MDR and city streets on different tile from POI place=city
Hi Gerd I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case-significance for the unicode/mdr errors) and finding discrepancies. I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing. If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems. This might be the reason for not being able to find some streets by city, rather than problems with case. I'm still trying to work it all out. Ticker
Hi Gerd Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase() Ticker On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Ticker, no idea what your problem is. Address search seems to work fine without any city POI. I think the code just tries to avoid duplicated entries in the LBL file? Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Dienstag, 2. November 2021 19:31 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city Hi Gerd Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase() Ticker On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd The scenario is: --lower-case, 2 tiles, mkmap:city "Aa" area that spans the tiles", streets with this mkgmap:city in either/both tiles, the City POI "Aa" on one of the tiles, another city, eg "AA" in the tile without the "Aa" POI, streets for this. I am finding streets in the tile with both regions are all assigned to one or the other and this behaviour is effected by the presence of the City POI. PlacesFile::CreateCity(not unique) will return the same City record for both "Aa" and "AA" so I propose to remove the toUpperCase() from it when --lower-case (unless ascii/cp0 which forces upper case) Ticker On Wed, 2021-11-03 at 09:17 +0000, Gerd Petermann wrote:
Hi Ticker,
no idea what your problem is. Address search seems to work fine without any city POI. I think the code just tries to avoid duplicated entries in the LBL file?
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Dienstag, 2. November 2021 19:31 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase()
Ticker
On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Ticker, yes, might be better to change that to be more precise. Question is if Aa and AA are two different cities or just different spellings of the same city. The latter is much more likely, so it seems justified to use only one. We have more code which tries to ignore case, e.g. for --housenumbers I often use String.equalsIgnoreCase() to work around different spelling in road names and addr:street or place names and addr:place. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Mittwoch, 3. November 2021 11:40 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city Hi Gerd The scenario is: --lower-case, 2 tiles, mkmap:city "Aa" area that spans the tiles", streets with this mkgmap:city in either/both tiles, the City POI "Aa" on one of the tiles, another city, eg "AA" in the tile without the "Aa" POI, streets for this. I am finding streets in the tile with both regions are all assigned to one or the other and this behaviour is effected by the presence of the City POI. PlacesFile::CreateCity(not unique) will return the same City record for both "Aa" and "AA" so I propose to remove the toUpperCase() from it when --lower-case (unless ascii/cp0 which forces upper case) Ticker On Wed, 2021-11-03 at 09:17 +0000, Gerd Petermann wrote:
Hi Ticker,
no idea what your problem is. Address search seems to work fine without any city POI. I think the code just tries to avoid duplicated entries in the LBL file?
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Dienstag, 2. November 2021 19:31 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase()
Ticker
On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd As per your example with "De Wijk", it could be the same city and different cities and difficult to decide programatically - esp. if different tiles. My attempts to cope with this in the MDR indexes didn't seem to work and I've abandoned this idea. I'd like to release a minimal patch that fixes the 2 assertion crashes. With --lower-case, which streets are attached to the different case variants of a city will not bear scrutiny - I don't think it did before. Then we should considering case-sensitivity of street, city, region and country as a new issue. Ticker On Wed, 2021-11-03 at 14:27 +0000, Gerd Petermann wrote:
Hi Ticker,
yes, might be better to change that to be more precise. Question is if Aa and AA are two different cities or just different spellings of the same city. The latter is much more likely, so it seems justified to use only one. We have more code which tries to ignore case, e.g. for --housenumbers I often use String.equalsIgnoreCase() to work around different spelling in road names and addr:street or place names and addr:place.
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Mittwoch, 3. November 2021 11:40 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
The scenario is: --lower-case, 2 tiles, mkmap:city "Aa" area that spans the tiles", streets with this mkgmap:city in either/both tiles, the City POI "Aa" on one of the tiles, another city, eg "AA" in the tile without the "Aa" POI, streets for this.
I am finding streets in the tile with both regions are all assigned to one or the other and this behaviour is effected by the presence of the City POI.
PlacesFile::CreateCity(not unique) will return the same City record for both "Aa" and "AA" so I propose to remove the toUpperCase() from it when --lower-case (unless ascii/cp0 which forces upper case)
Ticker
On Wed, 2021-11-03 at 09:17 +0000, Gerd Petermann wrote:
Hi Ticker,
no idea what your problem is. Address search seems to work fine without any city POI. I think the code just tries to avoid duplicated entries in the LBL file?
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Dienstag, 2. November 2021 19:31 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase()
Ticker
On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Ticker, I fear you lost me. My current understanding is that it doesn't matter much if there is one city or two as long as the index structures are correct and Garmin software finds the objects that exist. MapSource seems to ignore the case differences when it builds the list of city names in the drop down lists, it only showed one entry in my tests. I've not yet tried to understand which variant is choosen, might be the first in the MDR5 section. Anyway, it presents results with different spelling and they all seem to be correct as the correct object is shown when I click on an entry. Just post the minimal patch(es) and describe the expected effect(s) . Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Mittwoch, 3. November 2021 17:47 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city Hi Gerd As per your example with "De Wijk", it could be the same city and different cities and difficult to decide programatically - esp. if different tiles. My attempts to cope with this in the MDR indexes didn't seem to work and I've abandoned this idea. I'd like to release a minimal patch that fixes the 2 assertion crashes. With --lower-case, which streets are attached to the different case variants of a city will not bear scrutiny - I don't think it did before. Then we should considering case-sensitivity of street, city, region and country as a new issue. Ticker On Wed, 2021-11-03 at 14:27 +0000, Gerd Petermann wrote:
Hi Ticker,
yes, might be better to change that to be more precise. Question is if Aa and AA are two different cities or just different spellings of the same city. The latter is much more likely, so it seems justified to use only one. We have more code which tries to ignore case, e.g. for --housenumbers I often use String.equalsIgnoreCase() to work around different spelling in road names and addr:street or place names and addr:place.
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Mittwoch, 3. November 2021 11:40 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
The scenario is: --lower-case, 2 tiles, mkmap:city "Aa" area that spans the tiles", streets with this mkgmap:city in either/both tiles, the City POI "Aa" on one of the tiles, another city, eg "AA" in the tile without the "Aa" POI, streets for this.
I am finding streets in the tile with both regions are all assigned to one or the other and this behaviour is effected by the presence of the City POI.
PlacesFile::CreateCity(not unique) will return the same City record for both "Aa" and "AA" so I propose to remove the toUpperCase() from it when --lower-case (unless ascii/cp0 which forces upper case)
Ticker
On Wed, 2021-11-03 at 09:17 +0000, Gerd Petermann wrote:
Hi Ticker,
no idea what your problem is. Address search seems to work fine without any city POI. I think the code just tries to avoid duplicated entries in the LBL file?
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Ticker Berkin <rwb-mkgmap@jagit.co.uk> Gesendet: Dienstag, 2. November 2021 19:31 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] MDR and city streets on different tile from POI place=city
Hi Gerd
Actually, a major part of my problem is probably app/lbl/PlacesFile with its toUpperCase()
Ticker
On Tue, 2021-11-02 at 16:52 +0000, Ticker Berkin wrote:
Hi Gerd
I've been checking the street list in the Mdr20 index to make sure they are attached to the correct city (as part of checking case- significance for the unicode/mdr errors) and finding discrepancies.
I think this is because combiners/MdrBuilder, on a tile by tile basis takes the POI city list as the basis for Mdr5Records for attaching to streets and passing on for MDR processing.
If the POI is in an adjacent tile to a road with the same mkgmap:city, there seem to be problems.
This might be the reason for not being able to find some streets by city, rather than problems with case.
I'm still trying to work it all out.
Ticker
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
participants (2)
-
Gerd Petermann -
Ticker Berkin