Merge pull request #250 from github/mac-format

Handle Mac Format when splitting lines

3 files changed, 34 insertions, 1 deletions

diff --git a/lib/linguist/blob_helper.rb b/lib/linguist/blob_helper.rb
index 04c3f0a..b0a2062 100644
--- a/lib/linguist/blob_helper.rb
+++ b/lib/linguist/blob_helper.rb
@@ -204,7 +204,31 @@ module Linguist
     #
     # Returns an Array of lines
     def lines
-      @lines ||= (viewable? && data) ? data.split("\n", -1) : []
+      @lines ||=
+        if viewable? && data
+          data.split(line_split_character, -1)
+        else
+          []
+        end
+    end
+
+    # Character used to split lines. This is almost always "\n" except when Mac
+    # Format is detected in which case it's "\r".
+    #
+    # Returns a split pattern string.
+    def line_split_character
+      @line_split_character ||= (mac_format?? "\r" : "\n")
+    end
+
+    # Public: Is the data in ** Mac Format **. This format uses \r (0x0d) characters
+    # for line ends and does not include a \n (0x0a).
+    #
+    # Returns true when mac format is detected.
+    def mac_format?
+      return if !viewable?
+      if pos = data[0, 4096].index("\r")
+        data[pos + 1] != ?\n
+      end
     end
 
     # Public: Get number of lines of code
diff --git a/samples/Text/mac.txt b/samples/Text/mac.txt
new file mode 100644
index 0000000..c22612b
--- /dev/null
+++ b/samples/Text/mac.txt
@@ -0,0 +1 @@
+line 1
line 2
\ No newline at end of file
diff --git a/test/test_blob.rb b/test/test_blob.rb
index 17e9ef8..333b0ce 100644
--- a/test/test_blob.rb
+++ b/test/test_blob.rb
@@ -65,6 +65,14 @@ class TestBlob < Test::Unit::TestCase
     assert_equal ["module Foo", "end", ""], blob("Ruby/foo.rb").lines
   end
 
+  def test_mac_format
+    assert blob("Text/mac.txt").mac_format?
+  end
+
+  def test_lines_mac_format
+    assert_equal ["line 1", "line 2", ""], blob("Text/mac.txt").lines
+  end
+
   def test_size
     assert_equal 15, blob("Ruby/foo.rb").size
   end