- mkgmap-dev - The mkgmap lists

[Patch v1] style filter country-ISO
by Gerd Petermann 28 Nov '14

28 Nov '14

Hi all, see also http://gis.19327.n5.nabble.com/useless-rules-in-default-style-inc-address-t… attached is a patch that implements a style filter to convert country names to the 3 character ISO code which is used in e.g. the address rules. It also changes two rules in inc/address to use this filter and adds a simple check to ValueBuilder to report the case when a filter is expected but not recognized because of parser or syntax errors like in the case reported here: http://gis.19327.n5.nabble.com/Substring-returns-full-string-tp5825603.html Gerd

3 2

useless rules in default style inc/address
by Gerd Petermann 28 Nov '14

28 Nov '14

Hi all, the default style file inc/address starts with these three rules: # first set the country code mkgmap:country!=* & mkgmap:admin_level2=* { set mkgmap:country='${mkgmap:admin_level2}' } mkgmap:country!=* & addr:country=* { set mkgmap:country='${addr:country}' } mkgmap:country!=* & is_in:country=* { set mkgmap:country='${is_in:country}' } I found no code in the mkgmap java sources that sets mkgmap:country, so the first rule will always set mkgmap:country unless one manipulates the OSM data. The following rules evaluate mkgmap:country and expect the ISO 1366 code with 3 characters, e.g. DEU for Germany. The wiki for tag addr:country says that it should contain the 2 character ISO code, e.g. DE for Germany. The is_in:country tag contains the country name, not the iso code. I think we should remove the two useless rules unless someone finds a working alternative. Gerd

4 6

[Patch v1] use --country-name / --country-abbr to set mkgmap:admin_level2
by Gerd Petermann 27 Nov '14

27 Nov '14

Hi all, attached is a small patch which allows to use the options country-name or country-abbr to set a default value for the tag mkgmap:country which is important for the address rules. This should help those users who live in a big country like the USA and try to create a map without downloading the precompiled boundaries and instead use the boundaries contained in the OSM file for e.g. one state. The patch works like this: If the LocationHook did not set mkgmap:admin_level2 : if --country-abbr is given, use the value to set mkgmap:admin_level2 else if --country-name is given, get the ISO name from LocatorConfig.xml and set mkgmap:admin_level2 with that. Of course this happens before style evaluation. Gerd

1 0

Commit: r3356: commit styleStats-v5.patch + small change to make unit tests work
by svn commit 27 Nov '14

27 Nov '14

Version mkgmap-r3356 was committed by gerd on Thu, 27 Nov 2014 commit styleStats-v5.patch + small change to make unit tests work

1 0

Non-rectangular tiles
by Steve Ratcliffe 27 Nov '14

27 Nov '14

Hi I was meaning to write a message about this for a while and since the thread for drive-on-left/right has brought this up already today and since it is a mkgmap birthday (8 years!) this might be a good time to set the ball rolling. Currently we use rectangular tiles, but it is perfectly possible to have arbitrarily shaped tiles - in particular tiles can be shaped to exactly follow a country border or be conveniently split by straight lines between countries separated by sea. This means that every rule could be customised for each country. This is probably the biggest feature that we don't have that we can achieve with what we know about the format. I am sure that it can be done, but it is a large task and will need more than one person working on it. If there is any interest I shall outline my thoughts tomorrow. ..Steve

5 4

evaluation of boolean options
by Gerd Petermann 27 Nov '14

27 Nov '14

Hi programmers, the evaluation of boolean options in mkgmap is a bit confusing. It is possible to specify e.g. --drive-on-left or --drive-on-left=true or --drive-on-left=false or --drive-on-left=any-value In the current code both all have the same effect, because the evaluation is done like this: driveOnLeft = props.getProperty("drive-on-left") != null; On the other hand, --keep-going and --keep-going=false do not have the same effect, and that is true for most boolean options, because typically we use this kind of evaluation: linkPOIsToWays = props.getProperty("link-pois-to-ways", false); I think we should always use the latter method. Do you agree? Gerd

3 3

[Patch v2] manage drive-on-left/drive-on-right in resources\LocatorConfig.xml
by Gerd Petermann 26 Nov '14

26 Nov '14

Hi all, only change to v1: if neither drive-on-left nor drive-on-right is set by user and the number of "dol" roundabouts is smaller than the number of "dor" ones: set drive-on-right so that check-roundabouts doesn't set drive-on-left flag in NOD header if first roundabout is a drive-on-left one. Note that the counters are only set when the location hook works, so precompiled bounds are needed (--bounds) Binary is here: http://files.mkgmap.org.uk/download/232/mkgmap.jar Gerd

1 0

[Patch v1] manage drive-on-left/drive-on-right in resources\LocatorConfig.xml
by Gerd Petermann 26 Nov '14

26 Nov '14

Hi all, attached is a first version of a patch to implement this. I've used this list http://www.worldstandards.eu/cars/list-of-left-driving-countries/ to fill the LocatorConfig.xml, I hope it is a good source. A compiled binary based on r3355 is here: http://files.mkgmap.org.uk/download/231/mkgmap.jar Gerd

1 0

Line label text size
by gbbickerton 25 Nov '14

25 Nov '14

Are there different default text sizes for different line types in the way that there are different text sizes for poi labels -- View this message in context: http://gis.19327.n5.nabble.com/Line-label-text-size-tp5825360.html Sent from the Mkgmap Development mailing list archive at Nabble.com.

2 2

Track rendering on device
by Gerd Petermann 25 Nov '14

25 Nov '14

Hi all, I wonder if this is a problem that is specific for maps generated with mkgmap : When I chose a track as target for routing the track is displayed as a magenta line. When I follow this track on my bicycle the device sometimes fails to render the map, as a result large areas are flooded with magenta . The device is an Oregon 450t. The problem disappears when I move a few meters or chose a different zoom level. I'm never able to reproduce the problem at home, so I guess it is related to the movement, maybe combined with low batteries. I did not yet try if the latest changes regarding MapSplitter improved this. Did anybody else see this effect in original Garmin maps? Any idea what the reason is? Gerd

2 3