Elixir: Make regexps greedy, remove comment hacks. Update known failu…

…res and tests.

components/prism-elixir.js

@@ -1,13 +1,13 @@
 Prism.languages.elixir = {
-	// Negative look-ahead is needed for string interpolation
-	// Negative look-behind is needed to avoid highlighting markdown headers in
-	// multi-line doc strings
 	'comment': {
-		pattern: /(^|[^#])#(?![{#]).*/m,
+		pattern: /#.*/m,
 		lookbehind: true
 	},
 	// ~r"""foo""" (multi-line), ~r'''foo''' (multi-line), ~r/foo/, ~r|foo|, ~r"foo", ~r'foo', ~r(foo), ~r[foo], ~r{foo}, ~r<foo>
-	'regex': /~[rR](?:("""|''')(?:\\[\s\S]|(?!\1)[^\\])+\1|([\/|"'])(?:\\.|(?!\2)[^\\\r\n])+\2|\((?:\\.|[^\\)\r\n])+\)|\[(?:\\.|[^\\\]\r\n])+\]|\{(?:\\.|[^\\}\r\n])+\}|<(?:\\.|[^\\>\r\n])+>)[uismxfr]*/,
+	'regex': {
+		pattern: /~[rR](?:("""|''')(?:\\[\s\S]|(?!\1)[^\\])+\1|([\/|"'])(?:\\.|(?!\2)[^\\\r\n])+\2|\((?:\\.|[^\\)\r\n])+\)|\[(?:\\.|[^\\\]\r\n])+\]|\{(?:\\.|[^\\}\r\n])+\}|<(?:\\.|[^\\>\r\n])+>)[uismxfr]*/,
+		greedy: true
+	},
 	'string': [
 		{
 			// ~s"""foo""" (multi-line), ~s'''foo''' (multi-line), ~s/foo/, ~s|foo|, ~s"foo", ~s'foo', ~s(foo), ~s[foo], ~s{foo} (with interpolation care), ~s<foo>

examples/prism-elixir.html

@@ -458,9 +458,5 @@ <h2>Known failures</h2>
 	If a failure is listed here, it doesn’t mean it will never be fixed. This is more of a “known bugs” list, just with a certain type of bug.
 </p>
 
-<h3>Comments starting with an opening curly brace</h3>
-<pre><code>#{ This is not highlighted properly.</code></pre>
-
 <h3>String interpolation in single-quoted strings</h3>
-<pre><code>'#{:atom} &lt;- this should not be highligted'</code></pre>
-
+<pre><code>'#{:atom} &lt;- this should not be highligted'</code></pre>

tests/languages/elixir/comment_feature.test

@@ -1,11 +1,13 @@
 #
 # Foobar
+#{ This is a comment
 
 ----------------------------------------------------
 
 [
 	["comment", "#"],
-	["comment", "# Foobar"]
+	["comment", "# Foobar"],
+	["comment", "#{ This is a comment"]
 ]
 
 ----------------------------------------------------