bug#53896: 28.0.91; Icomplete Vertical Mode fails to preview "C-x 8 RET 8072"

[Eli Zaretskii]

> Date: Fri, 11 Feb 2022 23:11:13 +0000 (UTC)
> From: Van Ly <van.ly@sdf.org>
> cc: Eli Zaretskii <eliz@gnu.org>, 53896@debbugs.gnu.org
> 
> > The CJK UNIFIED IDEOGRAPH-XXXX names are utterly meaningless to any
> > speaker of an East Asian language that uses them.
> 
> LINEAR A SIGN A807 ; from C-x 8 RET 807
> 
> The above is utterly meaningless to me but maybe it is meaningful to 
> the community studying Linear A.

Yes.  More importantly, there are 341 of them, as opposed to almost
27,000 of CJK Ideographs.  Even UnicodeData.txt doesn't include them,
but gives only the first and the last codepoints (one more evidence
that they don't have meaningful names in Unicode DB).

> My use case sources the CJK to read from, for example
> 
> => https://ctext.org/er-ya/shi-gu
> => https://humanum.arts.cuhk.edu.hk/Lexis/lexi-mf/search.php

Sorry, I don't understand the relevance.  Do you see any names in
those URLs, and if so, are those names in any way derived from the
Unicode character DB?  I just see characters, and thus typing their
codepoints is the way to go.

The characters _are_ supported, and we do assign names to them (which
are trivially derived from their codepoints), so typing their names is
just an unnecessarily complicated way of typing their codepoint.  IOW,
why not just type 8072 at "C-x 8 RET"s prompt, instead of selecting
"CJK IDEOGRAPH-8072" from a list of 27,000 candidates?

> ; from C-x 8 RET a
> 
> The current name for "<control>" had the old name "LINE FEED (LF)".

I don't think I see the relevance.  The 'old-name' property comes from
the Unicode DB, and we include the old names in the completion
candidates of "C-x 8 RET".  But the characters that have a meaningful
old name have none of the problems the CJK Ideographs have.