Trunk throws warnings on php4 sites

[Denis-de-Bernardy]

On my customer's site, it was due to the RSS widgets grabbing eBay feeds. But there are other calls to html_entity_decode() around.

Symptom:

Trying to decode HTML entities into UTF-8 results in the following error message:

Warning: cannot yet handle MBCS in html_entity_decode()!

The line is repeated about 200 times, then html_entity_decode just uses ISO-8859-1 charset.

The related php bugs:

• ​http://bugs.php.net/bug.php?id=25670
• ​http://bugs.php.net/bug.php?id=27626

Also occurs in BB:

​http://bbpress.org/forums/topic/warning-cannot-yet-handle-mbcs-in-html_entity_decode

Based on this thread, there is a potential fix:

• ​http://pkp.sfu.ca/support/forum/viewtopic.php?f=3&t=2081
• ​http://pkp.sfu.ca/ocs_download (it's GPL)

PHP4 has reached EOL.

You may want to read: ​http://lesterchan.net/wordpress/2008/06/09/warning-cannot-yet-handle-mbcs-in-html_entity_decode/

yeah, I'm aware, but wp is still supporting it. plus, it might fix your plugin too. ;-)

another example...

Their implementation is in classes/core/String.inc.php, and goes like this:

function html2utf($str) {
	// convert named entities to numeric entities
	$str = strtr($str, String::getHTMLEntities());

	// use PCRE-aware replace function to replace numeric entities
	$str = String::regexp_replace('~&#x([0-9a-f]+);~ei', 'String::code2utf(hexdec("\\1"))', $str);
	$str = String::regexp_replace('~&#([0-9]+);~e', 'String::code2utf(\\1)', $str);

	return $str;
 }

Side note: best I'm aware, strtr is not mb safe, so the above code is potentially buggy in mb languages.

on a separate note, this class has a couple of useful functions

The code in the String.inc.php file might fix the issue as well...

http://core.trac.wordpress.org/attachment/ticket/9616/String.inc.php#L381

http://core.trac.wordpress.org/attachment/ticket/9616/String.inc.php#L487

http://core.trac.wordpress.org/attachment/ticket/9616/String.inc.php#L197

http://core.trac.wordpress.org/attachment/ticket/9616/String.inc.php#L323

... provided that this entity list is complete.

(In [11081]) Silence html_entity_decode warnings. Props Denis-de-Bernardy. fixes #9616