- Better handling of magic encoding comment edge cases (fixes #33).

whitequark · May 15, 2013 · 4ad308a · 4ad308a
1 parent d43338f
commit 4ad308a
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 6 deletions.
diff --git a/lib/parser/source/buffer.rb b/lib/parser/source/buffer.rb
@@ -8,12 +8,15 @@ class Buffer
 
       def self.recognize_encoding(string)
         if string.empty?
-          return Encoding::UTF_8
+          return Encoding::BINARY
         end
 
         # TODO: Make this more efficient.
         first_line, second_line = string.lines.first(2)
-        first_line.force_encoding(Encoding::ASCII_8BIT)
+
+        [first_line, second_line].each do |line|
+          line.force_encoding(Encoding::ASCII_8BIT) if line
+        end
 
         if first_line =~ /\A\xef\xbb\xbf/ # BOM
           return Encoding::UTF_8
@@ -23,9 +26,7 @@ def self.recognize_encoding(string)
           encoding_line = first_line
         end
 
-        encoding_line.force_encoding(Encoding::ASCII_8BIT)
-
-        if encoding_line =~ /coding[:=]?\s*([a-z0-9_-]+)/
+        if encoding_line =~ /^#.*coding[:=]?\s*([a-z0-9_-]+)/
           Encoding.find($1)
         else
           string.encoding

diff --git a/test/test_encoding.rb b/test/test_encoding.rb
@@ -25,7 +25,11 @@ def test_shebang
     end
 
     def test_empty
-      assert_equal Encoding::UTF_8, recognize("")
+      assert_equal Encoding::BINARY, recognize("")
+    end
+
+    def test_false_positive
+      assert_equal Encoding::BINARY, recognize(%{require 'cane/encoding_aware_iterator'})
     end
   end
 end