HTML block parser was confusing backtick fenced code blocks with code spans, which has no consequence most of the time but could be problematic if you're writing code spans inside of the code block. Code span markers are now matched only when backtick fence matching has failed. Fixed an issue were two consecutive fenced code blocks with no blank line between them which was causing the next paragraph to become a code block (with both tilde and backtick fences).
Removing the additional code that I applied previously to implement the `data-lang` attribute. In retrospect, this was a bad idea; as the existing implementation does a good job with this already when used properly. Many thanks to Michel Fortin for his help in accomplishing these changes :-) I hope this might be merged into the official copy at some point in the future. Conflicts: markdown.php
… by PCRE (depends on PCRE version used).
…xtra to 1.2.6. Conflicts: License.text PHP Markdown Extra Readme.text markdown.php
…and bBlog seems to be dead.
…parser-object variables too.
Merge remote-tracking branch 'coreyworrell/extra' into extra
… preceded by whitespace. Also made the regex a little more strict, so that </br/> wouldn't be seen as a valid tag.