Re: MARC::Charset problem

On Jun 22, 2006, at 5:34 AM, [email protected] wrote:
> I'm using MARC::Charset::marc8_to_utf8() v0.95 to transcode some
> Library of Congress data to utf8, however I'm finding a problem with
> character 'ø' (hex 0xB2 - lowercase scandinavian o / latin small
> letter o with stroke), this character is transcoding to 0xF8 -  
> which is
> not valid utf8 - when it should transcode to 0xC3B8. (According to the
> documentation, 0xF8 seems to be the ucs transcoding of this  
> character).
>
> Is this a bug in MARC::Charset or am I missing something?

Well I tried this out in the debugger with perl 5.8.7 and  
MARC::Charset v0.95:

--

   main::(-e:1):   1
     DB<1> use MARC::Charset qw(marc8_to_utf8)

     DB<2> $utf8 = marc8_to_utf8(chr(0xB2));

     DB<3> print "works" if $utf8 eq chr(0xF8);
   works

--

So it appears to be working fine. Perhaps when you are writing out  
your data you aren't preparing the filehandle for utf8? Can you  
provide a simple test script that demonstrates the problem so others  
can try to replicate?

//Ed


>
> Thanks,
>
> Michael
>
>