bug#24603: [RFC 02/18] Generate upcase and downcase tables from Unicode data

[Eli Zaretskii]

> From: Michal Nazarewicz <mina86@mina86.com>
> Date: Tue,  4 Oct 2016 03:10:25 +0200
> 
> +  ;; Set all Letter, uppercase; Letter, lowercase and Letter, titlecase 
> syntax
> +  ;; to word.  FIXME: Should this also be done for Letter, modifier and 
> Letter,
> +  ;; other?  What about other alphabetic characters?
> +  (let ((syn-tab (standard-syntax-table)))
> +    (map-char-table
> +     (lambda (ch cat)
> +       (when (memq cat '(Lu Ll Lt))
> +         (modify-syntax-entry ch "w   " syn-tab)))
> +     (unicode-property-table-internal 'general-category)))

The answer to these questions is "as required by backward
compatibility", i.e. compare with the manual setup we had until now.

If that criterion doesn't provide the full answer, I would go by
Unicode guidance, i.e. support all the case conversions specified in
the Unicode character database (UCD).

Thanks.