October 2009 - mkgmap-dev - The mkgmap lists

MDR success and a couple of questions
by Steve Ratcliffe 08 Oct '09

08 Oct '09

Hi I've managed to construct a working MDR file with all of 2 POIs in it. This is with sections 1 (with empty subsections), 4, 9, 10, 11, 15. That appears to be the minimum number of sections to get anything to work. Interestingly, MDR 15 doesn't appear to be used much. It supplies auto-completion results when typing in the search boxes, but the actual search results show the actual labels of the points. So 4 and 9 are required but I don't completely understand them, so does anyone know or speculate what the second byte in MDR 4 is? The wiki says 'kind of index', suggesting that it tells you which sections use that POI type. If that is what is meant then what is the mapping? Secondly any ideas on MDR 9? The first record always appears to be 01 01 00 00, and that is enough to make searching work in my simple example, but beyond that I have no idea. ..Steve

8 20

Commit: r1261: Implement the reverse index.
by svn commit 08 Oct '09

08 Oct '09

Version 1261 was commited by steve on 2009-10-05 16:04:55 +0100 (Mon, 05 Oct 2009) BRANCH: mdr Implement the reverse index. Should allow the index to be transferred to a GPS unit using m/s. However completely untested at this stage. Looks OK for a small number of maps, however I am not convinced it is working when 3 byte pointers are needed for some sections.

8 13

Naming motorway exit road from the exit ref
by Mark Burton 08 Oct '09

08 Oct '09

I have committed the Java tweak required to support the naming of motorway exit roads from the name/ref of a motorway_junction POI. I haven't committed the required change in the style file because maybe not everyone wants this to happen. So, if you do, please add lines like this to your points file. highway=motorway_junction & ref=* { add motorway_link_name = '${ref}' name '${ref}' } highway=motorway_junction & name=* { add motorway_link_name = '${name}' name '${name}'} These set the motorway_link_name tag to the name/ref and that gets applied to the motorway_link way. Cheers, Mark

1 0

Commit: r1267: Added code to set a highway's name from a POI tag.
by svn commit 08 Oct '09

08 Oct '09

Version 1267 was commited by markb on 2009-10-08 12:41:45 +0100 (Thu, 08 Oct 2009) Added code to set a highway's name from a POI tag. If a way (of highway type X) has a POI for its first point and that POI has a tag called X_name, set the way's name to that tag's value.

1 0

Commit: r1266: Add a character set aware label decoder.
by svn commit 07 Oct '09

07 Oct '09

Version 1266 was commited by steve on 2009-10-07 14:46:33 +0100 (Wed, 07 Oct 2009) BRANCH: mdr Add a character set aware label decoder. This is a fairly big change as the CharacterDecoder interface is changed to have a method that returns a DecodedText rather than an EncodedText. The new class DecodedText is more useful as it contains the actual string that you need. non-ascii characters will now be correctly represented in the index, although they will still not work as the sorting needs adjusting.

1 0

Memory problems since 1245
by Du Plessis, Bennie 07 Oct '09

07 Oct '09

Hi, I don't know where to post a ticket, but I'm sure you'll tell me if it's the wrong forum. Since version 1245 I had to increase the memory size in the MkGMap command line and the virtual memory of my machine to get mkGMap to work. (1245 was the last version I could get to work, but with the -index switch off) What am I doing wrong or is my hardware just too small? Did anyone else experience a large increase in memory resource usage since MDR & indexing was introduced? I have not been able to test the results of the new dev around indexing yet, because I am not able to compile a map to send mapsource. It takes 3 hours to compile with these errors, so I can barely run 2 tries per night. If someone can save me time with some tips, I will sincerely appreciate it. java -ea -Xmx1300M -jar mkgmap-mdr-r1257.jar (if I increase Xmx java cannot create VM (whoever that is)) --name-tag-list=name:af,name --route --remove-short-arcs --road-name-pois (I need this to search streets on device, but I can remove it & build a separate map for the device if is the cause of the problem) --tdbfile --style-file=MyStyle --description=01000000 --product-id=0001 --overview-mapname=SdAf1T1246 (should have been 1257 this time, but will not have an effect) --mapname=10000000 --family-name=GeoSdAf1T1246 (should have been 1257 this time, but will not have an effect --family-id=0001 --Region-name="SARegion" --generate-sea --ignore-osm-bounds (If I don't use this the land & sea is reversed) --index --gmapsupp 10000000.osm.gz (29 MB -1644587,605843 to -885464,1677722# : -35.289,12.99998 to -19,36 I can reduce this, but reluctantly) SdAf1T1246_mdr.img gets made; .mdx not; 10000000.img gets made; gmapsupp not Acer travelmate 4202WLMi Intel Core Duo processor T2300 1.66 Ghz; 667Mhz FSB; 2MB L2 Cache 2 x 512DDR2 (Support dual channel) Windows XP Home Virtual memory Max size 5000 MB (increased to this point to try) Initial size 1524 MB Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: -68 at java.util.ArrayList.get(Unknown Source) at uk.me.parabola.imgfmt.sys.BlockTable.physFromLogical(BlockTable.java: 112) at uk.me.parabola.imgfmt.sys.Dirent.getPhysicalBlock(Dirent.java:262) at uk.me.parabola.imgfmt.sys.FileNode.read(FileNode.java:156) at uk.me.parabola.imgfmt.app.BufferedImgFileReader.fillBuffer(BufferedIm gFileReader.java:256) at uk.me.parabola.imgfmt.app.BufferedImgFileReader.get(BufferedImgFileRe ader.java:86) at uk.me.parabola.imgfmt.app.BufferedImgFileReader.getChar(BufferedImgFi leReader.java:103) at uk.me.parabola.imgfmt.app.trergn.RGNFileReader$RgnOffsets.<init>(RGNF ileReader.java:227) at uk.me.parabola.imgfmt.app.trergn.RGNFileReader$RgnOffsets.<init>(RGNF ileReader.java:196) at uk.me.parabola.imgfmt.app.trergn.RGNFileReader.getOffsets(RGNFileRead er.java:185) at uk.me.parabola.imgfmt.app.trergn.RGNFileReader.pointsForSubdiv(RGNFil eReader.java:70) at uk.me.parabola.imgfmt.app.map.MapReader.pointsForLevel(MapReader.java :93) at uk.me.parabola.mkgmap.combiners.MdrBuilder.addPoints(MdrBuilder.java: 156) at uk.me.parabola.mkgmap.combiners.MdrBuilder.onMapEnd(MdrBuilder.java:1 09) at uk.me.parabola.mkgmap.main.Main.endOptions(Main.java:358) at uk.me.parabola.mkgmap.CommandArgsReader.readArgs(CommandArgsReader.ja va:124) at uk.me.parabola.mkgmap.main.Main.main(Main.java:121) If you could not tell by now: I have no IT savvy so laymans terms & patience required ;-\ BennieD ##################################################################################### Scanned by MailMarshal - Marshal's comprehensive email content security solution. Download a free evaluation of MailMarshal at www.marshal.com #####################################################################################

8 19

[PATCH v3] - tweeze arc headings to improve routing instructions
by Mark Burton 07 Oct '09

07 Oct '09

v3 - now based on r1265. Added --adjust-turn-headings option to enable this functionality. Reports now downgraded to info (from warn). ------------------------ v2 - various changes - should now work with junctions that have more than 1 side road - should be able to treat exits from roundabouts in the same way as side roads. ------------------------- The attached patch (which also includes the drive-on-left patch because it's based on that) attempts to improve the quality of routing instructions so that when a turn off a road that doesn't involve a big heading change is required, the GPS should say "turn left/right" rather than "keep left/right". It can also reduce the number of "keep left/right" instructions you get when you are not leaving the main road but passing a junction that has a shallow turn (e.g. sometimes happens on motorways when it tells you to keep left/right even though you are not leaving the motorway). I have tested it a little and it appears to work. I expect there are situations where it does the wrong thing so it would be great if people could give it a go and see if it behaves OK. There isn't any option to enable the feature at the moment. If you use the patch, it will be enabled. It's currently issuing a warning for each change it makes but later that can be downgraded to info. Cheers, Mark

1 0

Commit: r1264: Move code that determines bearing of one point from another into Coord class.
by svn commit 06 Oct '09

06 Oct '09

Version 1264 was commited by markb on 2009-10-06 21:29:57 +0100 (Tue, 06 Oct 2009) Move code that determines bearing of one point from another into Coord class.

1 0

Commit: r1263: Several fixes to the reverse index MDR 1.
by svn commit 06 Oct '09

06 Oct '09

Version 1263 was commited by steve on 2009-10-06 15:02:18 +0100 (Tue, 06 Oct 2009) BRANCH: mdr Several fixes to the reverse index MDR 1. Major error in the mdr10 record number fixed. Deal with sections that are less than 256 records in the reverse index. Now I am able to select a subset of the available map tiles and send them to the GPS unit.

1 0

Handling of Garmin magic label codes (highwayshields etc.) differs from mp-format input to osm-format ?
by Gert Münzel 06 Oct '09

06 Oct '09

Hi, i have already asked( "Problem with special Garmin label codes"). I try to use the Garmin magic label codes in the form ~[0x06]B51~[0x1f]Hauptstrasse If i pass this inside a mp-format file mkgmap could handle it. If i triy it with input in osm-format i can't get it to work. I have made an entry in stylefile like highway=* {name '${ref|highway-symbol:oval} ${name|highway-symbol:ele}' | '${ref|highway-symbol:oval}' | '${name}' } where i expect as result: ~[0x06]B51~[0x1f]Hauptstrasse if ref=B51 and name=Hauptstrasse But always the second magic codes is ignored by mkgamp. Is the format of my stylefile entry correct or false.? Or simply will mp-format and osm-format be treaten different by mkgamp in this case? Or what else i have to do to get it working also with osm-input. TIA Gert

4 8