[horde] IMP: Encoding problem in message view

[Deokgon Kim]

2016년 01월 12일 05:45에 Michael J Rubinsky 이(가) 쓴 글:
>
> Quoting 김 덕곤 <dgkim at dgkim.net>:
>
>> 2015. 12. 31. 오후 10:29에 Jan Schneider 이(가) 쓴 글:
>>>
>>> Zitat von 김 덕곤 <dgkim at dgkim.net>:
>>>
>>>> Hello.
>>>>
>>>> When I opening a message(that is not encoded in UTF-8),
>>>> character is not shown in correct encoding.
>>>>
>>>> But, when I reply or forward it, correct character is shown.
>>>> (And, the replied message is encoded in UTF-8)
>>>>
>>>> When I open a message as popup(Show all parts, and click on a part),
>>>> character is also in a wrong encoding,
>>>> but if I change character encoding to Unicode in a Browser option,
>>>> correct character is shown.
>>>> This means server already converted form encoding to encoding,
>>>> but wrong Content-Type is sent.
>>>>
>>>> A message's encoding is 'ks_c_5601-1987'
>>>> Web page's encoding is 'UTF-8'.
>>>>
>>>> Any suggestions?
>>>>
>>>> Versions.
>>>> Horde : 5.2.8
>>>> IMP : 6.2.11
>>>>
>>>> -- 
>>>> Horde mailing list
>>>> Frequently Asked Questions: http://horde.org/faq/
>>>> To unsubscribe, mail: horde-unsubscribe at lists.horde.org
>>>
>>> https://bugs.horde.org/ticket/14149
>>>
>> I think I figured out why this happens.
>>
>> In Domhtml, at the line of 87, there is call to "convertCharset". as 
>> follows.
>>
>>  86             $this->_loadHTML(
>>  87                 $this->_xmlencoding . 
>> Horde_String::convertCharset($text, $c    harset, 'UTF-8')
>>  88             );
>>
>> Here is the call-stack dumped from Horde_Domhtml::__construct().
>>
>> __construct: Called from /usr/share/php/Horde/Text/Filter/Xss.php:72
>> __construct   1. Horde_Text_Filter_Xss->postProcess() 
>> /usr/share/php/Horde/Text/Filter.php:108
>> __construct   2. Horde_Text_Filter->filter() 
>> /usr/share/php/Horde/Core/Factory/TextFilter.php:72
>> __construct   3. Horde_Core_Factory_TextFilter->filter() -:-
>> __construct   4. call_user_func() 
>> /usr/share/php/Horde/Mime/Viewer/Base.php:438
>> __construct   5. Horde_Mime_Viewer_Base->_textFilter() 
>> /usr/share/php/Horde/Mime/Viewer/Html.php:164
>> __construct   6. Horde_Mime_Viewer_Html->_cleanHTML() 
>> /var/www/html/horde/imp/lib/Mime/Viewer/Html.php:179
>> __construct   7. IMP_Mime_Viewer_Html->_IMPrender() 
>> /var/www/html/horde/imp/lib/Mime/Viewer/Html.php:75
>> __construct   8. IMP_Mime_Viewer_Html->_renderInline() 
>> /usr/share/php/Horde/Mime/Viewer/Base.php:156
>> __construct   9. Horde_Mime_Viewer_Base->render() 
>> /var/www/html/horde/imp/lib/Contents.php:654
>> __construct  10. IMP_Contents->renderMIMEPart() 
>> /var/www/html/horde/imp/lib/Mime/Viewer/Alternative.php:128
>> __construct  11. IMP_Mime_Viewer_Alternative->_IMPrender() 
>> /var/www/html/horde/imp/lib/Mime/Viewer/Alternative.php:54
>> __construct  12. IMP_Mime_Viewer_Alternative->_renderInline() 
>> /usr/share/php/Horde/Mime/Viewer/Base.php:156
>> __construct  13. Horde_Mime_Viewer_Base->render() 
>> /var/www/html/horde/imp/lib/Contents.php:654
>> __construct  14. IMP_Contents->renderMIMEPart() 
>> /var/www/html/horde/imp/lib/Contents.php:1462
>> __construct  15. IMP_Contents->getInlineOutput() 
>> /var/www/html/horde/imp/lib/Ajax/Application/ShowMessage.php:296
>> __construct  16. IMP_Ajax_Application_ShowMessage->showMessage() 
>> /var/www/html/horde/imp/lib/Dynamic/Message.php:64
>> __construct  17. IMP_Dynamic_Message->_init() 
>> /var/www/html/horde/imp/lib/Dynamic/Base.php:90
>> __construct  18. IMP_Dynamic_Base->__construct() 
>> /var/www/html/horde/imp/dynamic.php:33
>>
>> I tried to find where the 'convertCharset' occurs.
>>
>> I found in 'Horde/Mime/Viewer/Html.php'. (line 106 to 114)
>>
>> 106     protected function _renderInline()
>> 107     {
>> 108         $html = 
>> $this->_cleanHTML($this->_mimepart->getContents(), array('in    line' 
>> => true));
>> 109
>> 110         return $this->_renderReturn(
>> 111 Horde_String::convertCharset($html->returnHtml(), 
>> $this->_mimepa    rt->getCharset(), 'UTF-8'),
>> 112             'text/html; charset=UTF-8'
>> 113         );
>> 114     }
>>
>> So I think, 'convertCharset' is unintentionally called twice.
>
> Horde_Mime_Viewer::_renderInline() isn't (shouldn't) be called in this 
> context, since we are in IMP_Mime_Viewer_Html, which overrides the 
> parent's copy of the _renderInLine() method.
>
>
>

Hello.
Today I tried to figure out a little more.

I do understand Horde_Mime_Viewer::_renderinline() isn't called at this 
context.

And I still have the problem.

When I change code not to 'convertCharset', It works. 
(/usr/share/php/Horde/Domhtml.php)

  83             /* Convert/try with UTF-8 first. */
  84             $this->_origCharset = Horde_String::lower($charset);
  85             $this->_xmlencoding = '<?xml encoding="UTF-8"?>';
  86             $this->_loadHTML(
  87                 $this->_xmlencoding . 
Horde_String::convertCharset($text, $charset, 'UTF-8')
  88             );

At the line of 87.

Why, in the first place, converting to UTF-8?

And
what if body part of html($text) doesn't contains fully structured HTML,
Can Domhtml.php handle it?
(I tested DOMDocument object, if the $text doesn't contains [meta 
charset], 'dom->encoding' is not working.)

And what is the state of $text?
Does Content-Transfer-Encoding is processed?

To confirm, this problem is all related to Domhtml.php,
what can be the status of member variables?
Or is this problem related to php DOMDocument?