Subject: revised describe function now on cvs branch
From: Robert Dodier
Date: Sat, 25 Nov 2006 10:34:56 -0700
Hi Vadim,
> Great! Now it works almost fine except the length problem
> you mentioning. But I worry that enforcing :utf8 for open
> is not universal solution. It is OK with any utf8 locale and with
> English but what about other languages in 8-bit encodings?
> Maybe pt and es in iso-8859-1 are OK too since they have
> small number of non-ASCII character codes.
You're right, it is not quite correct to process iso-8859-1 as utf-8.
I see that the documentation seems to be displayed correctly
with LANG=es_ES or pt_BR, but I don't know if we can be assured
that will always work correctly.
> As far as I know Perl's open accept any encoding
> so maybe it is good idea to pass encoding
> to build_index script as extra parameter?
Yes, that makes sense to me.
All the best, & thanks for your help,
Robert