Skip to main content.
home | support | download

Back to List Archive

Re: non ISO-8859-1 headers

From: Peter Karman <karman(at)>
Date: Fri Feb 20 2004 - 22:33:43 GMT
Forgive me if I am misunderstanding. This sounds like a thread that went 
by in December 2003. Search the discussion archives. I believe the end 
result was that because swish-e currently does not save data in UTF-8, 
it can't display any of the indexed data in that format. If by "display 
the output" you mean the contents of StoreDescription from the swish 
index, then that currently can't happen. I don't think the issue is with 
the perl locale or encodings settings for the CGI scripts, I think it's 
with the data in the index, *as it was indexed*.

But then, I'm just learning all about encodings, and I may have 
misunderstood this thread the last time it went around.


Tim Freedom supposedly wrote on 2/20/04 1:51 PM:

> --- Bill Moseley <> wrote:
>>No there's not config for that.  For one thing Swish-e indexes only
>>8859-1 (when using libxml2 parser) so I didn't see any reason for
>>providing a config option for that.  If you don't want to hack the
>>source read the instructions in swish.cgi on how to copy the template
>>module to a new one, then edit and use that one instead.
> This is not really indexing related - this is more related to being
> able to display multiply encoded files as part of your results.  For
> instance, I have lots of files that have both English and Arabic in
> them (UTF-8), currently I can only index the english parts (again,
> I'm willing to help with adding UTF-8 abilities :-), yet when I display
> the output it would be nice to default to UTF-8 to see both texts.

Peter Karman - Software Publications Engineer - Cray Inc
phone: 651-605-9009 -
Received on Fri Feb 20 14:33:44 2004