language interpreters and shebang lines

Add an interpreter array to each language, and match interpreters found
in the shebang lines of scripts to this array to identify the language
of scripts.

With suggestions from tnm. https://github.com/github/linguist/pull/687

lib/linguist/samples.rb

@@ -52,6 +52,7 @@ module Linguist
             yield({
               :path     => File.join(dirname, filename),
               :language => category,
+              :interpreter => File.exist?(filename) ? Linguist.interpreter_from_shebang(File.read(filename)) : nil,
               :extname  => File.extname(filename)
             })
           end
@@ -67,6 +68,7 @@ module Linguist
     def self.data
       db = {}
       db['extnames'] = {}
+      db['interpreters'] = {}
       db['filenames'] = {}
 
       each do |sample|
@@ -80,6 +82,14 @@ module Linguist
           end
         end
 
+        if sample[:interpreter]
+          db['interpreters'][language_name] ||= []
+          if !db['interpreters'][language_name].include?(sample[:interpreter])
+            db['interpreters'][language_name] << sample[:interpreter]
+            db['interpreters'][language_name].sort!
+          end
+        end
+
         if sample[:filename]
           db['filenames'][language_name] ||= []
           db['filenames'][language_name] << sample[:filename]
@@ -95,4 +105,32 @@ module Linguist
       db
     end
   end
+
+  # Used to retrieve the interpreter from the shebang line of a file's
+  # data.
+  def self.interpreter_from_shebang(data)
+    lines = data.lines
+
+    if lines.any? && (match = lines[0].match(/(.+)\n?/)) && (bang = match[0]) =~ /^#!/
+      bang.sub!(/^#! /, '#!')
+      tokens = bang.split(' ')
+      pieces = tokens.first.split('/')
+
+      if pieces.size > 1
+        script = pieces.last
+      else
+        script = pieces.first.sub('#!', '')
+      end
+
+      script = script == 'env' ? tokens[1] : script
+
+      # "python2.6" -> "python"
+      if script =~ /((?:\d+\.?)+)/
+        script.sub! $1, ''
+      end
+
+      script
+    end
+  end
+
 end