Parsing HTML unescapes all escaped entities about hickory HOT 3 CLOSED

I was able to fix it like so:

diff --git a/src/hickory/core.clj b/src/hickory/core.clj
index 22fcdd2..5c75426 100644
--- a/src/hickory/core.clj
+++ b/src/hickory/core.clj
@@ -57,7 +57,7 @@
                                       (as-hiccup (.attributes this))]
                                      (map as-hiccup (.childNodes this)))))
   TextNode
-  (as-hiccup [this] (.text this))
+  (as-hiccup [this] (.outerHtml this))
   XmlDeclaration
   (as-hiccup [this] (str this)))

@@ -86,7 +86,7 @@
                       :content (not-empty (into [] (map as-hickory
                                                         (.childNodes this))))})
   TextNode
-  (as-hickory [this] (.text this)))
+  (as-hickory [this] (.outerHtml this)))

 (defn parse
   "Parse an entire HTML document into a DOM structure that can be

The .outerHtml function of nodes seems to return the text untouched. I noticed that .text also seems to be doing other weird formatting things like stripping line breaks and stuff. I'm not sure if outerHtml is what we want here, but I didn't see anything else in the API docs that wouldn't unescape stuff. http://jsoup.org/apidocs/org/jsoup/nodes/TextNode.html

Furthermore, I'm not entirely certain if you consider this a bug or not. I suppose unescaping makes sense for certain uses of the library, but if you're redistributing the HTML when you're finished mucking with it you definitely want what was escaped to stay that way.

from hickory.

You are absolutely correct to flag the issue of whether it should be parsing escaped or unescaped text. There are use cases in which you'd want either, and currently it is only supporting one of those. However, simply switching the code to the version that addresses your use case is not the best way to handle this.

I think the way to handle this is to change the parsing functions so they can take keyword arguments, of which we'd add one (for now): :unescaped-html, which defaults to false. This should leave user source code unchanged. It will also match the behavior of clojure.xml, which unescapes the XML it parses. Of course, passing the options map around will require a change to the function signatures in the protocol, but that should be fairly mechanical.

David

from hickory.

(After much discussion with you on IRC), I believe version 0.2.1 will address your problems.

from hickory.