Bug #49483 for Geo-Coder-GoogleMaps: Encoding problems with non-latin1 data and json transport

Mon Sep 07 06:38:51 2009 SREZIC [...] cpan.org - Ticket created

Subject:

Encoding problems with non-latin1 data and json transport

With the following script: #!/usr/bin/perl use strict; use Geo::Coder::GoogleMaps; my $apikey = ...; my $gmap = Geo::Coder::GoogleMaps->new(apikey => $apikey); my $location = $gmap->geocode(location => 'Djakovo, Croatia'); print $location->{data}{address}, "\n"; __END__ I get the result "\u0110akovo, Croatia". It seems that the first character (D with a stroke) was not decoded properly. I don't know if JSON::Syck does support javascript escapes at all, or if you better switch to JSON::XS instead? Also, if I try to use "Basks Voda, Croatia" as input, then the output is (using Data::Dumper notation): "Ba\x{9a}ka Voda, Croatia". This also does not right, as \x9a is not a valid (Unicode) character. It looks rather like a Windows-1252 fallback (for the letter "S with caron"). This is probably the fault of Google, but it would be nice if this could also be workarounded. Regards, Slaven

Mon Sep 07 07:05:12 2009 SREZIC [...] cpan.org - Correspondence added

On Mon Sep 07 03:38:51 2009, SREZIC wrote: Show quoted text

> With the following script: > > #!/usr/bin/perl > > use strict; > use Geo::Coder::GoogleMaps; > > my $apikey = ...; > my $gmap = Geo::Coder::GoogleMaps->new(apikey => $apikey); > my $location = $gmap->geocode(location => 'Djakovo, Croatia'); > print $location->{data}{address}, "\n"; > > __END__ > > I get the result "\u0110akovo, Croatia". It seems that the first > character (D with a stroke) was not decoded properly. I don't know if > JSON::Syck does support javascript escapes at all, or if you better > switch to JSON::XS instead? > > Also, if I try to use "Basks Voda, Croatia" as input, then the output is > (using Data::Dumper notation): "Ba\x{9a}ka Voda, Croatia". This also > does not right, as \x9a is not a valid (Unicode) character. It looks > rather like a Windows-1252 fallback (for the letter "S with caron"). > This is probably the fault of Google, but it would be nice if this could > also be workarounded.

I just stumbled over this ticket: https://rt.cpan.org/Ticket/Display.html?id=35173 describing a similar problem in Geo::Coder::Google, and the proposed solution there (setting a custom User-Agent string) also works here. So my current workaround is: my $gmap = Geo::Coder::GoogleMaps->new(apikey => $apikey, ua => LWP::UserAgent->new(agent =>"Mozilla/5.0 (compatible; Geo::Coder::GoogleMaps/$Geo::Coder::GoogleMaps::VERSION; Google, please stop smoking crack; http://rt.cpan.org/Public/Bug/Display.html?id=35173)") ); Regards, Slaven Regards, Slaven

Mon Sep 07 07:05:20 2009 SREZIC [...] cpan.org - Status changed from 'new' to 'open'

Mon Sep 07 08:31:09 2009 SREZIC [...] cpan.org - Correspondence added

On Mon Sep 07 06:38:51 2009, SREZIC wrote: [...] Show quoted text

> Also, if I try to use "Basks Voda, Croatia" as input, then the output is

Typo, this should be "Baska Voda, Croatia". Regards, Slaven

Sat Dec 05 17:53:12 2009 dupuisarn [...] cpan.org - Correspondence added

By the time I got through this ticket you give both the bug report and the solution :-) Thank you I am adding a default custom user agent right now. Regards, Arnaud Dupuis

Fri Dec 11 03:59:11 2009 dupuisarn [...] cpan.org - Correspondence added

I fixed the issue using the trick you provided, and now the result seems just fine : <?xml version="1.0" encoding="UTF-8" standalone="yes"?> <kml xmlns="http://earth.google.com/kml/2.1"> <Placemark id="p1"> <AddressDetails xmlns="urn:oasis:names:tc:ciq:xsdschema:xAL:2.0"> <Country> <CountryName>Hrvatska</CountryName> <AdministrativeArea> <AdministrativeAreaName>Osječko-baranjska</AdministrativeAreaName> <Locality/> </AdministrativeArea> <CountryNameCode>HR</CountryNameCode> </Country> </AddressDetails> <address>Đakovo, Croatia</address> <ExtendedData> <LatLonBox north="45.3707943" south="45.2452366" east="18.5360083" west="18.2798895"/> </ExtendedData> <Point> <coordinates>18.4079489,45.3080502,0</coordinates> </Point> </Placemark> </kml> The fix is not yet available on CPAN but it's available on the SVN at : https://code.google.com/p/geo-coder-googlemaps/source/browse/ This code is not yet ready for release but its fixing the issue already.

Fri Dec 11 13:17:28 2009 dupuisarn [...] cpan.org - Correspondence added

done in version 0.3.

Fri Dec 11 13:17:39 2009 dupuisarn [...] cpan.org - Correspondence added

done in version 0.3.

Fri Dec 11 13:17:40 2009 dupuisarn [...] cpan.org - Status changed from 'open' to 'resolved'