Fix issue with import of code containing '<'

stomar · stomar · commit fc5bb6441d9c · 2013-02-25T19:15:23.000+01:00
This fixes the problem that all code from an opening angle bracket
through to the next closing bracket or the end of the code element
is lost at import, since Nokogiri::XML::Node#replace interprets the
bracket as the start of an HTML tag.  By instead using the
content= method the new content string gets XML escaped.
diff --git a/Rakefile b/Rakefile
@@ -84,13 +84,13 @@ def html_to_markdown(content_div)
 
     # map all code elements to their inner_text
     pre.search('code').each do |code|
-      code.replace(highlight(code_to_text(code), lang))
+      code.content = highlight(code_to_text(code), lang)
     end
   end
 
   # map all code elements to their inner_text
   content_div.search('pre > code').each do |code|
-    code.replace(code_to_text(code))
+    code.content = code_to_text(code)
   end
 
   # remove the 'class' attribute from all pre tags
