Re: Encoding/decoding problems

[Jambunathan K <kjambunathan@gmail.com>]

Deniz Dogan <deniz@dogan.se> writes:

> On 2011-07-28 11:01, Eli Zaretskii wrote:
>>> Date: Thu, 28 Jul 2011 10:18:04 +0200
>>> From: Deniz Dogan<deniz@dogan.se>
>>>
>>> I'm fetching an XML document that's uses iso-8859-1 coding with
>>> `url-retrieve' and then I parse it using `xml-parse-region'.
>>>
>>> After that, I get the parts of the document that I want and insert them
>>> into a buffer.  However, the Swedish characters å, ä and ö are displayed
>>> as \345, \344 and \326 respectively.
>>>
>>> I've tried messing around with `encode-coding-region' and
>>> `decode-coding-region' but I'm really not sure what to do here.
>>
>> I suggest to start with describing a reproducible recipe for this
>> problem.  Not sure if this forum is appropriate, perhaps emacs-devel
>> is a better place (as it sounds like you are describing a bug).
>>
>
> Here is the code to reproduce it:
>
> (defun fetch-and-show ()
>   (interactive)
>   (let* ((old-buffer (current-buffer))
>          (url "http://dogan.se/sites/default/files/example.xml")
>          (buffer (url-retrieve-synchronously url)))
>     (with-current-buffer buffer
>       (let ((doc (car (xml-parse-region (point-min) (point-max)))))
>         (with-current-buffer old-buffer
>           (insert
>            (nth 2 (nth 2 (nth 3 doc)))))))))
>
> The XML file is encoded in iso-8859-1 with a bunch of Swedish
> characters here and there.  The buffer I'm testing this with is
> *scratch* with utf-8-unix.  It should insert "hallå" but inserts
> "hall\345".


FWIW, this works as expected. Note the use of decode coding string.

(defun fetch-and-show ()
  (interactive)
  (let* ((old-buffer (current-buffer))
         (url "http://dogan.se/sites/default/files/example.xml")
         (buffer (url-retrieve-synchronously url)))
    (with-current-buffer buffer
      (let ((doc (car (xml-parse-region (point-min) (point-max)))))
        (with-current-buffer old-buffer
	  (insert
	   (decode-coding-string (nth 2 (nth 2 (nth 3 doc))) 'iso-8859-1)))))))

>
> I have no idea whether I should use `encode-region-string' or
> decode-region-string' or what.  I'd doubt it's a bug to be honest,
> it's probably my lack of understanding that's causing this.
>
> Deniz
>
>

--