On Thu, Jan 17, 2008 at 09:28:54AM -0800, Dave Hansen wrote: > Well, I've done virtually the entire US's TIGER data with the script, > with no issues, but it finally choked on Puerto Rico. > > It gets this: > > not well-formed (invalid token) at line 330, column 38, byte 14569 > at /usr/local/lib/perl/5.8.8/XML/Parser.pm line 187 > > when running on this file: > > http://dev.openstreetmap.org/~daveh/tiger.files/counties/PR/Adjuntas.osm > > I think it's the crazy characters in tags like this: > > <tag k="name" v="Carr Sillo de Calder�n"/> > <tag k="tiger:name_base" v="Carr Sillo de Calder�n"/> > > Being a stupid American, I have no real knowledge of character sets and > that fun. Any idea what the right way to fix this is?
As others have pointed out, the Calderón bit is ISO-8859-15 while the XML file says it's UTF-8. A simple iconv run should make it work: $ iconv -f ISO-8859-15 -t UTF-8 ~/Adjuntas.osm | ./04to05.pl Gabriel. _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/dev
