Improve pingback page block parsing
|Reported by:||Otto42||Owned by:|
The code in pingback_ping which reads the remote page and parses the page into a suitable comment for the pingback needs some improvement with the paragraph parse.
This line of code converts blocks into paragraph separators, which are later split up and handled independently:
$linea = preg_replace( "/ <(h1|h2|h3|h4|h5|h6|p|th|td|li|dt|dd|pre|caption|input|textarea|button|body)[^>]*>/", "\n\n", $linea );
Problem is that it's only considering blocks with spaces at the beginning of them, and not also taking closing blocks into consideration as well.
This minor improvement produces better results, I think.
$linea = preg_replace( "/<\/*(h1|h2|h3|h4|h5|h6|p|th|td|li|dt|dd|pre|caption|input|textarea|button|body)[^>]*>/", "\n\n", $linea );
Change is to eliminate the required space in front of the < and to also allow the closing blocks to mark separation of paragraphs.
This tends to give much cleaner results, especially when the last paragraph in the content is the one that contains the pingback link (pretty commonplace). This happens because by allowing closing blocks to separate the paragraphs, stuff added after them (using the_content filters) doesn't get added into that last paragraph as well. Almost all content has a /p at the end of it, and filters that add to the_content area usually use div's or span's to add to it, not p's.
Patch attached for trunk.