From dae09d983371d5858e028b37afd7d2b3877b776d Mon Sep 17 00:00:00 2001
From: Masato Ikeda <masato.ikeda@gmail.com>
Date: Thu, 16 May 2013 02:16:38 +0900
Subject: [PATCH] Add LTSV scanner

LTSV is a text format described at: http://ltsv.org/
---
 lib/coderay/helpers/file_type.rb |  1 +
 lib/coderay/scanners/ltsv.rb     | 49 ++++++++++++++++++++++++++++++++
 2 files changed, 50 insertions(+)
 create mode 100644 lib/coderay/scanners/ltsv.rb

diff --git a/lib/coderay/helpers/file_type.rb b/lib/coderay/helpers/file_type.rb
index f52f17e3..c47ae551 100644
--- a/lib/coderay/helpers/file_type.rb
+++ b/lib/coderay/helpers/file_type.rb
@@ -96,6 +96,7 @@ def shebang filename
       'java'     => :java,
       'js'       => :java_script,
       'json'     => :json,
+      'ltsv'     => :ltsv,
       'mab'      => :ruby,
       'pas'      => :delphi,
       'patch'    => :diff,
diff --git a/lib/coderay/scanners/ltsv.rb b/lib/coderay/scanners/ltsv.rb
new file mode 100644
index 00000000..91d8c345
--- /dev/null
+++ b/lib/coderay/scanners/ltsv.rb
@@ -0,0 +1,49 @@
+module CodeRay
+module Scanners
+
+  # A scanner for LTSV.
+  # http://ltsv.org
+  class Ltsv < Scanner
+
+    register_for :ltsv
+    file_extension 'ltsv'
+
+  protected
+
+    def scan_tokens encoder, options
+
+      state = :initial
+
+      until eos?
+
+        if match = scan(/(?:\r?\n)+/)
+          encoder.text_token match, :space
+          state = :initial if match.index(/\r?\n/)
+
+        elsif (state == :initial || state == :tab) && match = scan(/[^:]+/)
+          encoder.text_token match, :key
+          state = :label
+
+        elsif state == :label && match = scan(/:/)
+          encoder.text_token match, :delimiter
+          state = :colon
+
+        elsif state == :colon && match = scan(/[^\t\r\n]*/)
+          encoder.text_token match, :value
+          state = :value
+
+        elsif state == :value && match = scan(/\t/)
+          encoder.text_token match, :space
+          state = :tab
+
+        else
+          raise
+        end
+      end
+
+      encoder
+    end
+  end
+
+end
+end