Re: 23.0.60; describe-char gives wrong information

[Kenichi Handa <handa@ni.aist.go.jp>]

In article <60200667-2BCD-4D24-BC0F-700361E632A2@Freenet.DE>, Peter Dyballa <Peter_Dyballa@Freenet.DE> writes:

>>> Arial Unicode has U+1F48. It does not have it in a gb18030.2000-0
>>> font encoding, because this code point is not defined in
>>> GB18030-2000. So one of the first mistakes is to assume U+1F48 is
>>> defined in GB18030-2000
> >
> > The charset GB18030-2000 surely contains U+1F48.  Actually
> > it contains all Unicode characters.
> >
>>> and another one is to use a partial font
>>> encoding like gb18030.2000-0
> >
> > What do you mean by "partial font encoding"?  Anyway, as I
> > wrote before, the bug of selecting a font that doesn't have
> > the character should be fixed now.

> Can be I misunderstood the standard and saw gaps in it when there  
> none.

A font labelled as GB18030.2000 doesn't necessarily contains
all characters of the charset GB18030.2000.

> Anyway: fact is that a few programmes show that arial unicode  
> ms has U+1F48 and GNU Emacs 23.0.60 does not display it from the  
> font's gb18030.2000-0 encoding. I am attaching two screenshots from  
> xfd. The gb18030.2000-0 encoding variant starts far behind Greek at U 
> +8140, which I understand as: this encoding does not provide glyphs  
> outside some Chinese block(s).

You are still confusing with the encoding and the repertory.
"GB18030.2000-0" is a part of a XLFD fontname showing the
encoding of the font, and it doesn't mean the repertory
(i.e. which characters are included in that font).  That is
the same as any "ISO10646-1" fonts.  AFAIK, none of them
contain all of Unicode BMP characters.

Anyway, have you tried the latest code?  Does it still try
to display U+1F48 using the font "...-gb18030.2000-0"?

---
Kenichi Handa
handa@ni.aist.go.jp