Decode URL entities to avoid broken links

<a href="http://example.com/foo?moo=kooh&affe=tiger">example</a>
shouldn't point to http://example.com/foo?moo=kooh&amp;affe=tiger

Signed-off-by: Julien Danjou <julien@danjou.info>

diff --git a/lisp/ChangeLog b/lisp/ChangeLog
index 71a642d..b240ad7 100644 (file)
--- a/lisp/ChangeLog
+++ b/lisp/ChangeLog
@@ -7,6 +7,11 @@
 
        * gnus-art.el (gnus-mime-delete-part): Fix Lisp type of byte(s).
 
+2010-09-26  Florian Ragwitz  <rafl@debian.org>
+
+       * gnus-html.el (gnus-html-wash-tags): Decode URL entities to avoid
+       handing broken links to browse-url.
+
 2010-09-26  Lars Magne Ingebrigtsen  <larsi@gnus.org>
 
        * nndoc.el (nndoc-request-list): Return success always.

diff --git a/lisp/gnus-html.el b/lisp/gnus-html.el
index 587c28e..4bac528 100644 (file)
--- a/lisp/gnus-html.el
+++ b/lisp/gnus-html.el
@@ -286,7 +286,7 @@ Use ALT-TEXT for the image string."
          (setq url (match-string 1 parameters))
           (gnus-message 8 "gnus-html-wash-tags: fetching link URL %s" url)
          (gnus-article-add-button start end
-                                  'browse-url url
+                                  'browse-url (mm-url-decode-entities-string url)
                                   url)
          (let ((overlay (gnus-make-overlay start end)))
            (gnus-overlay-put overlay 'evaporate t)