Re: [Unicode-2] `C-h f' error

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

In article <E1Iv1mx-0003q8-0I@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:

>     Thank you for investigating this problem.  But, as I don't
>     have a time to work on it at the moment, I fixed lread.c so
>     that it works as previously.

> What is the lread.c behavior that you changed?

Make the Lisp reader to generate a symbol of unibyte name
when it is read from a unibyte buffer.  Previously, the
multibyteness of a symbol name is determined by the byte
sequence (by using make_string).

> What is the new behavior?

The following phenomenon was reported.

Katsumi Yamaoka <yamaoka@jpl.org> writes:
> Isn't it a side effect of this change?  The `C-h f' command causes
> the following error, though it can be solved by reloading "help".

> Debugger entered--Lisp error: (setting-constant :validate)
>   function-called-at-point()
>   [...]
>   call-interactively(describe-function)

> It seems that the `with-syntax-table' macro, that
> `function-called-at-point' uses, was not expanded properly when
> dumping Emacs:

> (disassemble 'function-called-at-point)
> 0	constant  syntax-table
> 1	call	  0
> 2	current-buffer 
> 3	varbind	  :validate
> 4	varbind	  setup-function
> 5	constant  (<byte code>...)
>    0	   save-current-buffer 
>    1	   varref    :validate
>    2	   set-buffer 

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