid:image004.png@01D62D1F.D9B56C30] Same Route with my layout in Mapsource: [cid:image008.png@01D62D22.39C3CE80] And at the Edge 830: [cid:image007.png@01D62D20.3DFCCF30] Now the funny things: *Garmin Map is working, *If I delete a affeccted way in Openstreetmap and map it new, also only with 2 Tags like this: [cid:image009.png@01D62D21.72342900] ...the mistake is gone. I din´t "correct" the shown example yet. But I got this Problem all over whole germany, there must be an error. Maybe this is a personal probelm or mistake, but maybe that others has the same problems. If there is any information what Iam doing wrong or something else, please let me kown. There are so many hours Iam searching for a solution. The maps are built with extract from geofabrik, germany file splittet with splitter. I used the newest stable versions. Also the is more information you need, please lt me know. Viele Grüße / Best Regards Stephan Warzecha Logistic and Workplace Services | Logistic Support| Problem Manager Logistic Support T +49 (0) 6026 9762-6120 | F +49 (0) 6026 9762-6504 | stephan.warzecha(a)dpd.de<mailto:stephan.warzecha@dpd.de> DPD Deutschland GmbH Stockstädter Straße 10, 63762 Großostheim, Deutschland | dpd.de<http://www.dpd.de/> Besuchen Sie uns auf Facebook<http://www.facebook.com/dpdindeutschland>, Instagram <https://www.instagram.com/dpd.de/> und Twitter<http://twitter.com/dpd_de>. [cid:image002.gif@01D14D54.6725B0D0] Sitz der Gesellschaft: Aschaffenburg | Registergericht: Aschaffenburg HRB 8887 Geschäftsführung: Eric Malitzke (CEO), Dirk Müller, Thomas Ohnhaus, Andreas Reß [signature_304297842]

4 3

IS_IN tag - value as regular expression or wildcard
by Joris Bo 21 May '20

21 May '20

Hello I try to use the IS_IN function with regular expression or wildcard, but I get style-errors or false. When testing against the exact value it works perfectly. Is there a method for testing against building = * instead of building = yes ? On the same object I have the following results #leisure = pitch & is_in(building, *, all_in_or_on) = true {set debug0 = true} Style conversion error leisure = pitch & is_in(building, '.*', all_in_or_on) = true {set debug1 = true} false leisure = pitch & is_in(building, '\*', all_in_or_on) = true {set debug2 = true} false leisure = pitch & is_in(building, '*', all_in_or_on) = true {set debug3 = true} false leisure = pitch & is_in(building, '(.*)', all_in_or_on) = true {set debug4 = true} false leisure = pitch & is_in(building, '(\*)', all_in_or_on) = true {set debug5 = true} false leisure = pitch & is_in(building, '(*)', all_in_or_on) = true {set debug6 = true} false leisure = pitch & is_in(building, yes, all_in_or_on) = true {set debug7 = true} true mkgmap:cache_is_in_POLYLINE_building_(*)_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_(.*)_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_(\*)_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_*_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_.*_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_\*_all_in_or_on=false mkgmap:cache_is_in_POLYLINE_building_yes_all_in_or_on=true Many thx in advance, Kind Regards Joris jorisbo(a)hotmail.com<mailto:jorisbo@hotmail.com>

3 5

Commit r4494: fix some sonarlint issues, simplify code (no change in output expected)
by svn commit 21 May '20

21 May '20

Version mkgmap-r4494 was committed by gerd on Thu, 21 May 2020 fix some sonarlint issues, simplify code (no change in output expected) http://www.mkgmap.org.uk/websvn/revision.php?repname=mkgmap&rev=4494

1 0

Commit r4493: fix some sonarlint issues, simplify code (no change in output expected)
by svn commit 21 May '20

21 May '20

Version mkgmap-r4493 was committed by gerd on Thu, 21 May 2020 fix some sonarlint issues, simplify code (no change in output expected) http://www.mkgmap.org.uk/websvn/revision.php?repname=mkgmap&rev=4493

1 0

New User: Inconsistent results - help!
by Andrew Morris 18 May '20

18 May '20

(Sorry for the possible repeat, the previous message was sent from the wrong e-mail address) Recently I updated my GPS unit and decided to also update my maps (2011). After looking at Garmin's prices I decided to take a look at OSM, and things have come a long way since I looked last (2012?). I did some reading and headed down the road. Things seem to be working, but I'm getting inconsistent results - aka the gmapsupp.img file is not the same each time I create it. (Note: My Java isn't the latest-greatest, but it should be sufficient for this process: java version "1.8.0_202") Here's what I've done: 1) Downloaded the latest versions of mkgmap (r4486) and splitter (r597). I also downloaded and compiled osmconvert and osmfilter. I placed them in their own directory: (I'm a sysadmin by trade, this is what I do...) ls -l /opt/mkgmap/ /opt/mkgmap/lib/ /opt/mkgmap/bin/ /opt/mkgmap/: -rw-rw-r-- 1 bin bin 17989 May 6 02:37 LICENCE -rw-rw-r-- 1 bin bin 2784 May 6 02:37 README drwxr-xr-x 2 bin bin 4096 May 7 13:58 bin drwxrwxr-x 3 bin bin 4096 May 6 02:38 doc drwxrwxr-x 6 bin bin 4096 May 6 02:37 examples drwxrwxr-x 2 bin bin 4096 May 6 02:37 lib -rw-rw-r-- 1 bin bin 2350418 May 6 02:37 mkgmap.jar -rw-rw-r-- 1 bin bin 413879 Jan 30 07:11 splitter.jar /opt/mkgmap/bin/: -rwxr-xr-x 1 bin bin 182128 May 7 13:58 osmconvert -rwxr-xr-x 1 bin bin 100640 May 7 13:58 osmfilter /opt/mkgmap/lib/: -rw-rw-r-- 1 bin bin 475284 May 6 02:37 fastutil-6.5.15-mkg.1b.jar -rw-rw-r-- 1 bin bin 132654 May 6 02:37 osmpbf-1.3.3.jar -rw-rw-r-- 1 bin bin 533455 May 6 02:37 protobuf-java-2.5.0.jar -rw-rw-r-- 1 bin bin 120069 Jan 30 07:11 xpp3-1.1.4c.jar 2) I downloaded an OSM extract from Geofabrik, and using splitter carved up my canada-latest.osm.pbf into pieces. As per the wiki I used osmconvert and osmfilter to build the bounds data. These steps are consistent and repeatable. Here are the steps I've taken just in case they are consistently wrong... osmconvert canada-latest.osm.pbf -o=canada-latest.o5m osmfilter canada-latest.o5m \ --keep-nodes= \ --keep-ways-relations="boundary=administrative =postal_code postal_code=" \ -o=canada-latest-boundries.o5m mkdir bounds java -cp /opt/mkgmap/mkgmap.jar \ uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryPreprocessor \ canada-latest-boundries.o5m bounds mkdir work ( cd work ; java -jar /opt/mkgmap/splitter.jar ../canada-latest.osm.pbf ) The final step (afaik) is to combine the split files and the bounds information into a single gmapsupp.img file. ( cd work ; java -jar /opt/mkgmap/mkgmap.jar \ --route --add-pois-to-areas --bounds=../bounds \ --series-name='Canada OSM 2020' \ --family-name='Canada OSM 2020' \ --area-name='Canada OSM 2020' \ --country-name='Canada' --country-abbr='CA' \ --description='OpenStreetMap data for Garmin (2020)' \ --index --gmapsupp 6324*.osm.pbf ) This always runs to completion with MapFailedExceptions and ExitExceptions returning 0. Approximate time 19 minutes (4 core, 32G). I am assuming that the series-name, family-name, etc are just different tags used by different Garmin units. I haven't tried uploading it to my GPS unit yet as I'm concerned about the inconsistencies. Here are the results from multiple runs (osm/bounds/split data is all identical): 1260027904 May 7 16:54 gmapsupp.img 1260027904 May 7 17:18 gmapsupp2.img 1259929600 May 14 10:49 gmapsupp3.img 1259929600 May 14 11:44 gmapsupp4.img md5sum -b *.img 34053990aecc68a257e95115b2bb3f45 *gmapsupp.img 91e3d3e9f0012672da65dead8c752dd9 *gmapsupp2.img 821234ad26e17d0135bf613cf9faf161 *gmapsupp3.img abf39cef1430c4cbd74795fa11b571b8 *gmapsupp4.img Is this normal? If so why? Thank you for any help you can provide!

4 4

new feature request, New style command 'goto'
by Joris Bo 16 May '20

16 May '20

Hello, I finished a fancy but complex 'access evaluator' together with unittests prior to set the { mkgmap:car/truck/bicycle=yes/no/toll } It was not easy to keep it maintainable. The complexity makes it more difficult to keep in mind the impact of existing sections. The current mkgmap technique is processing line by line and first stops after a match is found resulting in an object to be visible in the map. Even if the match is a 1000 lines down the road (and even when I know its useless to try 999 of them) it will process them all, enabling the chance of a false positive (in between). I was wondering if it would be difficult to implement a new feature called : "goto' which immediately jumps to a certain section and continues from there. It is like the batch command 'goto, gosub'. Today for example I need to exclude rule after rule the possible existence of certain tags which are tested later on, to avoid an unexpected early match. It would be easier to just skip these parts and jump to the 'right section'. Because the software today also includes rules on the fly, and also uses 'if () then ... end' exclusions and also jumps to the finalizer if a match is found, it maybe just as simple as jump to another section? This will help a lot. It even implements the possibility of subroutines which I now do sequential bij calling the "include file" mechanism 9 times in a row for every transport type. Just a thought, it also works without this, but it would help a lot spending less debugging time. Especially after re-arranging sections which today for sure will end up with unexpected results. For example: include 'inc/myaccessrules'; myvariable:notransporttypeallowedatall = true {goto mynonroutablehighwaysection} <myhighwaysection> #don't bother the private and non-routable, it is already managed elsewhere building = * {goto mybuildingsection} natural = * {goto mynaturalsection} <myshopsection> shop = * & building !=* & natural !=* & amenity!=restaurant {name '${name}' | '${ housename}'} [0x0400 resolution 12-22 continue] <mybuildingsection> building = * {name '${name}' | '${addr:housename}'} [0x05 resolution 12-22 continue] <finalize> #reserved section name Kind reagards Jorisbo(a)hotmail.com<mailto:Jorisbo@hotmail.com>

1 0

Missing osm tags in the echo output
by Joris Bo 15 May '20

15 May '20

Hello, When debugging my map I encountered missing tags in the echotags output which are available at openstreetmap.org And also in the geofabric download because after assigning the tag to a variable in my style and after that call the echotags statement they do appear in the output. So I was thinking that mkgmap will only echo a part of the tags and not just 'all'? Can i do something about this behavior? Kind regards Joris Met vriendelijke groet, Joris Bovens +31 6 53233898 jorisbo(a)hotmail.com<mailto:jorisbo@hotmail.com>

4 4

nearby-POI: What to do with several POI arranged on a circle?
by Gerd Petermann 08 May '20

08 May '20

Hi all, while doing further tests I've found this edge case (see attached file) Draw a circle with a diameter of ~56m and tag all nodes amenity=bank. The default style has amenity=bank [0x2f06 resolution 24] With --nearby-poi-rules=0x2f06:30:merge-at-mid-point one might expect a single POI in the middle of the circle, but the result is different: Since there is no POI near the center of the circle the current algo finds 4 groups, all close to the edge of the circle. I think for merge-at-mid-point I need a different algo. I'm looking at https://en.wikipedia.org/wiki/Cluster_analysis now. Gerd

6 10

Open question reg. new nearby-poi options
by Gerd Petermann 07 May '20

07 May '20

Hi all, some days ago I expressed my concerns that the new option might have a bad influence on routing or the dead-end check. I was mistaken, there was and is no effect on routing or the results of the dead-end checks because the option only effects the POI, which is a different data structure. Sorry for the confusion. Gerd -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html

2 6

nearby-POI branch: please ignore r4489 and r4490
by Gerd Petermann 07 May '20

07 May '20

Hi all, sorry, see http://www.mkgmap.org.uk/websvn/revision.php?repname=mkgmap&rev=4491 Gerd

1 0