Does anyone have any thoughts on this? Apologies if the original post
didn’t concisely outline the issue. In short, the current libxml2 behavior
seems to result in well formed HTML being parsed in a way that is
round-tripped incorrectly and results in new elements being added.
{{{
-> libxml2 git:
Shoot..I see that the href from the example was stripped once it is
displayed on https://mail.gnome.org/archives/xml/2018-January/msg00010.html.
Here is a gist that preserves formatting:
https://gist.github.com/ptoomey3/4f684c7386229658b39b69756e262050.
On Fri, Jan 12, 2018 at 10:01 AM Patrick T
While triaging a reported cross site scripting bug we were analyzing the
behavior of our HTML sanitization code and noticed that it was parsing an
input in an unexpected way. The sanitization library itself eventually
wraps Nokogiri, which is a relatively thin wrapper around libxml2. We
reached ou
