bug#17865: 24.3.92; encode-coding-string returns 1

bug#17865: 24.3.92; encode-coding-string returns 1

[Leo Liu]

(with-temp-buffer
  (set-buffer-multibyte nil)
  (list (encode-coding-string "我" 'utf-8 nil (current-buffer))
        (buffer-string)))

In my emacs it inserts three bytes yet return just one. Bug?

Leo

bug#17865: 24.3.92; encode-coding-string returns 1

[Eli Zaretskii]

> From: Leo Liu <sdl.web@gmail.com>
> Date: Sat, 28 Jun 2014 01:30:54 +0800
> 
> 
> (with-temp-buffer
>   (set-buffer-multibyte nil)
>   (list (encode-coding-string "我" 'utf-8 nil (current-buffer))
>         (buffer-string)))
> 
> In my emacs it inserts three bytes yet return just one. Bug?

Not a bug: you inserted one character.

bug#17865: 24.3.92; encode-coding-string returns 1

[Leo Liu]

On 2014-06-27 23:15 +0300, Eli Zaretskii wrote:
> Not a bug: you inserted one character.

But without `(set-buffer-multibyte nil)' you get 3. Normal?

Leo

bug#17865: 24.3.92; encode-coding-string returns 1

[Eli Zaretskii]

> From:  Leo Liu <sdl.web@gmail.com>
> Cc: 17865@debbugs.gnu.org
> Date: Sat, 28 Jun 2014 07:56:53 +0800
> 
> On 2014-06-27 23:15 +0300, Eli Zaretskii wrote:
> > Not a bug: you inserted one character.
> 
> But without `(set-buffer-multibyte nil)' you get 3. Normal?

Yes, because then you insert 3 multibyte characters that represent 3
8-bit bytes.

The value you get counts characters produced by the encoding.  How to
count characters in a unibyte buffer is a question that has no easy
answers.

If, after all this, you will conclude that the return value is not
very useful for high-level programming, I will agree with you.  The
return value is set up to provide what some internal use cases need,
and that's it.  It just returns some member of the structure used to
convert the string; how that member is computed is an implementation
detail that rarely, if at all, is exposed to Lisp applications.

It could be that the return value of 1 in a unibyte buffer was never
bumped into before, and thus never scrutinized.  It could also be that
some Lisp application actually needs that value to be 1.

What is the real-life use case behind this?  Maybe if we talk in the
context of that use case, the discussion will be more meaningful.

bug#17865: 24.3.92; encode-coding-string returns 1

[Andreas Schwab]

Fixed in emacs-24.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."