Details
Bug
Description
The cause of this problem is utf8_encode() function in lib/html2text.php.(since 12th/Jun/2009)
function _convert()
{
// Variables used for building the link list
$this->_link_count = 0;
$this->_link_list = '';
$text = trim(stripslashes($this->html));
// Convert <PRE>
$this->_convert_pre($text);
// Run our defined search-and-replace
$text = preg_replace($this->search, $this->replace, $text);
$text = preg_replace_callback($this->callback_search, array(&$this, '_preg_callback'), $text);
// Replace known html entities
// $text = utf8_encode(html_entity_decode($text)); // Here!!!
$text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
// Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
$text = preg_replace('/&[^&;]+;/i', '', $text);
utf8_encode() function converts characters from ISO-8859-1 to UTF-8.
Then, Multibyte UTF-8 characters were convereted to wrong unexpected characters...
Issue Links
| This issue is duplicated by: | ||||
| MDL-2794 | html2text not compatible with utf-8 |
|
|
|
| MDLSITE-764 | Emails for forums subscribers deliver in broken encoding |
|
|
|
| This issue has been marked as being related by: | ||||
| MDL-19266 | Forum posts containing links not sent |
|
|
|
Activity
- All
- Comments
- History
- Activity
- Source
- Test Sessions
Moved to Moodle normal bugs (from site ones).
Assigned to Francois Marier, as far as I think this is related with the new html2text library recently introduced both in Moodle 1.9 and 2.0.
IMO, html_entity_decode() is 100% evil and that's the cause we have into our nice textlib library this function (that was the one used by old html2text library):
function entities_to_utf8($str, $htmlent=true)
it's able to convert both numerical and html entities to utf-8 as has been working fine since now, so something like:
$tl=textlib_get_instance();
$text = $tl->entities_to_utf8($text,true);
Should do the trick, without needing further processing out from the function (see weblib!). It will convert all entities.
Ciao 
Note that also there are some tests related to this:
function test_format_text_email()
that should pass to check that everything is in place, ciao
Hi Dmitry,
I guess moodle.org will be updated tomorrow (weekly build). So in 24h from now it should be working ok.
Thanks and ciao
I had posted the same comment for CVS.