WordPress fails to ping whole url if there's an ampersand

[unpix]

I have a custom blog and I noticed a problem with pingbacks from Wordpress blogs. When writing a post in Wordpress, it encodes the "&" properly to

&

html entity.

However, when it tries to ping the url, it uses the url where the ampersands are encoded as html entities. This causes Wordpress to strip out every GET variable after the first one, for example:

URL should be:

http://example.com/index.php?a=b&c=d

but it is

http://example.com/index.php?a=b&amp

So Wordpress misses everything after the first variable in the url and fails to ping the right url.

I can see this from my server logs.

[markjaquith]

To be clear, this is a Pingback issue.

[wonderboymusic]

The regex used to extract URLs for pinging is old and odd - I updated it to use this one, which I tend to prefer: ​http://daringfireball.net/2010/07/improved_regex_for_matching_urls

I also moved it into a function so the regex isn't duped across functions. Patch attached.

[wonderboymusic]

Duplicate of #9064.