Re: [mkgmap-dev] Index and equally named cities
Just to add another observation. Under the address search, you are prompted to spell a city. In my previous mkgmap compile, you can only find a city if it is within the tile of your current location. With city-region-index I can now select all cities in my map and search streetnames. On Thu, Jul 14, 2011 at 3:25 AM, Peter Lerner wrote: > Steve, > >> Done. Its at: http://files.mkgmap.org.uk/download/33/mkgmap.jar > > I think "locator + city-region-index branches (2)" is a milestone! > This is a strong "+1" from my side to commit the changes as soon as > possible and try to base a stable mkgmap release v1.0(?) on this version. > > For a long time ever I was able to load a 4 GB map witout problems to a > GPS. The map included cyrillic language tiles and tiles that were > mutually "incompatible" before due to index problems. > Also no more MDR_TRIM problems ... ;-) > > > My current findings on the search address funtion so far. I'm using an > "old" GPSmap 62 CSx. > > -- the region field is not available, only fields for > housenumber, street and city > > -- only a combination of housenumber and street can be found; > housenumber doesn't matter > (i'm not using locator-branch special rules) > > -- the found location is slightly offset from the > real course of the street displayed on the GPS > > -- searching for streetnames, GPSmap 62XSx lists only streets > up to ca. 40-50 km around the current location. This is > different from Mapsource, which lists all streets it finds. > > -- some streets appear multiple times in search list (2-3x); > this has different reasons, e.g.: > > - entry 1+2: each entry leads to the same three streets > equally named in three different cities some kms apart > > The behaviour that one entry leads to three > different streets is OK for the current state, > and subject to optimisation in the locator branch. > > The fact that the same entry is doubled in the search > index needs to be investigated further. > > - entry 3: this is a separate entry and possibly has to do > with a street tagged as cycleway=track and the > --make-all-cycleways option. I think this option creates > some problems when creating indices. > > I need to investigate the unwanted side effects of this > option on the index further. > > Peter > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Martin, read this: http://wiki.openstreetmap.org/wiki/Mkgmap/help/usage#Addressindex_with_locator-branch WanMil > WanMil, > > there were some changes on the boundaries around Stuttgart (suburbs were > tagged with admin-level=8), so that you get Stuttgart-Nord instead of > Stuttgart. > You can also give me the osmosis-command -line(s), so I can try it myself. I > think it's also good to update the wiki ;) > > Cheers, > Martin > > Am 12.07.2011 um 22:50 schrieb WanMil: > >>> >>> And now a special request to WanMil: >>> Could you please update the boundary-files? >>> >> >> Martin, >> >> at the moment Luxembourg is broken so it makes no sense for me to update >> the bounds files. >> Johan want to correct that within the next days. >> >> WanMil >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
WanMil, there were some changes on the boundaries around Stuttgart (suburbs were tagged with admin-level=8), so that you get Stuttgart-Nord instead of Stuttgart. You can also give me the osmosis-command -line(s), so I can try it myself. I think it's also good to update the wiki ;) Cheers, Martin Am 12.07.2011 um 22:50 schrieb WanMil: >> >> And now a special request to WanMil: >> Could you please update the boundary-files? >> > > Martin, > > at the moment Luxembourg is broken so it makes no sense for me to update > the bounds files. > Johan want to correct that within the next days. > > WanMil > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Steve, > Done. Its at: http://files.mkgmap.org.uk/download/33/mkgmap.jar I think "locator + city-region-index branches (2)" is a milestone! This is a strong "+1" from my side to commit the changes as soon as possible and try to base a stable mkgmap release v1.0(?) on this version. For a long time ever I was able to load a 4 GB map witout problems to a GPS. The map included cyrillic language tiles and tiles that were mutually "incompatible" before due to index problems. Also no more MDR_TRIM problems ... ;-) My current findings on the search address funtion so far. I'm using an "old" GPSmap 62 CSx. -- the region field is not available, only fields for housenumber, street and city -- only a combination of housenumber and street can be found; housenumber doesn't matter (i'm not using locator-branch special rules) -- the found location is slightly offset from the real course of the street displayed on the GPS -- searching for streetnames, GPSmap 62XSx lists only streets up to ca. 40-50 km around the current location. This is different from Mapsource, which lists all streets it finds. -- some streets appear multiple times in search list (2-3x); this has different reasons, e.g.: - entry 1+2: each entry leads to the same three streets equally named in three different cities some kms apart The behaviour that one entry leads to three different streets is OK for the current state, and subject to optimisation in the locator branch. The fact that the same entry is doubled in the search index needs to be investigated further. - entry 3: this is a separate entry and possibly has to do with a street tagged as cycleway=track and the --make-all-cycleways option. I think this option creates some problems when creating indices. I need to investigate the unwanted side effects of this option on the index further. Peter ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> > And now a special request to WanMil: > Could you please update the boundary-files? > Martin, at the moment Luxembourg is broken so it makes no sense for me to update the bounds files. Johan want to correct that within the next days. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Thank you Steve&WanMil,
after testing the new city-region-locator-branch I have to say: You've done a
great job. Now I can find more cities, for e.g. Neustadt in Brandenburg, in
Schleswig-Holstein, in Sachsen etc... Fantastic.
So, few more things I've seen. Zip-search still not work. If I use the
locator-branch with the following lines:
mkgmap:postal_code!=* & addr:postcode=* { set
mkgmap:postal_code='${addr:postcode}' }
mkgmap:postal_code!=* & openGeoDB:postal_codes=* { set
mkgmap:postal_code='${openGeoDB:postal_codes}' }
mkgmap:postal_code!=* & mkgmap:postcode=* { set
mkgmap:postal_code='${mkgmap:postcode}' }
I find long streets (that have more than on zip-code) a few times. If I leave
this lines out, I find this streets only once.
So it seems, that he create the table for the zip-code, but not correct.
Because in the adress-search the search for zip-code doesn't appear. (Any
idea?!) When I try to find the zip-codes in Mapsource he complete the line with
possible zip-codes, but also fail to find streets by zip-code.
And now a special request to WanMil:
Could you please update the boundary-files?
Cheers,
Martin
Am 12.07.2011 um 17:47 schrieb Steve Ratcliffe:
> Martin
>
>> I've tried to merge the locator-branch with your city-region-branch, but I
>> get a few errors and he failed to build the jar. Could you please merge it
>> for me?
>> I've tried:
>
> Done. Its at: http://files.mkgmap.org.uk/download/33/mkgmap.jar
>
> ..Steve
> ___
> mkgmap-dev mailing list
> mkgmap-dev@lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
>
___
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Martin > I've tried to merge the locator-branch with your city-region-branch, but I > get a few errors and he failed to build the jar. Could you please merge it > for me? > I've tried: Done. Its at: http://files.mkgmap.org.uk/download/33/mkgmap.jar ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
works on nuvi 1310 city search and address search On Tue, Jul 12, 2011 at 9:49 AM, maning sambale wrote: > Initial results: > > Compiled with mkgmap-city-region-index-r1992 tested on nuvi 255w. > All my San Fernandos are searcheable both in the address search and > city search. Streets are searchable as well. I also noticed a great > improvement in the street results which may have sole another bug > reported here: https://github.com/maning/osmphgps/issues/34 > > If anyone wants to test this in 4digit nuvi models, please report hem > here. Thanks! > My maps are here: > Links below: > > http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_img_latest_dev.zip > http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_macroadtrip_latest_dev.zip > http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_winmapsource_latest_dev.exe > > > > On Mon, Jul 11, 2011 at 10:42 PM, Steve Ratcliffe > wrote: >> On 11/07/11 15:26, maning sambale wrote: >>> Thanks will wait and test. >> >> You need wait no longer! >> >> The current city-region-index branch version has the changes. Please let >> me know what you think. >> >> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > > > -- > cheers, > maning > -- > "Freedom is still the most radical idea of all" -N.Branden > wiki: http://esambale.wikispaces.com/ > blog: http://epsg4253.wordpress.com/ > -- > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Initial results: Compiled with mkgmap-city-region-index-r1992 tested on nuvi 255w. All my San Fernandos are searcheable both in the address search and city search. Streets are searchable as well. I also noticed a great improvement in the street results which may have sole another bug reported here: https://github.com/maning/osmphgps/issues/34 If anyone wants to test this in 4digit nuvi models, please report hem here. Thanks! My maps are here: Links below: http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_img_latest_dev.zip http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_macroadtrip_latest_dev.zip http://dl.dropbox.com/u/607635/osm-ph_gps_maps/dev/osmph_winmapsource_latest_dev.exe On Mon, Jul 11, 2011 at 10:42 PM, Steve Ratcliffe wrote: > On 11/07/11 15:26, maning sambale wrote: >> Thanks will wait and test. > > You need wait no longer! > > The current city-region-index branch version has the changes. Please let > me know what you think. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Steve,
I've tried to merge the locator-branch with your city-region-branch, but I get
a few errors and he failed to build the jar. Could you please merge it for me?
I've tried:
svn co http://svn.parabola.me.uk/mkgmap/branches/locator mkgmap-locator
cd mkgmap-locator
svn merge ^/branches/city-region-index
ant dist
and then...
martins-macbook-pro-15:mkgmap-locator Martin$ ant dist
Buildfile: /Users/Martin/Desktop/map/mkgmap-locator/build.xml
prepare:
compile:
[javac] /Users/Martin/Desktop/map/mkgmap-locator/build.xml:93: warning:
'includeantruntime' was not set, defaulting to build.sysclasspath=last; set to
false for repeatable builds
[javac] Compiling 399 source files to
/Users/Martin/Desktop/map/mkgmap-locator/build/classes
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:27:
package crosby.binary does not exist
[javac] import crosby.binary.BinaryParser;
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:28:
package crosby.binary does not exist
[javac] import crosby.binary.Osmformat;
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinMapDataSource.java:24:
package crosby.binary.file does not exist
[javac] import crosby.binary.file.BlockInputStream;
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:42:
cannot find symbol
[javac] symbol : class BinaryParser
[javac] location: class uk.me.parabola.mkgmap.reader.osm.bin.OsmBinHandler
[javac] public class BinParser extends BinaryParser {
[javac]^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:44:
package Osmformat does not exist
[javac] protected void parse(Osmformat.HeaderBlock header) {
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:64:
package Osmformat does not exist
[javac] protected void parseNodes(List nodes) {
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:87:
package Osmformat does not exist
[javac] protected final void parseDense(Osmformat.DenseNodes
nodes) {
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:130:
package Osmformat does not exist
[javac] protected void parseWays(List ways) {
[javac]^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:164:
package Osmformat does not exist
[javac] protected void parseRelations(List
rels) {
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinMapDataSource.java:60:
cannot find symbol
[javac] symbol : class BlockInputStream
[javac] location: class
uk.me.parabola.mkgmap.reader.osm.bin.OsmBinMapDataSource
[javac] BlockInputStream stream = new
BlockInputStream(is, reader);
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinMapDataSource.java:60:
cannot find symbol
[javac] symbol : class BlockInputStream
[javac] location: class
uk.me.parabola.mkgmap.reader.osm.bin.OsmBinMapDataSource
[javac] BlockInputStream stream = new
BlockInputStream(is, reader);
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:65:
package Osmformat does not exist
[javac] for (Osmformat.Node binNode : nodes) {
[javac] ^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:93:
operator + cannot be applied to Osmformat.DenseNodes.getLat,long
[javac] long lat = nodes.getLat(nid) + lastLat;
[javac]^
[javac]
/Users/Martin/Desktop/map/mkgmap-locator/src/uk/me/parabola/mkgmap/reader/osm/bin/OsmBinHandler.java:93:
incompatible types
[javac] found :
[javac] required: long
[javac] long lat = nodes.getLat(nid) + lastLat;
Re: [mkgmap-dev] Index and equally named cities
On 11/07/11 15:26, maning sambale wrote: > Thanks will wait and test. You need wait no longer! The current city-region-index branch version has the changes. Please let me know what you think. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Thanks will wait and test. On Mon, Jul 11, 2011 at 4:51 PM, Steve Ratcliffe wrote: > Thanks, that was very helpful. I can probably work on a fix for that now. > > It looks like the city search, only uses the POI section of the index. > Therefore I need to try making the same change to the POI section and > retain multiple city POIs that have different region/country's. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> Hi > >> Oh, I wasn't aware of special svn merge information and merged changes >> from the trunk manually. Good to know. Will try that next time. > > You can fix up the merge information for the revisions that are already > merged from trunk with: > > cd > svn merge --record-only -r1891:1973 ^/trunk . > svn commit > > Then merging from trunk can be just: svn merge ^/trunk . > > ..Steve Thanks for the hints! Can you do me a favour and commit that for me? I am using the Eclipse SVN facility only and would have to try a lot how to achieve that with the special "record-only" parameter. Thanks! WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > Oh, I wasn't aware of special svn merge information and merged changes > from the trunk manually. Good to know. Will try that next time. You can fix up the merge information for the revisions that are already merged from trunk with: cd svn merge --record-only -r1891:1973 ^/trunk . svn commit Then merging from trunk can be just: svn merge ^/trunk . ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Thanks, that was very helpful. I can probably work on a fix for that now. It looks like the city search, only uses the POI section of the index. Therefore I need to try making the same change to the POI section and retain multiple city POIs that have different region/country's. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Yes, although minimal. --location-autofill=0 On Mon, Jul 11, 2011 at 12:53 PM, Charlie Ferrero wrote: > --location-autofill -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 11/07/2011 04:51, maning sambale wrote: > Dear steve, > > Here's a bit of explanation from one of our colleagues. > >> So can you explain exactly what is better and worse between the >> city-region-index branch r1867 and r1870. > >> Do the different "San Fernando's" have different regions? If not how do >> you tell them apart? > Out of curiosity, are you using --location-autofill? I've noticed that mkgmap isn't handling region, city and country names properly when location-autofill is used, but haven't quite got to the bottom of it yet. ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hallo Martin, hast du die neue Karte schon zum download eingestellt? Ich würde mal schauen ob meine "fehlenden" Straßen auftauchen. Ich konnte mit der letzten Karte ein paar Straßen finden, aber nicht alle. Der Straßenname hat vermutlich auch kein "Doppel" das irritieren könnte... Laas mal hören wo die Karte zu laden ist - Vieri Augen mappen mehr als zwei :-) Gruß Ludwich Am 10.07.2011 um 20:02 schrieb Martin: > Steve, > > thank you for merging both branches. I could make a map, but still get one > city that exist in 2 different regions. When I just search for city in the > city-search-function of my garmin I see both cities (Chemnitz, Sachsen and > Chemnitz, Mecklenburg-Vorpommern). But when I use the address-search function > I only find Chemnitz, Mecklenburg-Vorpommern. In the next menu I see all > streets, which are in Chemnitz, Sachsen, but when I choose these streets, No > results are found... Strange thing... > > Cheers, > Martin > > Am 09.07.2011 um 23:08 schrieb Steve Ratcliffe: > >> Hi >> >>> for public. I have no idea how I can download the city-region-& >>> location-branch and the r1973. I also didn't know how to patch this both >>> files and apply this patch to the locator branch. I think you know how >>> to do this. So if you find some time, it would be nice if you can >>> provide me this "special branch". >> >> Normally to merge a branches you would check out the locator branch and >> then merge the other one in like this: >> >> svn merge ^/branches/city-region-index . >> >> it doesn't work in this case, because the merge information is missing >> on the locator branch, probably because the merge command wasn't used to >> merge in the trunk changes. >> >> Anyway I fixed the merge information and created a merged tree locally >> and uploaded the result to: >> http://files.mkgmap.org.uk/download/32/mkgmap.jar >> >> I haven't tested it... >> >> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Steve, thank you for merging both branches. I could make a map, but still get one city that exist in 2 different regions. When I just search for city in the city-search-function of my garmin I see both cities (Chemnitz, Sachsen and Chemnitz, Mecklenburg-Vorpommern). But when I use the address-search function I only find Chemnitz, Mecklenburg-Vorpommern. In the next menu I see all streets, which are in Chemnitz, Sachsen, but when I choose these streets, No results are found... Strange thing... Cheers, Martin Am 09.07.2011 um 23:08 schrieb Steve Ratcliffe: > Hi > >> for public. I have no idea how I can download the city-region-& >> location-branch and the r1973. I also didn't know how to patch this both >> files and apply this patch to the locator branch. I think you know how >> to do this. So if you find some time, it would be nice if you can >> provide me this "special branch". > > Normally to merge a branches you would check out the locator branch and > then merge the other one in like this: > > svn merge ^/branches/city-region-index . > > it doesn't work in this case, because the merge information is missing > on the locator branch, probably because the merge command wasn't used to > merge in the trunk changes. > > Anyway I fixed the merge information and created a merged tree locally > and uploaded the result to: > http://files.mkgmap.org.uk/download/32/mkgmap.jar > > I haven't tested it... > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> > it doesn't work in this case, because the merge information is missing > on the locator branch, probably because the merge command wasn't used to > merge in the trunk changes. Oh, I wasn't aware of special svn merge information and merged changes from the trunk manually. Good to know. Will try that next time. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 07/07/11 09:35, maning sambale wrote: > Testing this version, > Uploading via mac mapinstall is a bit longer but not too long > > I only get two "San Fernando"s that are in a separate tile. The other > "San Fernando"s within one tile is not included in the search results. So can you explain exactly what is better and worse between the city-region-index branch r1867 and r1870. Do the different "San Fernando's" have different regions? If not how do you tell them apart? ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > for public. I have no idea how I can download the city-region-& > location-branch and the r1973. I also didn't know how to patch this both > files and apply this patch to the locator branch. I think you know how > to do this. So if you find some time, it would be nice if you can > provide me this "special branch". Normally to merge a branches you would check out the locator branch and then merge the other one in like this: svn merge ^/branches/city-region-index . it doesn't work in this case, because the merge information is missing on the locator branch, probably because the merge command wasn't used to merge in the trunk changes. Anyway I fixed the merge information and created a merged tree locally and uploaded the result to: http://files.mkgmap.org.uk/download/32/mkgmap.jar I haven't tested it... ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On Fri, Jul 08, 2011 at 02:49:17PM +0200, Martin wrote: >Hello Marko, > >thanks for your help: >I would like to merge locator with the city-region-branch. >Using your steps, I see, that Steve made the branch from 1979, or?! > >r1980 | steve | 2011-07-01 15:30:57 +0200 (Fre, 01. Jul 2011) | 2 Zeilen > >Create branch for city/region problem fix Subversion revision numbers are repository-wide. The common practice is that the repository contains subdirectories, such as "trunk" and "branches/locator". >So, now, that I know how I can download different versions (tried 1973 >and 1979), I tried to make a patch with >diff -r -u mkgmap-city newtrunk > patch1 "svn merge" should do that in a more intelligent way, resolving conflicts and taking into account renamed files. I took a look at the output of "svn diff", and I do not see any reference to BlockInputStream there. So, I guess that "ant dist" should fail even in the unmodified locator branch. Sure enough, I can repeat the build error after "svn revert -R ." Now I remember that I had a little tweak in my mkgmap/external.properties, replacing jars=/opt/jars with a different path. When I tweak that, the locator branch builds for me. The jars directory must contain the following: /opt/jars/osmprotobuf/osmprotobuf.jar /opt/jars/protobuf-2.3.0/protobuf.jar On my system, these point to the following files: osmosis-0.37/lib/default/osmbin-1.0-6d760534.jar osmosis-0.37/lib/default/protobuf-java-2.3.0.jar >So any idea what's wrong?! Just follow the steps that I posted earlier, but make sure that the jar files are available. Best regards, Marko ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hello Marko, thanks for your help: I would like to merge locator with the city-region-branch. Using your steps, I see, that Steve made the branch from 1979, or?! r1980 | steve | 2011-07-01 15:30:57 +0200 (Fre, 01. Jul 2011) | 2 Zeilen Create branch for city/region problem fix So, now, that I know how I can download different versions (tried 1973 and 1979), I tried to make a patch with diff -r -u mkgmap-city newtrunk > patch1 When I try to patch the locator with this patch I get a lot of errors: Hunk #1 FAILED at 1. 1 out of 1 hunk FAILED -- saving rejects to file all-wcprops.rej ... So any idea what's wrong?! Cheers, Martin Am 2011-07-08 14:02, schrieb Marko Mäkelä: > On Fri, Jul 08, 2011 at 01:42:13PM +0200, Martin wrote: >> I also didn't know how to patch this both files and apply this patch to >> the locator branch. > Here is how it should work: > > svn co http://svn.parabola.me.uk/mkgmap/branches/locator mkgmap-locator > cd mkgmap-locator > svn log|less # see which trunk revision it is based on > # The log reports for me that the svn:log of the current head (r1977) is > # "Synchronize to trunk r1973". So, you will need to merge changes from > # that on: > svn merge http://svn.parabola.me.uk/mkgmap/trunk -r1973:HEAD . > # This went without conflicts. > ant dist > # This gives compilation failures: "cannot find class BlockInputStream" > > So, it does not seem to be that simple. > > Best regards, > > Marko > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On Fri, Jul 08, 2011 at 01:42:13PM +0200, Martin wrote: >I also didn't know how to patch this both files and apply this patch to >the locator branch. Here is how it should work: svn co http://svn.parabola.me.uk/mkgmap/branches/locator mkgmap-locator cd mkgmap-locator svn log|less # see which trunk revision it is based on # The log reports for me that the svn:log of the current head (r1977) is # "Synchronize to trunk r1973". So, you will need to merge changes from # that on: svn merge http://svn.parabola.me.uk/mkgmap/trunk -r1973:HEAD . # This went without conflicts. ant dist # This gives compilation failures: "cannot find class BlockInputStream" So, it does not seem to be that simple. Best regards, Marko ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
WanMil, you are right. My fault, I didn't want you to merge this two branches for public. I have no idea how I can download the city-region- & location-branch and the r1973. I also didn't know how to patch this both files and apply this patch to the locator branch. I think you know how to do this. So if you find some time, it would be nice if you can provide me this "special branch". I think you can upload it to: http://files.mkgmap.org.uk/download Thank you in advance. Martin Am 2011-07-08 13:16, schrieb WanMil: > Martin, > > I don't want to merge changes from one branch to another branch. The > advantage of the branch is that one can concentrate on the development > of a few specific things only (better assignment of city/region/country > information in the locator branch). Merging unfinished developments from > another branch might introduce new problems. > > Of course once the changes from the city-region branch are merged back > to the trunk I will apply the changes to the locator branch soon. > > At the moment you can do that yourself: > * Create a patch between the city-region branch and r1973 > * Apply this patch to the locator branch > > WanMil > >> @WanMil: could you please merge the locator-branch with the last >> city-region-branch from Steve?! I have no idea, how can I make this. >> >> THX >> Martin >> >> Am 07.07.2011 um 10:35 schrieb maning sambale: >> >>> Testing this version, >>> Uploading via mac mapinstall is a bit longer but not too long >>> >>> I only get two "San Fernando"s that are in a separate tile. The other >>> "San Fernando"s within one tile is not included in the search results. >>> >>> >>> >>> args.list: >>> >>> code-page=1252 >>> tdbfile >>> latin1 >>> country-abbr=PHL >>> country-name=PHILIPPINES >>> remove-short-arcs=5 >>> route >>> add-pois-to-areas >>> family-id=639 >>> family-name=OSM_PHIL >>> overview-mapname=4001 >>> series-name=OSM_PHIL >>> description=OSM Philippines >>> style-file=/home/maning/osm/routable_garmin/git/osmphgps/styles/default >>> generate-sea=polygons,extend-sea-sectors,close-gaps=1000 >>> index >>> adjust-turn-headings >>> check-roundabouts >>> drive-on-right >>> check-roundabout-flares >>> report-dead-ends >>> ignore-maxspeeds >>> link-pois-to-ways >>> location-autofill=0 >>> >>> styles: >>> https://github.com/maning/osmphgps/tree/master/styles/default >>> >>> On Thu, Jul 7, 2011 at 4:01 PM, maning sambale >>>wrote: Got it! On Thu, Jul 7, 2011 at 3:58 PM, maning sambale wrote: > Where to download your branch? > > This is also a long standing issue for our PH maps. See issue report > here: > https://github.com/maning/osmphgps/issues/9 > > Basically, we have 6 "San Fernando" cities and city/address search > only works for this case when using mkgmap r1867. > > On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe > wrote: >> Hi >> >> I am away this week so didn't have time to mention it, but I created a >> branch (city-region-index) to fix the city/region bug since it turned >> out to be quite a complex change. >> >> It is now working as far as I can tell. Upload to a device works without >> mapsource crashes and giving both city and region in the search box >> works. >> >> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > > -- > cheers, > maning > -- > "Freedom is still the most radical idea of all" -N.Branden > wiki: http://esambale.wikispaces.com/ > blog: http://epsg4253.wordpress.com/ > -- > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- >>> >>> >>> -- >>> cheers, >>> maning >>> -- >>> "Freedom is still the most radical idea of all" -N.Branden >>> wiki: http://esambale.wikispaces.com/ >>> blog: http://epsg4253.wordpress.com/ >>> -- >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > _
Re: [mkgmap-dev] Index and equally named cities
Martin, I don't want to merge changes from one branch to another branch. The advantage of the branch is that one can concentrate on the development of a few specific things only (better assignment of city/region/country information in the locator branch). Merging unfinished developments from another branch might introduce new problems. Of course once the changes from the city-region branch are merged back to the trunk I will apply the changes to the locator branch soon. At the moment you can do that yourself: * Create a patch between the city-region branch and r1973 * Apply this patch to the locator branch WanMil > @WanMil: could you please merge the locator-branch with the last > city-region-branch from Steve?! I have no idea, how can I make this. > > THX > Martin > > Am 07.07.2011 um 10:35 schrieb maning sambale: > >> Testing this version, >> Uploading via mac mapinstall is a bit longer but not too long >> >> I only get two "San Fernando"s that are in a separate tile. The other >> "San Fernando"s within one tile is not included in the search results. >> >> >> >> args.list: >> >> code-page=1252 >> tdbfile >> latin1 >> country-abbr=PHL >> country-name=PHILIPPINES >> remove-short-arcs=5 >> route >> add-pois-to-areas >> family-id=639 >> family-name=OSM_PHIL >> overview-mapname=4001 >> series-name=OSM_PHIL >> description=OSM Philippines >> style-file=/home/maning/osm/routable_garmin/git/osmphgps/styles/default >> generate-sea=polygons,extend-sea-sectors,close-gaps=1000 >> index >> adjust-turn-headings >> check-roundabouts >> drive-on-right >> check-roundabout-flares >> report-dead-ends >> ignore-maxspeeds >> link-pois-to-ways >> location-autofill=0 >> >> styles: >> https://github.com/maning/osmphgps/tree/master/styles/default >> >> On Thu, Jul 7, 2011 at 4:01 PM, maning sambale >> wrote: >>> Got it! >>> >>> On Thu, Jul 7, 2011 at 3:58 PM, maning sambale >>> wrote: Where to download your branch? This is also a long standing issue for our PH maps. See issue report here: https://github.com/maning/osmphgps/issues/9 Basically, we have 6 "San Fernando" cities and city/address search only works for this case when using mkgmap r1867. On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe wrote: > Hi > > I am away this week so didn't have time to mention it, but I created a > branch (city-region-index) to fix the city/region bug since it turned > out to be quite a complex change. > > It is now working as far as I can tell. Upload to a device works without > mapsource crashes and giving both city and region in the search box works. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- >>> >>> >>> >>> -- >>> cheers, >>> maning >>> -- >>> "Freedom is still the most radical idea of all" -N.Branden >>> wiki: http://esambale.wikispaces.com/ >>> blog: http://epsg4253.wordpress.com/ >>> -- >>> >> >> >> >> -- >> cheers, >> maning >> -- >> "Freedom is still the most radical idea of all" -N.Branden >> wiki: http://esambale.wikispaces.com/ >> blog: http://epsg4253.wordpress.com/ >> -- >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
@WanMil: could you please merge the locator-branch with the last city-region-branch from Steve?! I have no idea, how can I make this. THX Martin Am 07.07.2011 um 10:35 schrieb maning sambale: > Testing this version, > Uploading via mac mapinstall is a bit longer but not too long > > I only get two "San Fernando"s that are in a separate tile. The other > "San Fernando"s within one tile is not included in the search results. > > > > args.list: > > code-page=1252 > tdbfile > latin1 > country-abbr=PHL > country-name=PHILIPPINES > remove-short-arcs=5 > route > add-pois-to-areas > family-id=639 > family-name=OSM_PHIL > overview-mapname=4001 > series-name=OSM_PHIL > description=OSM Philippines > style-file=/home/maning/osm/routable_garmin/git/osmphgps/styles/default > generate-sea=polygons,extend-sea-sectors,close-gaps=1000 > index > adjust-turn-headings > check-roundabouts > drive-on-right > check-roundabout-flares > report-dead-ends > ignore-maxspeeds > link-pois-to-ways > location-autofill=0 > > styles: > https://github.com/maning/osmphgps/tree/master/styles/default > > On Thu, Jul 7, 2011 at 4:01 PM, maning sambale > wrote: >> Got it! >> >> On Thu, Jul 7, 2011 at 3:58 PM, maning sambale >> wrote: >>> Where to download your branch? >>> >>> This is also a long standing issue for our PH maps. See issue report here: >>> https://github.com/maning/osmphgps/issues/9 >>> >>> Basically, we have 6 "San Fernando" cities and city/address search >>> only works for this case when using mkgmap r1867. >>> >>> On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe >>> wrote: Hi I am away this week so didn't have time to mention it, but I created a branch (city-region-index) to fix the city/region bug since it turned out to be quite a complex change. It is now working as far as I can tell. Upload to a device works without mapsource crashes and giving both city and region in the search box works. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >>> >>> >>> >>> -- >>> cheers, >>> maning >>> -- >>> "Freedom is still the most radical idea of all" -N.Branden >>> wiki: http://esambale.wikispaces.com/ >>> blog: http://epsg4253.wordpress.com/ >>> -- >>> >> >> >> >> -- >> cheers, >> maning >> -- >> "Freedom is still the most radical idea of all" -N.Branden >> wiki: http://esambale.wikispaces.com/ >> blog: http://epsg4253.wordpress.com/ >> -- >> > > > > -- > cheers, > maning > -- > "Freedom is still the most radical idea of all" -N.Branden > wiki: http://esambale.wikispaces.com/ > blog: http://epsg4253.wordpress.com/ > -- > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Testing this version, Uploading via mac mapinstall is a bit longer but not too long I only get two "San Fernando"s that are in a separate tile. The other "San Fernando"s within one tile is not included in the search results. args.list: code-page=1252 tdbfile latin1 country-abbr=PHL country-name=PHILIPPINES remove-short-arcs=5 route add-pois-to-areas family-id=639 family-name=OSM_PHIL overview-mapname=4001 series-name=OSM_PHIL description=OSM Philippines style-file=/home/maning/osm/routable_garmin/git/osmphgps/styles/default generate-sea=polygons,extend-sea-sectors,close-gaps=1000 index adjust-turn-headings check-roundabouts drive-on-right check-roundabout-flares report-dead-ends ignore-maxspeeds link-pois-to-ways location-autofill=0 styles: https://github.com/maning/osmphgps/tree/master/styles/default On Thu, Jul 7, 2011 at 4:01 PM, maning sambale wrote: > Got it! > > On Thu, Jul 7, 2011 at 3:58 PM, maning sambale > wrote: >> Where to download your branch? >> >> This is also a long standing issue for our PH maps. See issue report here: >> https://github.com/maning/osmphgps/issues/9 >> >> Basically, we have 6 "San Fernando" cities and city/address search >> only works for this case when using mkgmap r1867. >> >> On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe wrote: >>> Hi >>> >>> I am away this week so didn't have time to mention it, but I created a >>> branch (city-region-index) to fix the city/region bug since it turned >>> out to be quite a complex change. >>> >>> It is now working as far as I can tell. Upload to a device works without >>> mapsource crashes and giving both city and region in the search box works. >>> >>> ..Steve >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >>> >> >> >> >> -- >> cheers, >> maning >> -- >> "Freedom is still the most radical idea of all" -N.Branden >> wiki: http://esambale.wikispaces.com/ >> blog: http://epsg4253.wordpress.com/ >> -- >> > > > > -- > cheers, > maning > -- > "Freedom is still the most radical idea of all" -N.Branden > wiki: http://esambale.wikispaces.com/ > blog: http://epsg4253.wordpress.com/ > -- > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Got it! On Thu, Jul 7, 2011 at 3:58 PM, maning sambale wrote: > Where to download your branch? > > This is also a long standing issue for our PH maps. See issue report here: > https://github.com/maning/osmphgps/issues/9 > > Basically, we have 6 "San Fernando" cities and city/address search > only works for this case when using mkgmap r1867. > > On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe wrote: >> Hi >> >> I am away this week so didn't have time to mention it, but I created a >> branch (city-region-index) to fix the city/region bug since it turned >> out to be quite a complex change. >> >> It is now working as far as I can tell. Upload to a device works without >> mapsource crashes and giving both city and region in the search box works. >> >> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > > > -- > cheers, > maning > -- > "Freedom is still the most radical idea of all" -N.Branden > wiki: http://esambale.wikispaces.com/ > blog: http://epsg4253.wordpress.com/ > -- > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Where to download your branch? This is also a long standing issue for our PH maps. See issue report here: https://github.com/maning/osmphgps/issues/9 Basically, we have 6 "San Fernando" cities and city/address search only works for this case when using mkgmap r1867. On Thu, Jul 7, 2011 at 3:55 PM, Steve Ratcliffe wrote: > Hi > > I am away this week so didn't have time to mention it, but I created a > branch (city-region-index) to fix the city/region bug since it turned > out to be quite a complex change. > > It is now working as far as I can tell. Upload to a device works without > mapsource crashes and giving both city and region in the search box works. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > -- cheers, maning -- "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi, how can I merge them?! Thx Martin Am 07.07.2011 um 09:50 schrieb Steve Ratcliffe : > On 06/07/11 06:42, Martin wrote: >> Today I've found the mkgmap-city-region-index-r1984.jar-branch on the >> snapshot-site. Sound interesting, but the bounds-options didn't work. >> "Invalid option: 'boundsdirectory'" >> Or has something changed in the options? > > boundsdirectory is an option that is in the locator branch. > > You would have to merge the two branches together to get both sets of > changes. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi I am away this week so didn't have time to mention it, but I created a branch (city-region-index) to fix the city/region bug since it turned out to be quite a complex change. It is now working as far as I can tell. Upload to a device works without mapsource crashes and giving both city and region in the search box works. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 06/07/11 06:42, Martin wrote: > Today I've found the mkgmap-city-region-index-r1984.jar-branch on the > snapshot-site. Sound interesting, but the bounds-options didn't work. > "Invalid option: 'boundsdirectory'" > Or has something changed in the options? boundsdirectory is an option that is in the locator branch. You would have to merge the two branches together to get both sets of changes. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Today I've found the mkgmap-city-region-index-r1984.jar-branch on the snapshot-site. Sound interesting, but the bounds-options didn't work. "Invalid option: 'boundsdirectory'" Or has something changed in the options? Cheers Martin Am 30.06.2011 um 16:03 schrieb Steve Ratcliffe: > On 24/06/11 07:55, navmaps wrote: >> My 2nd attempt, running mkgmap again, also failed and produced the same >> error message. Index building on smaller sized maps (< 100 Mb) is ok. >> Yet, I suppose that somehow a bug has been introduced between locator >> 1969 and locator 1977 > > I'm going to check in a patch to fix this since trunk is broken anyway, > that largely goes back to how it was before the change that broke it. > > It will re-introduce a problem that this thread is about, but I have > seen that mdr27 does not look right, so I will then see if that makes > any difference with the original problem. > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 24/06/11 07:55, navmaps wrote: > My 2nd attempt, running mkgmap again, also failed and produced the same > error message. Index building on smaller sized maps (< 100 Mb) is ok. > Yet, I suppose that somehow a bug has been introduced between locator > 1969 and locator 1977 I'm going to check in a patch to fix this since trunk is broken anyway, that largely goes back to how it was before the change that broke it. It will re-introduce a problem that this thread is about, but I have seen that mdr27 does not look right, so I will then see if that makes any difference with the original problem. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi, is this problem already fixed and I missed some emails? Cheers, Martin Am 2011-06-26 14:56, schrieb Steve Ratcliffe: > Hi > > Ahh, I remembered a keyword that helped me find it. I had no idea it was > just Feb this year, thought is was years ago, so not looking in the > right place at all! > > It was this thread I was thinking of: > http://www.mkgmap.org.uk/pipermail/mkgmap-dev/2011q1/010163.html > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > is this problem already fixed and I missed some emails? The problem in Feb was fixed, there is currently a problem in the latest svn version with similar symptoms that is not fixed. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi Ahh, I remembered a keyword that helped me find it. I had no idea it was just Feb this year, thought is was years ago, so not looking in the right place at all! It was this thread I was thinking of: http://www.mkgmap.org.uk/pipermail/mkgmap-dev/2011q1/010163.html ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> Hi > >>> r1971 also crashes. I wonder if r1970 does not... > >> r1970 is ok and works. > > I am not surprised ;) > > I have managed to reproduce this with two tiles, each one by itself > transfers fine, but together they fail. > > I know that there was a bug like this a long time ago but I cannot find > the email exchanges where it was discussed. It would probably be very > helpful to know what I did to fix it then. > > Can anyone remember of find the problem in the list? > > ..Steve Do you mean: http://www.mkgmap.org.uk/pipermail/mkgmap-dev/2011q2/011612.html ? WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi >> r1971 also crashes. I wonder if r1970 does not... > r1970 is ok and works. I am not surprised ;) I have managed to reproduce this with two tiles, each one by itself transfers fine, but together they fail. I know that there was a bug like this a long time ago but I cannot find the email exchanges where it was discussed. It would probably be very helpful to know what I did to fix it then. Can anyone remember of find the problem in the list? ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Am 24.06.2011 17:14, schrieb WanMil: I tried to reproduce and got the same problem with the trunk r1973. So the problem must have been introduced with r1971/2/3 and is not locator specific. @Steve: any ideas? >>> >>> Yes, it is bound to be on trunk, I've just seen something that might be >>> a problem, so it may be easy to fix after all. I'll look into it. >>> >>> ..Steve >> >> Great! >> FYI: r1972 crashes too. Will continue to try r1971. >> >> WanMil > > r1971 also crashes. I wonder if r1970 does not... > WanMil r1970 is ok and works. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
>>> I tried to reproduce and got the same problem with the trunk r1973. So >>> the problem must have been introduced with r1971/2/3 and is not locator >>> specific. >>> >>> @Steve: any ideas? >> >> Yes, it is bound to be on trunk, I've just seen something that might be >> a problem, so it may be easy to fix after all. I'll look into it. >> >> ..Steve > > Great! > FYI: r1972 crashes too. Will continue to try r1971. > > WanMil r1971 also crashes. I wonder if r1970 does not... WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
>> I tried to reproduce and got the same problem with the trunk r1973. So >> the problem must have been introduced with r1971/2/3 and is not locator >> specific. >> >> @Steve: any ideas? > > Yes, it is bound to be on trunk, I've just seen something that might be > a problem, so it may be easy to fix after all. I'll look into it. > > ..Steve Great! FYI: r1972 crashes too. Will continue to try r1971. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 24/06/11 14:13, WanMil wrote: > I tried to reproduce and got the same problem with the trunk r1973. So > the problem must have been introduced with r1971/2/3 and is not locator > specific. > > @Steve: any ideas? Yes, it is bound to be on trunk, I've just seen something that might be a problem, so it may be easy to fix after all. I'll look into it. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 24/06/11 07:55, navmaps wrote: > My 2nd attempt, running mkgmap again, also failed and produced the same > error message. Index building on smaller sized maps (< 100 Mb) is ok. > Yet, I suppose that somehow a bug has been introduced between locator > 1969 and locator 1977 There were several fixes on trunk between those revisions that were merged in with 1977. It would be good to know which one it was. Unfortunately the error message means nothing to me, you get a similar message when there is anything it thinks is wrong. I will only be able to fix it if you can give me a way of reproducing it, preferably with a small example. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
I tried to reproduce and got the same problem with the trunk r1973. So the problem must have been introduced with r1971/2/3 and is not locator specific. @Steve: any ideas? WanMil > I can confirm this. > I've attached the error-message. > > Cheers > > > Am 24.06.2011 um 08:55 schrieb navmaps: > >> My 2nd attempt, running mkgmap again, also failed and produced the same >> error message. Index building on smaller sized maps (< 100 Mb) is ok. >> Yet, I suppose that somehow a bug has been introduced between locator >> 1969 and locator 1977 >> >> Johan >> >> >> On Thu, 23 Jun 2011 23:45:06 +0200, navmaps wrote: >>> However, for the first time (ever in my short mkgmap map making life) >>> I >>> get an error when Mapsource builds the index. Everything was fine >>> till >>> locator version 1969. Locator version 1977 gives a >>> MDR_TRIM_ADDR.CXX-440.6.16.3.0 error code. >>> >>> I'll try another test run tomorrow >>> >>> Cheers, Johan >>> >>> On Thu, 23 Jun 2011 15:17:47 +0200, WanMil wrote: Hi Steve, can you commit the patch or do you have some more ideas you want to implement and test first? WanMil > Hi > >> Now I get one result when searching for StreetX/CityX/Region1 >> which >> is >> good. When using Region2 I always get two results when searching >> for the >> StreetX/CityX/Region2 combination. > > > I tried with region 2 and 4 too. > > I found that if you search for street+city+region you always get > Region1 > as am extra (incorrect) result. But never any other region, so for > example if the street+city combination is not in region1 then you > just > get the correct item and not any of the other regions instead. > > So with lots of regions in a real map it is probably not much of a > problem. Anyway, I cannot see any way to fix it - it may even be a > mapsource bug, you never know. > > So I will just commit the previous patches which fix the main > problems. > > ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >>> >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
I can confirm this. I've attached the error-message. Cheers http://www.garmin.com/xmlschemas/ApplicationErrors/v1";> MDR_TRIM_ADDR.CXX 440 Products Installed: Name: Germany Id: 262189 Version: 1.0 Type: Standard MapInstall 3 14 4 0 Release Nov 10 2010, 16:27:21 sqa 1031 006-A0083-00 Windows XP Service Pack 3 x86, Processor Level: 6, Processors:1, Model: 15 Stepping: 6 1048048 2011-06-24T08:29:57Z MDR_TRIM_ADDR.CXX 440 Products Installed: Name: Germany Id: 262189 Version: 1.0 Type: Standard MapInstall 3 14 4 0 Release Nov 10 2010, 16:27:21 sqa 1031 006-A0083-00 Windows XP Service Pack 3 x86, Processor Level: 6, Processors:1, Model: 15 Stepping: 6 1048048 2011-06-24T08:31:17Z Am 24.06.2011 um 08:55 schrieb navmaps: > My 2nd attempt, running mkgmap again, also failed and produced the same > error message. Index building on smaller sized maps (< 100 Mb) is ok. > Yet, I suppose that somehow a bug has been introduced between locator > 1969 and locator 1977 > > Johan > > > On Thu, 23 Jun 2011 23:45:06 +0200, navmaps wrote: >> However, for the first time (ever in my short mkgmap map making life) >> I >> get an error when Mapsource builds the index. Everything was fine >> till >> locator version 1969. Locator version 1977 gives a >> MDR_TRIM_ADDR.CXX-440.6.16.3.0 error code. >> >> I'll try another test run tomorrow >> >> Cheers, Johan >> >> On Thu, 23 Jun 2011 15:17:47 +0200, WanMil wrote: >>> Hi Steve, >>> >>> can you commit the patch or do you have some more ideas you want to >>> implement and test first? >>> >>> WanMil >>> Hi > Now I get one result when searching for StreetX/CityX/Region1 > which > is > good. When using Region2 I always get two results when searching > for the > StreetX/CityX/Region2 combination. I tried with region 2 and 4 too. I found that if you search for street+city+region you always get Region1 as am extra (incorrect) result. But never any other region, so for example if the street+city combination is not in region1 then you just get the correct item and not any of the other regions instead. So with lots of regions in a real map it is probably not much of a problem. Anyway, I cannot see any way to fix it - it may even be a mapsource bug, you never know. So I will just commit the previous patches which fix the main problems. ..Steve >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
My 2nd attempt, running mkgmap again, also failed and produced the same error message. Index building on smaller sized maps (< 100 Mb) is ok. Yet, I suppose that somehow a bug has been introduced between locator 1969 and locator 1977 Johan On Thu, 23 Jun 2011 23:45:06 +0200, navmaps wrote: > However, for the first time (ever in my short mkgmap map making life) > I > get an error when Mapsource builds the index. Everything was fine > till > locator version 1969. Locator version 1977 gives a > MDR_TRIM_ADDR.CXX-440.6.16.3.0 error code. > > I'll try another test run tomorrow > > Cheers, Johan > > On Thu, 23 Jun 2011 15:17:47 +0200, WanMil wrote: >> Hi Steve, >> >> can you commit the patch or do you have some more ideas you want to >> implement and test first? >> >> WanMil >> >>> Hi >>> Now I get one result when searching for StreetX/CityX/Region1 which is good. When using Region2 I always get two results when searching for the StreetX/CityX/Region2 combination. >>> >>> >>> I tried with region 2 and 4 too. >>> >>> I found that if you search for street+city+region you always get >>> Region1 >>> as am extra (incorrect) result. But never any other region, so for >>> example if the street+city combination is not in region1 then you >>> just >>> get the correct item and not any of the other regions instead. >>> >>> So with lots of regions in a real map it is probably not much of a >>> problem. Anyway, I cannot see any way to fix it - it may even be a >>> mapsource bug, you never know. >>> >>> So I will just commit the previous patches which fix the main >>> problems. >>> >>> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
However, for the first time (ever in my short mkgmap map making life) I get an error when Mapsource builds the index. Everything was fine till locator version 1969. Locator version 1977 gives a MDR_TRIM_ADDR.CXX-440.6.16.3.0 error code. I'll try another test run tomorrow Cheers, Johan On Thu, 23 Jun 2011 15:17:47 +0200, WanMil wrote: > Hi Steve, > > can you commit the patch or do you have some more ideas you want to > implement and test first? > > WanMil > >> Hi >> >>> Now I get one result when searching for StreetX/CityX/Region1 which >>> is >>> good. When using Region2 I always get two results when searching >>> for the >>> StreetX/CityX/Region2 combination. >> >> >> I tried with region 2 and 4 too. >> >> I found that if you search for street+city+region you always get >> Region1 >> as am extra (incorrect) result. But never any other region, so for >> example if the street+city combination is not in region1 then you >> just >> get the correct item and not any of the other regions instead. >> >> So with lots of regions in a real map it is probably not much of a >> problem. Anyway, I cannot see any way to fix it - it may even be a >> mapsource bug, you never know. >> >> So I will just commit the previous patches which fix the main >> problems. >> >> ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > can you commit the patch or do you have some more ideas you want to > implement and test first? Sure, sorry, I thought that I had already done so. Thanks for reminding me. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi Steve, can you commit the patch or do you have some more ideas you want to implement and test first? WanMil > Hi > >> Now I get one result when searching for StreetX/CityX/Region1 which is >> good. When using Region2 I always get two results when searching for the >> StreetX/CityX/Region2 combination. > > > I tried with region 2 and 4 too. > > I found that if you search for street+city+region you always get Region1 > as am extra (incorrect) result. But never any other region, so for > example if the street+city combination is not in region1 then you just > get the correct item and not any of the other regions instead. > > So with lots of regions in a real map it is probably not much of a > problem. Anyway, I cannot see any way to fix it - it may even be a > mapsource bug, you never know. > > So I will just commit the previous patches which fix the main problems. > > ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> Hi > >> Now I get one result when searching for StreetX/CityX/Region1 which is >> good. When using Region2 I always get two results when searching for the >> StreetX/CityX/Region2 combination. > > > I tried with region 2 and 4 too. > > I found that if you search for street+city+region you always get Region1 > as am extra (incorrect) result. But never any other region, so for > example if the street+city combination is not in region1 then you just > get the correct item and not any of the other regions instead. > > So with lots of regions in a real map it is probably not much of a > problem. Anyway, I cannot see any way to fix it - it may even be a > mapsource bug, you never know. > > So I will just commit the previous patches which fix the main problems. Yep, that's fine! I think this fix will bring the index generation a big step forward. WanMil > > ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > Now I get one result when searching for StreetX/CityX/Region1 which is > good. When using Region2 I always get two results when searching for the > StreetX/CityX/Region2 combination. I tried with region 2 and 4 too. I found that if you search for street+city+region you always get Region1 as am extra (incorrect) result. But never any other region, so for example if the street+city combination is not in region1 then you just get the correct item and not any of the other regions instead. So with lots of regions in a real map it is probably not much of a problem. Anyway, I cannot see any way to fix it - it may even be a mapsource bug, you never know. So I will just commit the previous patches which fix the main problems. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Okay, I've made a new map of Germany on a Windows-machine. On Mapsource I can find a few more streets, which I couldn't found before. But, on my Garmin, I still couldn't find the streets like Schafblumenhalde in Horb. If I leave out the City, I can find the street, but when I click on it, the Garmin says, that it couldn't found this street. When I make just the tile containing the Schafblumenhalde, I can find it on Mapsource, my Garmin but not in Basecamp (used Gmapi-Builder and the Garmin Map Converter). Strange things... Am 02.06.2011 um 19:51 schrieb Martin: > You are right... I've converted the map, produced on the windows machine, to > the Mac-Format, and the same behavior: > I couldn't find the streets nor the city, which is outside of the tile. > I didn't think this is an error from the Garmin-Software, I think there is a > difference between the index-table produced with mkgmap and a original > Garmin-Map. > Maybe a character or something like this. > > Cheers > Martin > > Am 02.06.2011 um 18:24 schrieb Clinton Gladstone: > >> On Jun 2, 2011, at 16:12, Martin wrote: >> >>> I use the following commands: >>> python gmapi-builder.py -t osmmap.tdb -b osmmap.img -s ./master/basemap.TYP >>> -i osmmap.mdx -m osmmap_mdr.img osmmap.img 63240*.img osmmap_mdr.img >> >> I compared the output from gmapi-builder with that produced by Garmin's >> MapConverter program: they appear to be identical. >> >> Perhaps you could do the following to help isolate the problem: >> >> - Run Garmin's MapConverter on the map which you created on your Windows >> machine. >> >> - Install the resulting gmapi map on your Mac OS machine. >> >> - Does the problem still occur? >> >> If so, it may be a general problem with converting maps to gmapi format, or >> (perhaps more likely) it could be a problem in Garmin's Mac OS programs >> (MapInstall, Basecamp, etc.). >> >> - Which versions of MapInstall and Basecamp do you have on your Mac? >> >> Cheers. >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi WanMil > Now I get one result when searching for StreetX/CityX/Region1 which is > good. When using Region2 I always get two results when searching for the > StreetX/CityX/Region2 combination. Weird, I see that on my up-to-date mapsource but not on the older version. I'll have another think! ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
>> Steve, >> >> no matter if there are some minor issues left your patch is a big >> improvement and is worth being committed. > > OK thanks, but I now have a patch that works with your new example. > > Attached. > > ..Steve That's one further step :-) Now I get one result when searching for StreetX/CityX/Region1 which is good. When using Region2 I always get two results when searching for the StreetX/CityX/Region2 combination. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
You are right... I've converted the map, produced on the windows machine, to the Mac-Format, and the same behavior: I couldn't find the streets nor the city, which is outside of the tile. I didn't think this is an error from the Garmin-Software, I think there is a difference between the index-table produced with mkgmap and a original Garmin-Map. Maybe a character or something like this. Cheers Martin Am 02.06.2011 um 18:24 schrieb Clinton Gladstone: > On Jun 2, 2011, at 16:12, Martin wrote: > >> I use the following commands: >> python gmapi-builder.py -t osmmap.tdb -b osmmap.img -s ./master/basemap.TYP >> -i osmmap.mdx -m osmmap_mdr.img osmmap.img 63240*.img osmmap_mdr.img > > I compared the output from gmapi-builder with that produced by Garmin's > MapConverter program: they appear to be identical. > > Perhaps you could do the following to help isolate the problem: > > - Run Garmin's MapConverter on the map which you created on your Windows > machine. > > - Install the resulting gmapi map on your Mac OS machine. > > - Does the problem still occur? > > If so, it may be a general problem with converting maps to gmapi format, or > (perhaps more likely) it could be a problem in Garmin's Mac OS programs > (MapInstall, Basecamp, etc.). > > - Which versions of MapInstall and Basecamp do you have on your Mac? > > Cheers. > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On Jun 2, 2011, at 16:12, Martin wrote: > I use the following commands: > python gmapi-builder.py -t osmmap.tdb -b osmmap.img -s ./master/basemap.TYP > -i osmmap.mdx -m osmmap_mdr.img osmmap.img 63240*.img osmmap_mdr.img I compared the output from gmapi-builder with that produced by Garmin's MapConverter program: they appear to be identical. Perhaps you could do the following to help isolate the problem: - Run Garmin's MapConverter on the map which you created on your Windows machine. - Install the resulting gmapi map on your Mac OS machine. - Does the problem still occur? If so, it may be a general problem with converting maps to gmapi format, or (perhaps more likely) it could be a problem in Garmin's Mac OS programs (MapInstall, Basecamp, etc.). - Which versions of MapInstall and Basecamp do you have on your Mac? Cheers. ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 02/06/11 15:13, WanMil wrote:
Steve,
no matter if there are some minor issues left your patch is a big
improvement and is worth being committed.
OK thanks, but I now have a patch that works with your new example.
Attached.
..Steve
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision 1870)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision )
@@ -64,16 +64,18 @@
cities.clear();
int count = 0;
int lastMapId = 0;
+ int lastRegion = 0;
String lastName = null;
for (SortKey key : sortKeys) {
Mdr5Record c = key.getObject();
- if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName)) {
+ if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName) || c.getRegionIndex() != lastRegion) {
count++;
c.setGlobalCityIndex(count);
cities.add(c);
lastName = c.getName();
lastMapId = c.getMapIndex();
+lastRegion = c.getRegionIndex();
} else {
c.setGlobalCityIndex(count);
}
@@ -116,12 +118,14 @@
*/
private void fix20() {
String lastName = null;
+ int lastRegion = 0;
for (int index = 1; index < cities.size(); index++) {
Mdr5Record city = cities.get(index-1);
+ int region = city.getRegionIndex();
assert city.getGlobalCityIndex() == index : index + "/" + city.getGlobalCityIndex();
String name = city.getName();
- if (name.equals(lastName)) {
+ if (name.equals(lastName) && region == lastRegion) {
int last = findLastRepeat(index, name);
int mdr20val = findMin20(index - 1, last);
for (int j = index-1; j < last; j++) {
@@ -133,6 +137,7 @@
}
lastName = name;
+ lastRegion = region;
}
}
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (revision 1870)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (revision )
@@ -73,19 +73,23 @@
String lastCityName = null;
int lastMapid = 0;
int record = 0;
+ int lastRegion = 0;
for (SortKey key : keys) {
Mdr7Record street = key.getObject();
int mapid = street.getMapIndex();
String name = street.getName();
String cityName = street.getCity().getName();
- if (mapid != lastMapid || !name.equals(lastName) || !cityName.equals(lastCityName)) {
+ int region = street.getCity().getRegionIndex();
+
+ if (mapid != lastMapid || !name.equals(lastName) || !cityName.equals(lastCityName) || region != lastRegion) {
record++;
streets.add(street);
lastMapid = mapid;
lastName = name;
lastCityName = cityName;
+lastRegion = region;
}
int gci = street.getCity().getGlobalCityIndex();
___
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On Jun 2, 2011, at 16:12, Martin wrote: > I couldn't find any version in the file, so I attached the file. Thanks, I'll take a look and see if I can find anything. Right now I'm comparing the output produced by Garmin's MapConvert with that of gmapi-builder. Cheers. ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Steve, no matter if there are some minor issues left your patch is a big improvement and is worth being committed. WanMil P.S.: I forgot to mention that I have done my tests in combination with the translit_first.patch. > Looks good to me. > There is one crude thing left but hopefully you can fix that too. > > I have modified the testcase (attached). It contains now one street with > each permutation of > name=Street1 / Street2 > city=City1 / City2 > region=Region / Region2 > > Performing some searches gives the following results: > * Searching for street only gives the expected 4 results > * Searching for a street/city combination gives the expected 2 results > * Searching for a street/region combination gives the expected 2 results > * Adding a country to the searches is working too (might have to add > another permutation with a 2nd country) > > * Searching for street, city and region gives one expected result only > when searching for Street1/City2/Region1 or Street2/City2/Region1. All > other combinations return two results with both regions. > > WanMil > >> >> Hi >> >>> Did something happen to the patch file, or am I missing something? >> >> Yes something weird happened to the patch file. There was a change to >> Mdr20. Trying again... >> >> ..Steve >> >> >> >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hello Clinton,
I couldn't find any version in the file, so I attached the file.
I use the following commands:
python gmapi-builder.py -t osmmap.tdb -b osmmap.img -s ./master/basemap.TYP -i
osmmap.mdx -m osmmap_mdr.img osmmap.img 63240*.img osmmap_mdr.img
Thanks
Martin
#!/usr/bin/python
"""
Simple convertor for files in Garmins MapSource format to the directory
structure RoadTrip for OS X uses.
Copyright (c) 2009, Berteun Damman
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the OpenStreetMap Project nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE AUTHOR ''AS IS'' AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
First of all: This program is not a replacement for Garmins MapConverter.exe.
It was specifically written to convert a bunch of .IMG and .TDB files of
the OpenStreetmap project to a .gmapi directory structure which can be
installed under OS X. As such it is geared towards the versions of the format
used by OSM. This means it will most certain not work on non-OSM files (this
has not been tested in anyway!).
The conversion itself is fairly simple, under Windows you have:
* Registry entries indicating the TDB-file, the Base Image and the Image dir,
and optionally a TYP file and some other types.
* A single TDB-file
* A whole collection of IMG files.
Under Mac you have:
* An XML file which contains the information of the Windows Register
* For each IMG file a directory with the same basename as the IMG-file,
and with the subfiles of this IMG extracted into this directory; the IMG-files
are in fact containers (they are similar to disk images).
This program does the conversion.
"""
import logging
import optparse
import os
import shutil
import struct
import sys
import StringIO
# Logger setup
logging.NORMAL = logging.INFO + 5
logging.addLevelName(logging.NORMAL, 'NORMAL')
logger = logging.getLogger('logger')
logger.setLevel(logging.DEBUG)
# Program options can increase the verbosity
cons = logging.StreamHandler(sys.stdout)
cons.setLevel(logging.NORMAL)
logger.addHandler(cons)
def error_exit(msg):
sys.exit(msg)
def write(msg, *args, **kwargs):
logger.log(logging.NORMAL, msg, *args, **kwargs)
class EndOfFile(IOError):
pass
class FileFormatError(Exception):
pass
# Auxiliary functions that make life easier to read bytes, shorts, int and
# so on from a file.
def getX(length, fmt):
def get(f):
v = f.read(length)
if len(v) < length:
raise EndOfFile("End of file reached on '%s'." % f.name)
return struct.unpack(fmt, v)[0]
return get
# See the documentation of struct for an explanation of these
# format specifiers. Notably, the < indicates little endian
# format.
get_short = getX(2, '> 7, l[2] & 0x7F
#v = -sign * (1<<23) + (l[2] << 16) + (l[1] << 8) + l[0]
def todegrees(n):
return (n * 360.0) / (2 ** 32)
# A TDB file consists of a sequence of blocks, the format of each
# block is simply:
# 1 Byte: Block ID
# 2 Bytes: Block length (l):
# l Bytes: Data
class Block(object):
def __init__(self, f):
self.bid = get_byte(f)
self.length = get_ushort(f)
self.data = StringIO.StringIO(f.read(self.length))
if len(self.data.buf) < self.length:
raise EndOfFile("End of file reached early on '%s'." % f.name)
class TDBFile(object):
"""This represents the TDBFile with its known blocks, such as
the header, copyright block, overview block and detailed maps."""
def __init__(self, filename):
self.f = open(filename, 'rb')
self.header_block = None
self.copyright_block =
Re: [mkgmap-dev] Index and equally named cities
On Jun 2, 2011, at 15:49, Martin wrote: > I've made a map on a Windows-Machine, using the same files I used on my > Macbook and the nsis-compiler... And now it works... > Damn... So the problem comes from the gmapi-builder. > Anyone here who understand python to fix this problem?! > And sorry for bothering you with this problem. I've never expected that the > problem comes from gmapi. Which version of gmapi-builder are you using? Please also send the command lines you use. Cheers. ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
So I've made a map on a Windows-Machine, using the same files I used on my Macbook and the nsis-compiler... And now it works... Damn... So the problem comes from the gmapi-builder. Anyone here who understand python to fix this problem?! And sorry for bothering you with this problem. I've never expected that the problem comes from gmapi. Cheers Martin Am 02.06.2011 um 14:59 schrieb Martin: > Hallo WanMil, > > could you please give me your settings for the splitter and mkgmap. > Maybe something goes wrong with gmapi.py?! > > Thanks > Martin > > Am 02.06.2011 um 14:36 schrieb WanMil: > >> Martin, Steve, >> >> compiling only the given tile I cannot reproduce the problem. MapSource >> search is ok and searching on the Oregon finds Schafblumenhalde too. >> >> Anyhow I have reviewed some relevant source code parts and have a >> question Steve maybe can answer. >> >> The class PlacesFile contains the methods: >> City createCity(Country country, String name, boolean unique) >> City createCity(Region region, String name, boolean unique) >> >> They are called when creating a city from a place tag (unique=true) and >> when creating a road (unique=false). >> I don't understand in detail what the unique parameter should do. >> >> WanMil >> >> >>> Hello Steve, >>> >>> thank you for your fast reply. >>> My problems with not findable streets still exist. >>> If a town is devided into 2 or more pieces, only streets could be find on >>> the tile which contains the place-tag. Tested on Basecamp for Mac and my >>> Garmin Oregon. On Mapsource it seems to work. I don't know how the index is >>> generated (tile by tile or all-in-one), but I hope this helps you to find >>> the problem. >>> When I copy the place-tag into this tile then I can find all that streets. >>> >>> Testcase: >>> Tile: >>> 63240403: 2250752,378880 to 2271232,403456 >>> # : 48.295898,8.129883 to 48.735352,8.657227 >>> >>> Streets that are not findable: e.g Schafblumenhalde (City: Horb am Neckar) >>> >>> Generated with locator-branch and your patched jar-file. >>> >>> Cheers >>> Martin >>> >>> >>> Am 01.06.2011 um 12:52 schrieb Steve Ratcliffe: >>> On 01/06/11 09:24, Martin wrote: > Hello, > > how can I patch a precompiled jar-file? I would like to test this patch > with the locator-branch to see if this fix some problems I've found. Or > is it already included in the locator-branch? Its not included in the branch yet. Here is a jar with the patch applied. URL: http://files.mkgmap.org.uk/download/23/mkgmap.jar Description: mkgmap from locator branch with city_region3.patch applied. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >>> >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hallo WanMil, could you please give me your settings for the splitter and mkgmap. Maybe something goes wrong with gmapi.py?! Thanks Martin Am 02.06.2011 um 14:36 schrieb WanMil: > Martin, Steve, > > compiling only the given tile I cannot reproduce the problem. MapSource > search is ok and searching on the Oregon finds Schafblumenhalde too. > > Anyhow I have reviewed some relevant source code parts and have a > question Steve maybe can answer. > > The class PlacesFile contains the methods: > City createCity(Country country, String name, boolean unique) > City createCity(Region region, String name, boolean unique) > > They are called when creating a city from a place tag (unique=true) and > when creating a road (unique=false). > I don't understand in detail what the unique parameter should do. > > WanMil > > >> Hello Steve, >> >> thank you for your fast reply. >> My problems with not findable streets still exist. >> If a town is devided into 2 or more pieces, only streets could be find on >> the tile which contains the place-tag. Tested on Basecamp for Mac and my >> Garmin Oregon. On Mapsource it seems to work. I don't know how the index is >> generated (tile by tile or all-in-one), but I hope this helps you to find >> the problem. >> When I copy the place-tag into this tile then I can find all that streets. >> >> Testcase: >> Tile: >> 63240403: 2250752,378880 to 2271232,403456 >> # : 48.295898,8.129883 to 48.735352,8.657227 >> >> Streets that are not findable: e.g Schafblumenhalde (City: Horb am Neckar) >> >> Generated with locator-branch and your patched jar-file. >> >> Cheers >> Martin >> >> >> Am 01.06.2011 um 12:52 schrieb Steve Ratcliffe: >> >>> On 01/06/11 09:24, Martin wrote: Hello, how can I patch a precompiled jar-file? I would like to test this patch with the locator-branch to see if this fix some problems I've found. Or is it already included in the locator-branch? >>> >>> Its not included in the branch yet. >>> >>> Here is a jar with the patch applied. >>> >>> URL: http://files.mkgmap.org.uk/download/23/mkgmap.jar >>> >>> Description: mkgmap from locator branch with city_region3.patch applied. >>> >>> >>> ..Steve >>> ___ >>> mkgmap-dev mailing list >>> mkgmap-dev@lists.mkgmap.org.uk >>> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >>> >> >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Martin, Steve, compiling only the given tile I cannot reproduce the problem. MapSource search is ok and searching on the Oregon finds Schafblumenhalde too. Anyhow I have reviewed some relevant source code parts and have a question Steve maybe can answer. The class PlacesFile contains the methods: City createCity(Country country, String name, boolean unique) City createCity(Region region, String name, boolean unique) They are called when creating a city from a place tag (unique=true) and when creating a road (unique=false). I don't understand in detail what the unique parameter should do. WanMil > Hello Steve, > > thank you for your fast reply. > My problems with not findable streets still exist. > If a town is devided into 2 or more pieces, only streets could be find on the > tile which contains the place-tag. Tested on Basecamp for Mac and my Garmin > Oregon. On Mapsource it seems to work. I don't know how the index is > generated (tile by tile or all-in-one), but I hope this helps you to find the > problem. > When I copy the place-tag into this tile then I can find all that streets. > > Testcase: > Tile: > 63240403: 2250752,378880 to 2271232,403456 > # : 48.295898,8.129883 to 48.735352,8.657227 > > Streets that are not findable: e.g Schafblumenhalde (City: Horb am Neckar) > > Generated with locator-branch and your patched jar-file. > > Cheers > Martin > > > Am 01.06.2011 um 12:52 schrieb Steve Ratcliffe: > >> On 01/06/11 09:24, Martin wrote: >>> Hello, >>> >>> how can I patch a precompiled jar-file? I would like to test this patch >>> with the locator-branch to see if this fix some problems I've found. Or is >>> it already included in the locator-branch? >> >> Its not included in the branch yet. >> >> Here is a jar with the patch applied. >> >> URL: http://files.mkgmap.org.uk/download/23/mkgmap.jar >> >> Description: mkgmap from locator branch with city_region3.patch applied. >> >> >> ..Steve >> ___ >> mkgmap-dev mailing list >> mkgmap-dev@lists.mkgmap.org.uk >> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev >> > > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hello Steve, thank you for your fast reply. My problems with not findable streets still exist. If a town is devided into 2 or more pieces, only streets could be find on the tile which contains the place-tag. Tested on Basecamp for Mac and my Garmin Oregon. On Mapsource it seems to work. I don't know how the index is generated (tile by tile or all-in-one), but I hope this helps you to find the problem. When I copy the place-tag into this tile then I can find all that streets. Testcase: Tile: 63240403: 2250752,378880 to 2271232,403456 # : 48.295898,8.129883 to 48.735352,8.657227 Streets that are not findable: e.g Schafblumenhalde (City: Horb am Neckar) Generated with locator-branch and your patched jar-file. Cheers Martin Am 01.06.2011 um 12:52 schrieb Steve Ratcliffe: > On 01/06/11 09:24, Martin wrote: >> Hello, >> >> how can I patch a precompiled jar-file? I would like to test this patch with >> the locator-branch to see if this fix some problems I've found. Or is it >> already included in the locator-branch? > > Its not included in the branch yet. > > Here is a jar with the patch applied. > > URL: http://files.mkgmap.org.uk/download/23/mkgmap.jar > > Description: mkgmap from locator branch with city_region3.patch applied. > > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Looks good to me. There is one crude thing left but hopefully you can fix that too. I have modified the testcase (attached). It contains now one street with each permutation of name=Street1 / Street2 city=City1 / City2 region=Region / Region2 Performing some searches gives the following results: * Searching for street only gives the expected 4 results * Searching for a street/city combination gives the expected 2 results * Searching for a street/region combination gives the expected 2 results * Adding a country to the searches is working too (might have to add another permutation with a 2nd country) * Searching for street, city and region gives one expected result only when searching for Street1/City2/Region1 or Street2/City2/Region1. All other combinations return two results with both regions. WanMil Hi Did something happen to the patch file, or am I missing something? Yes something weird happened to the patch file. There was a change to Mdr20. Trying again... ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On 01/06/11 09:24, Martin wrote: > Hello, > > how can I patch a precompiled jar-file? I would like to test this patch with > the locator-branch to see if this fix some problems I've found. Or is it > already included in the locator-branch? Its not included in the branch yet. Here is a jar with the patch applied. URL: http://files.mkgmap.org.uk/download/23/mkgmap.jar Description: mkgmap from locator branch with city_region3.patch applied. ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hello, how can I patch a precompiled jar-file? I would like to test this patch with the locator-branch to see if this fix some problems I've found. Or is it already included in the locator-branch? Thanks Martin Am 30.05.2011 um 23:56 schrieb Steve Ratcliffe: > > Hi > >> Did something happen to the patch file, or am I missing something? > > Yes something weird happened to the patch file. There was a change to Mdr20. > Trying again... > > ..Steve > ___ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi
Did something happen to the patch file, or am I missing something?
Yes something weird happened to the patch file. There was a change to
Mdr20. Trying again...
..Steve
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision 1957)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (working copy)
@@ -64,16 +64,18 @@
cities.clear();
int count = 0;
int lastMapId = 0;
+ int lastRegion = 0;
String lastName = null;
for (SortKey key : sortKeys) {
Mdr5Record c = key.getObject();
- if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName)) {
+ if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName) || c.getRegionIndex() != lastRegion) {
count++;
c.setGlobalCityIndex(count);
cities.add(c);
lastName = c.getName();
lastMapId = c.getMapIndex();
+lastRegion = c.getRegionIndex();
} else {
c.setGlobalCityIndex(count);
}
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (revision 1957)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (working copy)
@@ -73,19 +73,23 @@
String lastCityName = null;
int lastMapid = 0;
int record = 0;
+ int lastRegion = 0;
for (SortKey key : keys) {
Mdr7Record street = key.getObject();
int mapid = street.getMapIndex();
String name = street.getName();
String cityName = street.getCity().getName();
- if (mapid != lastMapid || !name.equals(lastName) || !cityName.equals(lastCityName)) {
+ int region = street.getCity().getRegionIndex();
+
+ if (mapid != lastMapid || !name.equals(lastName) || !cityName.equals(lastCityName) || region != lastRegion) {
record++;
streets.add(street);
lastMapid = mapid;
lastName = name;
lastCityName = cityName;
+lastRegion = region;
}
int gci = street.getCity().getGlobalCityIndex();
___
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
On May 30, 2011, at 23:23, Steve Ratcliffe wrote: > The attached patch seems to work better. > > ..Steve > This patch appears to be identical to the first one, with the exception of an empty src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java patch. Did something happen to the patch file, or am I missing something? Cheers. ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi
I've just tried with an up-to-date version of mapsource and I see the
same thing. Back to the drawing board then!
The attached patch seems to work better.
..Steve
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision 1870)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision )
@@ -64,16 +64,18 @@
cities.clear();
int count = 0;
int lastMapId = 0;
+ int lastRegion = 0;
String lastName = null;
for (SortKey key : sortKeys) {
Mdr5Record c = key.getObject();
- if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName)) {
+ if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName) || c.getRegionIndex() != lastRegion) {
count++;
c.setGlobalCityIndex(count);
cities.add(c);
lastName = c.getName();
lastMapId = c.getMapIndex();
+lastRegion = c.getRegionIndex();
} else {
c.setGlobalCityIndex(count);
}
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (revision 1870)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr20.java (revision )
@@ -1,0 +1,0 @@
___
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi > with the patch I get the street from Region 2 only when I search for > "Hauptstrasse" and "Testcity". It seems as if the city "Testcity" in the > search box is linked to the city in Region 2 using the patch and to the > city in Region 1 without this patch. I've just tried with an up-to-date version of mapsource and I see the same thing. Back to the drawing board then! ..Steve ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
> Hi > >> Steve, do you have any idea where the problem might be located? > > Well it might be that there should be separate city entries for each > region, as in attached patch. > > The patch works for the given example, although I wouldn't be surprised > if it caused another problem somewhere else - but give it a go see what > happens! > > ..Steve > Hi, with the patch I get the street from Region 2 only when I search for "Hauptstrasse" and "Testcity". It seems as if the city "Testcity" in the search box is linked to the city in Region 2 using the patch and to the city in Region 1 without this patch. Searching for "Hauptstrasse / Region1" and "Hauptstrasse / Region2" without the cityname works fine. WanMil ___ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Re: [mkgmap-dev] Index and equally named cities
Hi
Steve, do you have any idea where the problem might be located?
Well it might be that there should be separate city entries for each
region, as in attached patch.
The patch works for the given example, although I wouldn't be surprised
if it caused another problem somewhere else - but give it a go see what
happens!
..Steve
Index: src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java
===
--- src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision 1870)
+++ src/uk/me/parabola/imgfmt/app/mdr/Mdr5.java (revision )
@@ -64,16 +64,18 @@
cities.clear();
int count = 0;
int lastMapId = 0;
+ int lastRegion = 0;
String lastName = null;
for (SortKey key : sortKeys) {
Mdr5Record c = key.getObject();
- if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName)) {
+ if (c.getMapIndex() != lastMapId || !c.getName().equals(lastName) || c.getRegionIndex() != lastRegion) {
count++;
c.setGlobalCityIndex(count);
cities.add(c);
lastName = c.getName();
lastMapId = c.getMapIndex();
+lastRegion = c.getRegionIndex();
} else {
c.setGlobalCityIndex(count);
}
___
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
