Hi,

> From: TAKAHASHI Masayoshi [mailto:maki / inac.co.jp]
> Sent: Tuesday, July 17, 2001 1:22 PM

> "NAKAMURA, Hiroshi" <nahi / keynauts.com> wrote:
> > Regards to NQXML:  Though it is very interesting idea to
> > add it to std lib in point of pure Ruby library, I think
> > we should wait ruby/1.8 with m17n and NQXML's CES support.
> 
> Why should we wait m17n-ruby? ruby-1.6.x has already supported
> utf-8 string :-)

See SOAP4R implementation if you are interested.  UTF-8 has
already been supported.

Hmm. "UTF-8 only" might not be a problem for some system.
But my intention is "XML as a glue" so XML processor should
not restrict CES usage of XML application.

$KCODE = "UTF8" with ruby/1.6 + NQXML works well but I want
to use EUC/SJIS/iso-8859-1 in my CGI, desktop tool, and so on.

> According to NQXML's CES support:
>  * at least, should be able to parse UTF-8 documents.

Under $KCODE="UTF8" condition.

>  * But some restriction of character(ex. NameChar)
>    may not be needed.
>    (Do you know XML Blueberry and discussion on XML-dev?
>     cf. http://www.xml.com/pub/a/2001/07/11/sunshine.html )
>    So it's not difficult to support (limited) UTF-8, isn't it?

No and I don't know, sorry.

> By the way, I think NQXML is slow for large documents. Is it
> serious problem?

Can you hack it more deep?

// NaHi