shp2pgsql fail load localized character

[aperi2007]

using the experimental version for windows:

POSTGIS="2.0.0SVN" GEOS="3.3.0-CAPI-1.7.0" PROJ="Rel. 4.6.1, 21 August 2008" LIBXML="2.7.6" USE_STATS

the shp2pgsql.exe don't load this shapefile where there is this local characters: "àèìòù".

I try with this settings: SRID:3003 Encoding "UTF-8"

I verify that removing the localized character the shapefile is correctly loaded.

Even on Linux debian version of Linux this shapefile with localized character isn't load.

I attach to this ticket the shapefile.

[aperi2007]

sample shapefile don't load from shp2pgsql

[strk]

Did you try passing "-W Latin1" ?

[aperi2007]

line shapefile version UTF8

[aperi2007]

hi strk, using Latin1 it work !

But I guess there is something other problem .

Because effectively using the "Latin1" the original shapefile is loaded. But I notice another strange phenomenon.

I have try to convert the line.shp to UTF8 version (using qgis) call it "line_utf8.shp" (add it to this ticket).

I notice that using encoding "UTF8" shp2pgsql load only the "line_utf8.shp" and dont load the "line.shp" reporting no error. I guess an error would be reported. The sql file generate from shp2pgsql is simply stop to the first line of COPY.

The second question is: Is right that shp2pgsql skip the load of a shapefile if encoding is set to UTF8 and the shapefile use a different encoding. Instead if the encoding is set to Latin1 it seem to be always loaded. Latin1 seem to be a special encoding setting.

[strk]

While UTF8 is a possibly multi-byte encoding, Latin1 (like others) are single byte. For single-byte encodings there's no way to tell if the encoding is correct, as each byte correspond to a character, so at worst you get wrong character. For multi-byte, there can be _wrong_ sequences.