Hello Latest versions of mkgmap stopped working correctly with windows-1257 encoding (I no longer see Lithuanian letters on Colorado and QLandkarteGT). Is this related to the fact that sort/cp1257.txt is not provided yet? I'm NOT using --index option yet but using --lower-case and --charset=windows-1257. Thank you P.S. r1877 -- Tomas Straupis
On 03/03/11 21:33, Tomas Straupis wrote:
Hello
Latest versions of mkgmap stopped working correctly with windows-1257 encoding (I no longer see Lithuanian letters on Colorado and QLandkarteGT).
looks like i'm having the same problem (with latvian characters) - but as i only recently started using mkgmap, i haven't had a version showing those characters successfully ;) is this something that can be worked around using different options or not ? which is the latest version that such characters are expected to work ?
Is this related to the fact that sort/cp1257.txt is not provided yet?
I'm NOT using --index option yet but using --lower-case and --charset=windows-1257.
Thank you
P.S. r1877 -- Rich
HI
is this something that can be worked around using different options or not ? which is the latest version that such characters are expected to work ?
I expect the latest version to work, but if it doesn't...
Is this related to the fact that sort/cp1257.txt is not provided yet?
I'm NOT using --index option yet but using --lower-case and --charset=windows-1257.
I would try --code-page=1257 instead of --charset A quick check seems to show that --code-page still works, but if it doesn't then since it fails with qlandkartegt and that is open-source we should be able to work out what is going wrong. I would not expect that the lack of a sort/cp1257.txt file would affect anything that worked before, since there has never been any special support for sorting names written with 1257 characters. ..Steve
On 03/07/11 11:08, Steve Ratcliffe wrote:
HI
is this something that can be worked around using different options or not ? which is the latest version that such characters are expected to work ?
I expect the latest version to work, but if it doesn't...
Is this related to the fact that sort/cp1257.txt is not provided yet?
I'm NOT using --index option yet but using --lower-case and --charset=windows-1257.
I would try --code-page=1257 instead of --charset
so what is the difference between the two in the case of mkgmap ? :) --help=options not-too-helpfully says about "--code-page" : "TODO: explain what this does, and why one would or would not want to do it."
A quick check seems to show that --code-page still works, but if it doesn't then since it fails with qlandkartegt and that is open-source we should be able to work out what is going wrong.
I would not expect that the lack of a sort/cp1257.txt file would affect anything that worked before, since there has never been any special support for sorting names written with 1257 characters.
..Steve -- Rich
Hi
--help=options not-too-helpfully says about "--code-page" :
"TODO: explain what this does, and why one would or would not want to do it."
Yes, that is not very useful... There are two options concerned with character sets which were added before I really understood the possibilities and so they were to allow for experimentation. The code-page option is the one to use now. It is a number 1252, 1257 etc. The charset option is currently not useful, it should probably be removed and/or renamed. If given it is a string. I'll update the documentation to reflect this. ..Steve
2011-03-07 Steve Ratcliffe:
is this something that can be worked around using different options or not ? which is the latest version that such characters are expected to work ? I expect the latest version to work, but if it doesn't...
charsets stopped working (at least for cp1257) after index branch merge. It used to work for at least half a year before that (on stable versions). -- Tomas Straupis
Why cant we make all Names UTF-8 ? -- Wikipedia -- http://tools.wikimedia.de/~flacus/IWLC/ OSM -- http://osm.flacus.de/
On 03/07/11 12:06, flabot@googlemail.com wrote:
Why cant we make all Names UTF-8 ?
i tried that and only got garbled characters on garmin. either i did not do it correctly or they are a bit backwards with their understanding of international market :) i'd be glad to perform specific tests if needed. -- Rich
I expect the latest version to work, but if it doesn't...
charsets stopped working (at least for cp1257) after index branch merge. It used to work for at least half a year before that (on stable versions).
-- Tomas Straupis _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Is that using code-page or charset? I would not expect --charset to work.
On 03/07/11 19:49, Tomas Straupis wrote:
2011-03-07 Steve Ratcliffe:
Is that using code-page or charset?
I would not expect --charset to work.
Yep. Changing to --code-page=1257 fixes Lithuanian letters!
hmm. i might be doing something wrong... but --code-page=1257 with mkgmap 1867 still results in garbled latvian characters. in which revision/version does it work for you ? what else could i be doing incorrectly ? full options to the jar are : --mapname=13131313 --description=Latvia-2011.03.06 --country-name=Latvia --country-abbr=LV --code-page=1257 --net --route latvia-2011.03.06.osm
Thank you! -- Rich
On Wed, Mar 09, 2011 at 01:45:02PM +0200, Rich wrote:
Yep. Changing to --code-page=1257 fixes Lithuanian letters!
hmm. i might be doing something wrong... but --code-page=1257 with mkgmap 1867 still results in garbled latvian characters.
Do you have an example? Can you check the file with a hex editor. For example, if you have a string Test*#% that shows up as Test&/!, search for "Test" in the img file and show the hexdump. Also, what encoding is the input in? UTF-8 like XML? What chars should they be? You might find the GNU Recode tool useful when diagnosing character set translation problems. For example, if your terminal emulator is configured to use the UTF-8 encoding, the following should encode your input properly, or complain if it cannot be done: echo 'your utf-8 string'|recode utf8..cp1257|od -t x1 Best regards, Marko
On 03/09/11 14:26, Marko Mäkelä wrote:
On Wed, Mar 09, 2011 at 01:45:02PM +0200, Rich wrote:
Yep. Changing to --code-page=1257 fixes Lithuanian letters!
hmm. i might be doing something wrong... but --code-page=1257 with mkgmap 1867 still results in garbled latvian characters.
Do you have an example? Can you check the file with a hex editor. For example, if you have a string Test*#% that shows up as Test&/!, search for "Test" in the img file and show the hexdump. Also, what encoding is
Ū -> DB Ā -> C2 Ņ -> D2 Ļ -> CF Ž -> DE
the input in? UTF-8 like XML? What chars should they be?
it's an osm xml from geofabrik downloads, so i'd expect it to be proper utf8 :)
You might find the GNU Recode tool useful when diagnosing character set translation problems. For example, if your terminal emulator is configured to use the UTF-8 encoding, the following should encode your input properly, or complain if it cannot be done:
echo 'your utf-8 string'|recode utf8..cp1257|od -t x1
echo 'Ž'|recode utf8..cp1257|od -t x1 0000000 de 0d 0a 0000003
so it might be that the file has properly encoded characters, but the device does not properly display them ? if i "manually" compare what they look like on the gps device and onscreen, result on gps device seems to match what i get if i look at cp1257 (or iso-8859-13) encoded text in a unicode terminal...
Best regards,
Marko -- Rich
On Wed, Mar 09, 2011 at 04:02:38PM +0200, Rich wrote:
so it might be that the file has properly encoded characters, but the device does not properly display them ?
That is a possibility. Do the older Garmins (such as GPSMap 60 and Edge 605/705) support anything else than the ISO 8859-1 like CP1252? I think that I saw some hacked Russian firmware somewhere, with some 8-bit Cyrillic encoding support, I suppose.
if i "manually" compare what they look like on the gps device and onscreen, result on gps device seems to match what i get if i look at cp1257 (or iso-8859-13) encoded text in a unicode terminal...
Do you mean UTF-8 interpreted as if it were cp1257 (multiple characters per non-ASCII character)? Anything non-UTF-8, non-ASCII interpreted as UTF-8 should show as � (U+FFFD). echo -e '\xdb\xc2\xd2\xcf\xde\r'|recode cp1257..utf8 ŪĀŅĻŽ echo -e '\xdb\xc2\xd2\xcf\xde\r'|recode cp1252..utf8 ÛÂÒÏÞ The former interprets the codes as CP1257 and the latter as CP1252. Did you see the latter on your GPS? Marko
On 03/09/11 17:00, Marko Mäkelä wrote:
On Wed, Mar 09, 2011 at 04:02:38PM +0200, Rich wrote:
so it might be that the file has properly encoded characters, but the device does not properly display them ?
That is a possibility. Do the older Garmins (such as GPSMap 60 and Edge 605/705) support anything else than the ISO 8859-1 like CP1252? I think that I saw some hacked Russian firmware somewhere, with some 8-bit Cyrillic encoding support, I suppose.
no idea about older ones, but this specimen is etrex vista hcx
if i "manually" compare what they look like on the gps device and onscreen, result on gps device seems to match what i get if i look at cp1257 (or iso-8859-13) encoded text in a unicode terminal...
Do you mean UTF-8 interpreted as if it were cp1257 (multiple characters per non-ASCII character)? Anything non-UTF-8, non-ASCII interpreted as UTF-8 should show as � (U+FFFD).
echo -e '\xdb\xc2\xd2\xcf\xde\r'|recode cp1257..utf8 ŪĀŅĻŽ echo -e '\xdb\xc2\xd2\xcf\xde\r'|recode cp1252..utf8 ÛÂÒÏÞ
The former interprets the codes as CP1257 and the latter as CP1252. Did you see the latter on your GPS?
the latter seems like what i'm seeing on the device with --codepage=cp-1257
Marko -- Rich
I've tried downloading Latvian extract from geofabrik and creating a map. I then browsed through Riga and haven't noticed any problems (using Colorado). I do see Latvian letters properly... P.S. Note that Lithuanian letters (and the same should apply to Latvian ones) does NOT work on GPSMAP 60csx, eTrex Vista HCx and it DOES work with Colorado, Oregon, MobileXT 5.00(60), Nuvi 250w. -- Tomas Straupis
On 03/09/11 18:31, Tomas Straupis wrote:
I've tried downloading Latvian extract from geofabrik and creating a map. I then browsed through Riga and haven't noticed any problems (using Colorado). I do see Latvian letters properly...
P.S. Note that Lithuanian letters (and the same should apply to Latvian ones) does NOT work on GPSMAP 60csx, eTrex Vista HCx and it
i... don't like vista hcx being mentioned in that list :) do you have any source for this info ?
DOES work with Colorado, Oregon, MobileXT 5.00(60), Nuvi 250w. -- Rich
2011-03-09 Rich:
i... don't like vista hcx being mentioned in that list :) do you have any source for this info ?
I've placed a map on the internet and asked people to send me an e-mail telling if they see Lithuanian letters or not. That is the "source" of above-mentioned information... -- Tomas Straupis
Am 09.03.2011 16:10, schrieb Rich:
On 03/09/11 17:00, Marko Mäkelä wrote:
On Wed, Mar 09, 2011 at 04:02:38PM +0200, Rich wrote:
so it might be that the file has properly encoded characters, but the device does not properly display them ? That is a possibility. Do the older Garmins (such as GPSMap 60 and Edge 605/705) support anything else than the ISO 8859-1 like CP1252? I think that I saw some hacked Russian firmware somewhere, with some 8-bit Cyrillic encoding support, I suppose. no idea about older ones, but this specimen is etrex vista hcx
Do you have a garmin map, which shows the characters as expected? This will prove that the device could display the chars at least.
On 03/09/11 22:58, Johann Gail wrote:
Am 09.03.2011 16:10, schrieb Rich:
On 03/09/11 17:00, Marko Mäkelä wrote:
On Wed, Mar 09, 2011 at 04:02:38PM +0200, Rich wrote:
so it might be that the file has properly encoded characters, but the device does not properly display them ? That is a possibility. Do the older Garmins (such as GPSMap 60 and Edge 605/705) support anything else than the ISO 8859-1 like CP1252? I think that I saw some hacked Russian firmware somewhere, with some 8-bit Cyrillic encoding support, I suppose. no idea about older ones, but this specimen is etrex vista hcx
Do you have a garmin map, which shows the characters as expected? This will prove that the device could display the chars at least.
nope, default garmin map was quite a joke. i submitted a question at garmin techsupport. -- Rich
participants (6)
-
flabot＠googlemail.com -
Johann Gail -
Marko Mäkelä -
Rich -
Steve Ratcliffe -
Tomas Straupis