Bug#365029: binmode not good enough

[Carl Johnstone]

reopen 365029 !

Hi,

I've run ran into this problem too.

Unfortunately setting binary=>1 isn't enough. Although the module will now 
seem to parse the file without errors, it doesn't correctly deal with the 
character set of the input and treats it only as a bunch of bytes. This is 
even the case where you have explicitly (correctly) set the character set of 
the CSV file.

This is fine in situations where your code is doing something trivial, but 
needs workarounds in other cases. For example reading in a CSV file in 
ISO-8859-1, and outputting UTF-8 XML.

It would seem that the XS parts of the module need updating in line with 
Perl's Unicode support. I note that somebody has taken ownership of the 
module upstream with the intent of fixing this, and has already started to 
release updated versions of the module.

Carl