Hi Steve, On Wed, Jun 24, 2009 at 12:56 AM, Steve Ratcliffe < steve@parabola.demon.co.uk> wrote:

...

With no options (i.e. no --code-page), transliteration works but resulting map is in lowercase letters. This works in MapSource but I get only the upper-case letters as abbreviated street names/poi in Garmin Mobile XT.

The only translitterations currently in mkgmap are from unicode to ascii. It is possible to target other code-pages if it makes sense and you can supply me with the transliteration table.

I'm not sure how you get lowercase letters. Mapsource shows upper-case letters in lowercase, but they are always upper case in the map itself.

What version are you using? There was a problem with an older version where you only saw the first letter of each word on certain devices.

I am using r1064 and command line I use is: java -Xmx1024m -ea -jar mkgmap.jar -n 62233665 --family-id=773 --series-name="OSM Macedonia-INT" \ --family-name="OSM_MKD_INT" --overview-mapname=40000773 --net --route --country-name=MACEDONIA \ --country-abbr=MKD --road-name-pois --add-pois-to-areas --no-poi-address --tdbfile --gmapsupp \ --description="OSM_MKD_INT" --name-tag-list="int_name,name:en,name" OSM_MKD.osm Which produces gmasupp.img (and same with 62233665.img) with lowercase letters inside In my mobile device, only capital letters are shown, i.e. first letters of names of cities, streets on the map or only capital letters of POIs under POI search. Mapsource shows first letter as capital and shows rest as lower-case. The OSM data is in unicode (i.e. bbox=21.32,41.92,21.55,42.08 where you can find lots of it). I also use slightly modified row04.trans (attached) to add transliteration to some extended cyrilic letters (i.e. U+0403 = Gj) plus slight change to transliteration standard (i.e. U+0425 = 'H' istead of 'Kh'), which unfortunately depends on each country... I guess it is possible to modify row tables to transiterate everything to captial ASCII, however i was hoping it is more systematic that everithing is captilized after transliteration pass if --lower-case option is not given. Thanks, Ivan

Hi Steve, I used that command line but still got upper case letters in the file.

I tries with your row04.trans too.

So I am puzzled at the moment as to what is happening.

I downloaded fresh r1067. Unpacked it new directory, cleaned up mapsource registry, etc.. added attached files (reg, cmd file renamed to .txt and xml file to fetch OSM data) and I still get lower case letters in resulting .img file Java interpreter is Sun JRE, running on XP sp2: java version "1.6.0_14" Java(TM) SE Runtime Environment (build 1.6.0_14-b08) Java HotSpot(TM) Client VM (build 14.0-b16, mixed mode, sharing) There are actually two maps I generate, one is with --code-page=1251 and is with all uppercase letters and the other one without CP specified is with lowercase letters.

...

plus slight change to transliteration standard (i.e. U+0425 = 'H' istead of 'Kh'), which unfortunately depends on each country...

Any thoughts about the best way to build this into mkgmap?

I would suggest style-file override lines for each letter (i.e. u+0425 = 'H' ) as there are few letters that need to be changed to adapt to local 'flavor' of transliteration/romanzation/etc... You should than apply style changes after reading base row tables. Kind regards, Ivan