There's a bug in link_if_authorized that includes the :wrap_in hash as an
attribute of the A tag, even though it tries to delete it. There's also
a noted TODO that things like :id will propagate to both the A and the wrapper
element.
I think this patch should solve that. :wrap_in now takes either the old
string element, or a new hash, consisting of (a) :tag=>"li/div/etc", and
(b) anything else.
I'm a very novice Ruby programmer, so if someone (hi James) could eyeball
this and tell me if it does anything obviously stupid, I'll code up some
tests and submit it in Trac.
Index:
C:/dev/src/eclipse/radio/radiostar/vendor/plugins/user_engine/lib/user_engine.rb
===================================================================
---
C:/dev/src/eclipse/radio/radiostar/vendor/plugins/user_engine/lib/user_engine.rb
(revision 39)
+++
C:/dev/src/eclipse/radio/radiostar/vendor/plugins/user_engine/lib/user_engine.rb
(working copy)
@@ -156,13 +156,22 @@
end
(block.nil? || (yield block)) && authorized?(auth_options) {
- #result = link_to_with_current_styling(name, options, html_options,
*params)
+ wrap_in = html_options.delete(:wrap_in)
result = link_to(name, options, html_options, *params)
-
- # TODO: won't this pass other things like html_options[:id], which is
EVIL since two
- # things shouldn't share the same ID.
- wrap_tag = html_options.delete(:wrap_in)
- result = content_tag(wrap_tag, result, html_options) if wrap_tag != nil
+ if wrap_in != nil
+ if Hash === wrap_in
+ wrap_tag = wrap_in.delete(:tag)
+ wrap_options = wrap_in
+ p "it's a hash"
+ else
+ wrap_tag = wrap_in
+ wrap_options = nil
+ p "wrap_tag: "
+ p wrap_tag
+ end
+
+ result = content_tag(wrap_tag, result, wrap_options)
+ end
}
result
end
Jay Levitt
_______________________________________________
engine-users mailing list
[email protected]
http://lists.rails-engines.org/listinfo.cgi/engine-users-rails-engines.org
