Merge branch 'master' into 2512-local

2025-12-08 04:18:48 +00:00 · 2016-03-17 14:00:34 -06:00
parent ed4bbe97d1 e6dabd59ad
commit d377e23193
334 changed files with 53887 additions and 1152 deletions
--- a/lib/linguist/blob.rb
+++ b/lib/linguist/blob.rb
@@ -0,0 +1,73 @@
+require 'linguist/blob_helper'
+
+module Linguist
+  # A Blob is a wrapper around the content of a file to make it quack
+  # like a Grit::Blob. It provides the basic interface: `name`,
+  # `data`, `path` and `size`.
+  class Blob
+    include BlobHelper
+
+    # Public: Initialize a new Blob.
+    #
+    # path    - A path String (does not necessarily exists on the file system).
+    # content - Content of the file.
+    #
+    # Returns a Blob.
+    def initialize(path, content)
+      @path = path
+      @content = content
+    end
+
+    # Public: Filename
+    #
+    # Examples
+    #
+    #   Blob.new("/path/to/linguist/lib/linguist.rb", "").path
+    #   # =>  "/path/to/linguist/lib/linguist.rb"
+    #
+    # Returns a String
+    attr_reader :path
+
+    # Public: File name
+    #
+    # Returns a String
+    def name
+      File.basename(@path)
+    end
+
+    # Public: File contents.
+    #
+    # Returns a String.
+    def data
+      @content
+    end
+
+    # Public: Get byte size
+    #
+    # Returns an Integer.
+    def size
+      @content.bytesize
+    end
+
+    # Public: Get file extension.
+    #
+    # Returns a String.
+    def extension
+      extensions.last || ""
+    end
+
+    # Public: Return an array of the file extensions
+    #
+    #     >> Linguist::Blob.new("app/views/things/index.html.erb").extensions
+    #     => [".html.erb", ".erb"]
+    #
+    # Returns an Array
+    def extensions
+      basename, *segments = name.downcase.split(".")
+
+      segments.map.with_index do |segment, index|
+        "." + segments[index..-1].join(".")
+      end
+    end
+  end
+end
--- a/lib/linguist/blob_helper.rb
+++ b/lib/linguist/blob_helper.rb
@@ -6,7 +6,7 @@ require 'yaml'

 module Linguist
  # DEPRECATED Avoid mixing into Blob classes. Prefer functional interfaces
-  # like `Language.detect` over `Blob#language`. Functions are much easier to
+  # like `Linguist.detect` over `Blob#language`. Functions are much easier to
  # cache and compose.
  #
  # Avoid adding additional bloat to this module.
@@ -325,7 +325,7 @@ module Linguist
    #
    # Returns a Language or nil if none is detected
    def language
-      @language ||= Language.detect(self)
+      @language ||= Linguist.detect(self)
    end

    # Internal: Get the TextMate compatible scope for the blob
--- a/lib/linguist/documentation.yml
+++ b/lib/linguist/documentation.yml
@@ -13,11 +13,18 @@
 - (^|/)[Dd]ocumentation/
 - (^|/)javadoc/
 - ^man/
+- ^[Ee]xamples/

 ## Documentation files ##

+- (^|/)CHANGE(S|LOG)?(\.|$)
 - (^|/)CONTRIBUTING(\.|$)
 - (^|/)COPYING(\.|$)
 - (^|/)INSTALL(\.|$)
 - (^|/)LICEN[CS]E(\.|$)
+- (^|/)[Ll]icen[cs]e(\.|$)
 - (^|/)README(\.|$)
+- (^|/)[Rr]eadme(\.|$)
+
+# Samples folders
+- ^[Ss]amples/
--- a/lib/linguist/file_blob.rb
+++ b/lib/linguist/file_blob.rb
@@ -1,10 +1,11 @@
 require 'linguist/blob_helper'
+require 'linguist/blob'

 module Linguist
  # A FileBlob is a wrapper around a File object to make it quack
  # like a Grit::Blob. It provides the basic interface: `name`,
  # `data`, `path` and `size`.
-  class FileBlob
+  class FileBlob < Blob
    include BlobHelper

    # Public: Initialize a new FileBlob from a path
@@ -18,20 +19,6 @@ module Linguist
      @path = base_path ? path.sub("#{base_path}/", '') : path
    end

-    # Public: Filename
-    #
-    # Examples
-    #
-    #   FileBlob.new("/path/to/linguist/lib/linguist.rb").path
-    #   # =>  "/path/to/linguist/lib/linguist.rb"
-    #
-    #   FileBlob.new("/path/to/linguist/lib/linguist.rb",
-    #                "/path/to/linguist").path
-    #   # =>  "lib/linguist.rb"
-    #
-    # Returns a String
-    attr_reader :path
-
    # Public: Read file permissions
    #
    # Returns a String like '100644'
@@ -39,13 +26,6 @@ module Linguist
      File.stat(@fullpath).mode.to_s(8)
    end

-    # Public: File name
-    #
-    # Returns a String
-    def name
-      File.basename(@fullpath)
-    end
-
    # Public: Read file contents.
    #
    # Returns a String.
@@ -59,26 +39,5 @@ module Linguist
    def size
      File.size(@fullpath)
    end
-
-    # Public: Get file extension.
-    #
-    # Returns a String.
-    def extension
-      extensions.last || ""
-    end
-
-    # Public: Return an array of the file extensions
-    #
-    #     >> Linguist::FileBlob.new("app/views/things/index.html.erb").extensions
-    #     => [".html.erb", ".erb"]
-    #
-    # Returns an Array
-    def extensions
-      basename, *segments = name.downcase.split(".")
-
-      segments.map.with_index do |segment, index|
-        "." + segments[index..-1].join(".")
-      end
-    end
  end
 end
--- a/lib/linguist/generated.rb
+++ b/lib/linguist/generated.rb
@@ -71,7 +71,10 @@ module Linguist
      generated_jni_header? ||
      vcr_cassette? ||
      generated_module? ||
-      generated_unity3d_meta?
+      generated_unity3d_meta? ||
+      generated_racc? ||
+      generated_jflex? ||
+      generated_grammarkit?
    end

    # Internal: Is the blob an Xcode file?
@@ -241,22 +244,26 @@ module Linguist
      return lines[0].include?("Code generated by")
    end

+    PROTOBUF_EXTENSIONS = ['.py', '.java', '.h', '.cc', '.cpp']
+
    # Internal: Is the blob a C++, Java or Python source file generated by the
    # Protocol Buffer compiler?
    #
    # Returns true of false.
    def generated_protocol_buffer?
-      return false unless ['.py', '.java', '.h', '.cc', '.cpp'].include?(extname)
+      return false unless PROTOBUF_EXTENSIONS.include?(extname)
      return false unless lines.count > 1

      return lines[0].include?("Generated by the protocol buffer compiler.  DO NOT EDIT!")
    end

+    APACHE_THRIFT_EXTENSIONS = ['.rb', '.py', '.go', '.js', '.m', '.java', '.h', '.cc', '.cpp']
+
    # Internal: Is the blob generated by Apache Thrift compiler?
    #
    # Returns true or false
    def generated_apache_thrift?
-      return false unless ['.rb', '.py', '.go', '.js', '.m', '.java', '.h', '.cc', '.cpp'].include?(extname)
+      return false unless APACHE_THRIFT_EXTENSIONS.include?(extname)
      return false unless lines.count > 1

      return lines[0].include?("Autogenerated by Thrift Compiler") || lines[1].include?("Autogenerated by Thrift Compiler")
@@ -355,5 +362,45 @@ module Linguist
      return false unless lines.count > 1
      return lines[0].include?("fileFormatVersion: ")
    end
+
+    # Internal: Is this a Racc-generated file?
+    #
+    # A Racc-generated file contains:
+    # # This file is automatically generated by Racc x.y.z
+    # on the third line.
+    #
+    # Return true or false
+    def generated_racc?
+      return false unless extname == '.rb'
+      return false unless lines.count > 2
+      return lines[2].start_with?("# This file is automatically generated by Racc")
+    end
+
+    # Internal: Is this a JFlex-generated file?
+    #
+    # A JFlex-generated file contains:
+    # /* The following code was generated by JFlex x.y.z on d/at/e ti:me */
+    # on the first line.
+    #
+    # Return true or false
+    def generated_jflex?
+      return false unless extname == '.java'
+      return false unless lines.count > 1
+      return lines[0].start_with?("/* The following code was generated by JFlex ")
+    end
+
+    # Internal: Is this a GrammarKit-generated file?
+    #
+    # A GrammarKit-generated file typically contain:
+    # // This is a generated file. Not intended for manual editing.
+    # on the first line. This is not always the case, as it's possible to
+    # customize the class header.
+    #
+    # Return true or false
+    def generated_grammarkit?
+      return false unless extname == '.java'
+      return false unless lines.count > 1
+      return lines[0].start_with?("// This is a generated file. Not intended for manual editing.")
+    end
  end
 end
--- a/lib/linguist/grammars.rb
+++ b/lib/linguist/grammars.rb
@@ -1,6 +1,3 @@
-# Note: This file is included in the github-linguist-grammars gem, not the
-# github-linguist gem.
-
 module Linguist
  module Grammars
    # Get the path to the directory containing the language grammar JSON files.
--- a/lib/linguist/heuristics.rb
+++ b/lib/linguist/heuristics.rb
@@ -13,11 +13,14 @@ module Linguist
    #   ])
    #
    # Returns an Array of languages, or empty if none matched or were inconclusive.
-    def self.call(blob, languages)
+    def self.call(blob, candidates)
      data = blob.data

      @heuristics.each do |heuristic|
-        return Array(heuristic.call(data)) if heuristic.matches?(languages)
+        if heuristic.matches?(blob.name)
+          languages = Array(heuristic.call(data))
+          return languages if languages.any? || languages.all? { |l| candidates.include?(l) }
+        end
      end

      [] # No heuristics matched
@@ -30,7 +33,7 @@ module Linguist
    #
    # Examples
    #
-    #     disambiguate "Perl", "Prolog" do |data|
+    #     disambiguate ".pm" do |data|
    #       if data.include?("use strict")
    #         Language["Perl"]
    #       elsif /^[^#]+:-/.match(data)
@@ -38,22 +41,23 @@ module Linguist
    #       end
    #     end
    #
-    def self.disambiguate(*languages, &heuristic)
-      @heuristics << new(languages, &heuristic)
+    def self.disambiguate(*extensions, &heuristic)
+      @heuristics << new(extensions, &heuristic)
    end

    # Internal: Array of defined heuristics
    @heuristics = []

    # Internal
-    def initialize(languages, &heuristic)
-      @languages = languages
+    def initialize(extensions, &heuristic)
+      @extensions = extensions
      @heuristic = heuristic
    end

    # Internal: Check if this heuristic matches the candidate languages.
-    def matches?(candidates)
-      candidates.any? && candidates.all? { |l| @languages.include?(l.name) }
+    def matches?(filename)
+      filename = filename.downcase
+      @extensions.any? { |ext| filename.end_with?(ext) }
    end

    # Internal: Perform the heuristic
@@ -62,99 +66,9 @@ module Linguist
    end

    # Common heuristics
-    ObjectiveCRegex = /^[ \t]*@(interface|class|protocol|property|end|synchronised|selector|implementation)\b/
+    ObjectiveCRegex = /^\s*(@(interface|class|protocol|property|end|synchronised|selector|implementation)\b|#import\s+.+\.h[">])/

-    disambiguate "BitBake", "BlitzBasic" do |data|
-      if /^\s*; /.match(data) || data.include?("End Function")
-        Language["BlitzBasic"]
-      elsif /^\s*(# |include|require)\b/.match(data)
-        Language["BitBake"]
-      end
-    end
-
-    disambiguate "C#", "Smalltalk" do |data|
-      if /![\w\s]+methodsFor: /.match(data)
-        Language["Smalltalk"]
-      elsif /^\s*namespace\s*[\w\.]+\s*{/.match(data) || /^\s*\/\//.match(data)
-        Language["C#"]
-      end
-    end
-
-    disambiguate "Objective-C", "C++", "C" do |data|
-      if ObjectiveCRegex.match(data)
-        Language["Objective-C"]
-      elsif (/^\s*#\s*include <(cstdint|string|vector|map|list|array|bitset|queue|stack|forward_list|unordered_map|unordered_set|(i|o|io)stream)>/.match(data) ||
-        /^\s*template\s*</.match(data) || /^[ \t]*try/.match(data) || /^[ \t]*catch\s*\(/.match(data) || /^[ \t]*(class|(using[ \t]+)?namespace)\s+\w+/.match(data) || /^[ \t]*(private|public|protected):$/.match(data) || /std::\w+/.match(data))
-        Language["C++"]
-      end
-    end
-
-    disambiguate "Perl", "Perl6", "Prolog" do |data|
-      if data.include?("use v6")
-        Language["Perl6"]
-      elsif data.match(/use strict|use\s+v?5\./)
-        Language["Perl"]
-      elsif /^[^#]+:-/.match(data)
-        Language["Prolog"]
-      end
-    end
-
-    disambiguate "ECL", "Prolog" do |data|
-      if /^[^#]+:-/.match(data)
-        Language["Prolog"]
-      elsif data.include?(":=")
-        Language["ECL"]
-      end
-    end
-
-    disambiguate "IDL", "Prolog", "INI", "QMake" do |data|
-      if /^[^#]+:-/.match(data)
-        Language["Prolog"]
-      elsif data.include?("last_client=")
-        Language["INI"]
-      elsif data.include?("HEADERS") && data.include?("SOURCES")
-        Language["QMake"]
-      elsif /^\s*function[ \w,]+$/.match(data)
-        Language["IDL"]
-      end
-    end
-
-    disambiguate "GAP", "Scilab" do |data|
-      if (data.include?("gap> "))
-        Language["GAP"]
-      # Heads up - we don't usually write heuristics like this (with no regex match)
-      else
-        Language["Scilab"]
-      end
-    end
-
-    disambiguate "Common Lisp", "OpenCL", "Cool" do |data|
-      if /^\s*\((defun|in-package|defpackage) /i.match(data)
-        Language["Common Lisp"]
-      elsif /^class/x.match(data)
-        Language["Cool"]
-      elsif /\/\* |\/\/ |^\}/.match(data)
-        Language["OpenCL"]
-      end
-    end
-
-    disambiguate "Hack", "PHP" do |data|
-      if data.include?("<?hh")
-        Language["Hack"]
-      elsif /<?[^h]/.match(data)
-        Language["PHP"]
-      end
-    end
-
-    disambiguate "Scala", "SuperCollider" do |data|
-      if /\^(this|super)\./.match(data) || /^\s*(\+|\*)\s*\w+\s*{/.match(data) || /^\s*~\w+\s*=\./.match(data)
-        Language["SuperCollider"]
-      elsif /^\s*import (scala|java)\./.match(data) || /^\s*val\s+\w+\s*=/.match(data) || /^\s*class\b/.match(data)
-        Language["Scala"]
-      end
-    end
-
-    disambiguate "AsciiDoc", "AGS Script", "Public Key" do |data|
+    disambiguate ".asc" do |data|
      if /^(----[- ]BEGIN|ssh-(rsa|dss)) /.match(data)
        Language["Public Key"]
      elsif /^[=-]+(\s|\n)|{{[A-Za-z]/.match(data)
@@ -164,15 +78,75 @@ module Linguist
      end
    end

-    disambiguate "FORTRAN", "Forth", "Formatted" do |data|
+    disambiguate ".bb" do |data|
+      if /^\s*; /.match(data) || data.include?("End Function")
+        Language["BlitzBasic"]
+      elsif /^\s*(# |include|require)\b/.match(data)
+        Language["BitBake"]
+      end
+    end
+
+    disambiguate ".ch" do |data|
+      if /^\s*#\s*(if|ifdef|ifndef|define|command|xcommand|translate|xtranslate|include|pragma|undef)\b/i.match(data)
+        Language["xBase"]
+      end
+    end
+
+    disambiguate ".cl" do |data|
+      if /^\s*\((defun|in-package|defpackage) /i.match(data)
+        Language["Common Lisp"]
+      elsif /^class/x.match(data)
+        Language["Cool"]
+      elsif /\/\* |\/\/ |^\}/.match(data)
+        Language["OpenCL"]
+      end
+    end
+
+    disambiguate ".cs" do |data|
+      if /![\w\s]+methodsFor: /.match(data)
+        Language["Smalltalk"]
+      elsif /^\s*namespace\s*[\w\.]+\s*{/.match(data) || /^\s*\/\//.match(data)
+        Language["C#"]
+      end
+    end
+
+    disambiguate ".d" do |data|
+      if /^module /.match(data)
+        Language["D"]
+      elsif /^((dtrace:::)?BEGIN|provider |#pragma (D (option|attributes)|ident)\s)/.match(data)
+        Language["DTrace"]
+      elsif /(\/.*:( .* \\)$| : \\$|^ : |: \\$)/.match(data)
+        Language["Makefile"]
+      end
+    end
+
+    disambiguate ".ecl" do |data|
+      if /^[^#]+:-/.match(data)
+        Language["ECLiPSe"]
+      elsif data.include?(":=")
+        Language["ECL"]
+      end
+    end
+
+    disambiguate ".for", ".f" do |data|
      if /^: /.match(data)
        Language["Forth"]
-      elsif /^([c*][^a-z]|      (subroutine|program)\s|\s*!)/i.match(data)
+      elsif /^([c*][^abd-z]|      (subroutine|program|end)\s|\s*!)/i.match(data)
        Language["FORTRAN"]
      end
    end

-    disambiguate "F#", "Forth", "GLSL", "Filterscript" do |data|
+    disambiguate ".fr" do |data|
+      if /^(: |also |new-device|previous )/.match(data)
+        Language["Forth"]
+      elsif /^\s*(import|module|package|data|type) /.match(data)
+        Language["Frege"]
+      else
+        Language["Text"]
+      end
+    end
+
+    disambiguate ".fs" do |data|
      if /^(: |new-device)/.match(data)
        Language["Forth"]
      elsif /^\s*(#light|import|let|module|namespace|open|type)/.match(data)
@@ -184,7 +158,48 @@ module Linguist
      end
    end

-    disambiguate "Limbo", "M", "MUF", "Mathematica", "Matlab", "Mercury", "Objective-C" do |data|
+    disambiguate ".gs" do |data|
+      Language["Gosu"] if /^uses java\./.match(data)
+    end
+
+    disambiguate ".h" do |data|
+      if ObjectiveCRegex.match(data)
+        Language["Objective-C"]
+      elsif (/^\s*#\s*include <(cstdint|string|vector|map|list|array|bitset|queue|stack|forward_list|unordered_map|unordered_set|(i|o|io)stream)>/.match(data) ||
+        /^\s*template\s*</.match(data) || /^[ \t]*try/.match(data) || /^[ \t]*catch\s*\(/.match(data) || /^[ \t]*(class|(using[ \t]+)?namespace)\s+\w+/.match(data) || /^[ \t]*(private|public|protected):$/.match(data) || /std::\w+/.match(data))
+        Language["C++"]
+      end
+    end
+
+    disambiguate ".l" do |data|
+      if /\(def(un|macro)\s/.match(data)
+        Language["Common Lisp"]
+      elsif /^(%[%{}]xs|<.*>)/.match(data)
+        Language["Lex"]
+      elsif /^\.[a-z][a-z](\s|$)/i.match(data)
+        Language["Groff"]
+      elsif /^\((de|class|rel|code|data|must)\s/.match(data)
+        Language["PicoLisp"]
+      end
+    end
+
+    disambiguate ".ls" do |data|
+      if /^\s*package\s*[\w\.\/\*\s]*\s*{/.match(data)
+        Language["LoomScript"]
+      else
+        Language["LiveScript"]
+      end
+    end
+
+    disambiguate ".lsp", ".lisp" do |data|
+      if /^\s*\((defun|in-package|defpackage) /i.match(data)
+        Language["Common Lisp"]
+      elsif /^\s*\(define /.match(data)
+        Language["NewLisp"]
+      end
+    end
+
+    disambiguate ".m" do |data|
      if ObjectiveCRegex.match(data)
        Language["Objective-C"]
      elsif data.include?(":- module")
@@ -202,46 +217,144 @@ module Linguist
      end
    end

-    disambiguate "Gosu", "JavaScript" do |data|
-      Language["Gosu"] if /^uses java\./.match(data)
-    end
-
-    disambiguate "LoomScript", "LiveScript" do |data|
-      if /^\s*package\s*[\w\.\/\*\s]*\s*{/.match(data)
-        Language["LoomScript"]
-      else
-        Language["LiveScript"]
+    disambiguate ".ml" do |data|
+      if /(^\s*module)|let rec |match\s+(\S+\s)+with/.match(data)
+        Language["OCaml"]
+      elsif /=> |case\s+(\S+\s)+of/.match(data)
+        Language["Standard ML"]
      end
    end

-    disambiguate "Common Lisp", "NewLisp" do |data|
-      if /^\s*\((defun|in-package|defpackage) /i.match(data)
-        Language["Common Lisp"]
-      elsif /^\s*\(define /.match(data)
-        Language["NewLisp"]
-      end
-    end
-
-    disambiguate "TypeScript", "XML" do |data|
-      if data.include?("<TS ")
+    disambiguate ".mod" do |data|
+      if data.include?('<!ENTITY ')
        Language["XML"]
+      elsif /MODULE\s\w+\s*;/i.match(data) || /^\s*END \w+;$/i.match(data)
+        Language["Modula-2"]
      else
-        Language["TypeScript"]
+        [Language["Linux Kernel Module"], Language["AMPL"]]
      end
    end

-    disambiguate "Frege", "Forth", "Text" do |data|
-      if /^(: |also |new-device|previous )/.match(data)
-        Language["Forth"]
-      elsif /^\s*(import|module|package|data|type) /.match(data)
-        Language["Frege"]
+    disambiguate ".ms" do |data|
+      if /^[.'][a-z][a-z](\s|$)/i.match(data)
+        Language["Groff"]
+      elsif /(?<!\S)\.(include|globa?l)\s/.match(data) || /(?<!\/\*)(\A|\n)\s*\.[A-Za-z]/.match(data.gsub(/"([^\\"]|\\.)*"|'([^\\']|\\.)*'|\\\s*(?:--.*)?\n/, ""))
+        Language["GAS"]
      else
+        Language["MAXScript"]
+      end
+    end
+
+    disambiguate ".n" do |data|
+      if /^[.']/.match(data)
+        Language["Groff"]
+      elsif /^(module|namespace|using)\s/.match(data)
+        Language["Nemerle"]
+      end
+    end
+
+    disambiguate ".ncl" do |data|
+      if data.include?("THE_TITLE")
        Language["Text"]
      end
    end

-    disambiguate "PLSQL", "SQLPL", "PLpgSQL", "SQL" do |data|
-      if /^\\i\b|AS \$\$|LANGUAGE '+plpgsql'+/i.match(data) || /SECURITY (DEFINER|INVOKER)/i.match(data) || /BEGIN( WORK| TRANSACTION)?;/i.match(data)
+    disambiguate ".nl" do |data|
+      if /^(b|g)[0-9]+ /.match(data)
+        Language["NL"]
+      else
+        Language["NewLisp"]
+      end
+    end
+
+    disambiguate ".php" do |data|
+      if data.include?("<?hh")
+        Language["Hack"]
+      elsif /<?[^h]/.match(data)
+        Language["PHP"]
+      end
+    end
+
+    disambiguate ".pl" do |data|
+      if /^[^#]+:-/.match(data)
+        Language["Prolog"]
+      elsif /use strict|use\s+v?5\./.match(data)
+        Language["Perl"]
+      elsif /^(use v6|(my )?class|module)/.match(data)
+        Language["Perl6"]
+      end
+    end
+
+    disambiguate ".pm", ".t" do |data|
+      if /use strict|use\s+v?5\./.match(data)
+        Language["Perl"]
+      elsif /^(use v6|(my )?class|module)/.match(data)
+        Language["Perl6"]
+      end
+    end
+
+    disambiguate ".pod" do |data|
+      if /^=\w+$/.match(data)
+        Language["Pod"]
+      else
+        Language["Perl"]
+      end
+    end
+
+    disambiguate ".pro" do |data|
+      if /^[^#]+:-/.match(data)
+        Language["Prolog"]
+      elsif data.include?("last_client=")
+        Language["INI"]
+      elsif data.include?("HEADERS") && data.include?("SOURCES")
+        Language["QMake"]
+      elsif /^\s*function[ \w,]+$/.match(data)
+        Language["IDL"]
+      end
+    end
+
+    disambiguate ".props" do |data|
+      if /^(\s*)(<Project|<Import|<Property|<?xml|xmlns)/i.match(data)
+        Language["XML"]
+      elsif /\w+\s*=\s*/i.match(data)
+        Language["INI"]
+      end
+    end
+
+    disambiguate ".r" do |data|
+      if /\bRebol\b/i.match(data)
+        Language["Rebol"]
+      elsif data.include?("<-")
+        Language["R"]
+      end
+    end
+
+    disambiguate ".rpy" do |data|
+      if /(^(import|from|class|def)[\s\S])/m.match(data)
+        Language["Python"]
+      else
+        Language["Ren'Py"]
+      end
+    end
+
+    disambiguate ".rs" do |data|
+      if /^(use |fn |mod |pub |macro_rules|impl|#!?\[)/.match(data)
+        Language["Rust"]
+      elsif /#include|#pragma\s+(rs|version)|__attribute__/.match(data)
+        Language["RenderScript"]
+      end
+    end
+
+    disambiguate ".sc" do |data|
+      if /\^(this|super)\./.match(data) || /^\s*(\+|\*)\s*\w+\s*{/.match(data) || /^\s*~\w+\s*=\./.match(data)
+        Language["SuperCollider"]
+      elsif /^\s*import (scala|java)\./.match(data) || /^\s*val\s+\w+\s*=/.match(data) || /^\s*class\b/.match(data)
+        Language["Scala"]
+      end
+    end
+
+    disambiguate ".sql" do |data|
+      if /^\\i\b|AS \$\$|LANGUAGE '?plpgsql'?/i.match(data) || /SECURITY (DEFINER|INVOKER)/i.match(data) || /BEGIN( WORK| TRANSACTION)?;/i.match(data)
        #Postgres
        Language["PLpgSQL"]
      elsif /(alter module)|(language sql)|(begin( NOT)+ atomic)/i.match(data)  || /signal SQLSTATE '[0-9]+'/i.match(data)
@@ -256,81 +369,20 @@ module Linguist
      end
    end

-    disambiguate "D", "DTrace", "Makefile" do |data|
-      if /^module /.match(data)
-        Language["D"]
-      elsif /^((dtrace:::)?BEGIN|provider |#pragma (D (option|attributes)|ident)\s)/.match(data)
-        Language["DTrace"]
-      elsif /(\/.*:( .* \\)$| : \\$|^ : |: \\$)/.match(data)
-        Language["Makefile"]
-      end
-    end
-
-    disambiguate "OCaml", "Standard ML" do |data|
-      if /(^\s*module)|let rec |match\s+(\S+\s)+with/.match(data)
-        Language["OCaml"]
-      elsif /=> |case\s+(\S+\s)+of/.match(data)
-        Language["Standard ML"]
-      end
-    end
-
-    disambiguate "XML", "Modula-2", "Linux Kernel Module", "AMPL" do |data|
-      if data.include?('<!ENTITY ')
+    disambiguate ".ts" do |data|
+      if data.include?("<TS ")
        Language["XML"]
-      elsif /MODULE\s\w+\s*;/i.match(data) || /^\s*END \w+;$/i.match(data)
-        Language["Modula-2"]
      else
-        [Language["Linux Kernel Module"], Language["AMPL"]]
+        Language["TypeScript"]
      end
    end

-    disambiguate "Text", "NCL" do |data|
-      if data.include?("THE_TITLE")
-        Language["Text"]
-      end
-    end
-
-    disambiguate "NL", "NewLisp" do |data|
-      if /^(b|g)[0-9]+ /.match(data)
-        Language["NL"]
+    disambiguate ".tst" do |data|
+      if (data.include?("gap> "))
+        Language["GAP"]
+      # Heads up - we don't usually write heuristics like this (with no regex match)
      else
-        Language["NewLisp"]
-      end
-    end
-
-    disambiguate "Rust", "RenderScript" do |data|
-      if /^(use |fn |mod |pub |macro_rules|impl|#!?\[)/.match(data)
-        Language["Rust"]
-      elsif /#include|#pragma\s+(rs|version)|__attribute__/.match(data)
-        Language["RenderScript"]
-      end
-    end
-
-    disambiguate "Common Lisp", "Lex", "Groff", "PicoLisp" do |data|
-      if /\(def(un|macro)\s/.match(data)
-        Language["Common Lisp"]
-      elsif /^(%[%{}]xs|<.*>)/.match(data)
-        Language["Lex"]
-      elsif /^\.[a-z][a-z](\s|$)/i.match(data)
-        Language["Groff"]
-      elsif /^\((de|class|rel|code|data|must)\s/.match(data)
-        Language["PicoLisp"]
-      end
-    end
-
-    disambiguate "Groff", "Nemerle" do |data|
-      if /^[.']/.match(data)
-        Language["Groff"]
-      elsif /^(module|namespace|using)\s/.match(data)
-        Language["Nemerle"]
-      end
-    end
-
-    disambiguate "GAS", "Groff" do |data|
-      if /^[.'][a-z][a-z](\s|$)/i.match(data)
-        Language["Groff"]
-      elsif /((^|\s)move?[. ])|\.(include|globa?l)\s/.match(data)
-        Language["GAS"]
+        Language["Scilab"]
      end
    end
  end
--- a/lib/linguist/language.rb
+++ b/lib/linguist/language.rb
@@ -87,14 +87,6 @@ module Linguist
      language
    end

-    STRATEGIES = [
-      Linguist::Strategy::Modeline,
-      Linguist::Shebang,
-      Linguist::Strategy::Filename,
-      Linguist::Heuristics,
-      Linguist::Classifier
-    ]
-
    # Public: Detects the Language of the blob.
    #
    # blob - an object that includes the Linguist `BlobHelper` interface;
@@ -102,34 +94,8 @@ module Linguist
    #
    # Returns Language or nil.
    def self.detect(blob)
-      # Bail early if the blob is binary or empty.
-      return nil if blob.likely_binary? || blob.binary? || blob.empty?
-
-      Linguist.instrument("linguist.detection", :blob => blob) do
-        # Call each strategy until one candidate is returned.
-        languages = []
-        returning_strategy = nil
-
-        STRATEGIES.each do |strategy|
-          returning_strategy = strategy
-          candidates = Linguist.instrument("linguist.strategy", :blob => blob, :strategy => strategy, :candidates => languages) do
-            strategy.call(blob, languages)
-          end
-          if candidates.size == 1
-            languages = candidates
-            break
-          elsif candidates.size > 1
-            # More than one candidate was found, pass them to the next strategy.
-            languages = candidates
-          else
-            # No candidates, try the next strategy
-          end
-        end
-
-        Linguist.instrument("linguist.detected", :blob => blob, :strategy => returning_strategy, :language => languages.first)
-
-        languages.first
-      end
+      warn "[DEPRECATED] `Linguist::Language.detect` is deprecated. Use `Linguist.detect`. #{caller[0]}"
+      Linguist.detect(blob)
    end

    # Public: Get all Languages
@@ -150,7 +116,8 @@ module Linguist
    #
    # Returns the Language or nil if none was found.
    def self.find_by_name(name)
-      name && @name_index[name.downcase]
+      return nil if name.to_s.empty?
+      name && (@name_index[name.downcase] || @name_index[name.split(',').first.downcase])
    end

    # Public: Look up Language by one of its aliases.
@@ -164,7 +131,8 @@ module Linguist
    #
    # Returns the Language or nil if none was found.
    def self.find_by_alias(name)
-      name && @alias_index[name.downcase]
+      return nil if name.to_s.empty?
+      name && (@alias_index[name.downcase] || @alias_index[name.split(',').first.downcase])
    end

    # Public: Look up Languages by filename.
@@ -240,7 +208,8 @@ module Linguist
    #
    # Returns the Language or nil if none was found.
    def self.[](name)
-      name && @index[name.downcase]
+      return nil if name.to_s.empty?
+      name && (@index[name.downcase] || @index[name.split(',').first.downcase])
    end

    # Public: A List of popular languages
--- a/lib/linguist/languages.yml
+++ b/lib/linguist/languages.yml
@@ -8,7 +8,8 @@
 #                     Use "text" if a mode does not exist.
 # wrap              - Boolean wrap to enable line wrapping (default: false)
 # extensions        - An Array of associated extensions (the first one is
-#                     considered the primary extension)
+#                     considered the primary extension, the others should be
+#                     listed alphabetically)
 # interpreters      - An Array of associated interpreters
 # searchable        - Boolean flag to enable searching (defaults to true)
 # search_term       - Deprecated: Some languages maybe indexed under a
@@ -141,7 +142,7 @@ Agda:

 Alloy:
  type: programming  # 'modeling' would be more appropiate
-  color: "#cc5c24"
+  color: "#64C800"
  extensions:
  - .als
  ace_mode: text
@@ -182,6 +183,7 @@ AppleScript:
  interpreters:
  - osascript
  ace_mode: applescript
+  color: "#101F1F"

 Arc:
  type: programming
@@ -214,7 +216,7 @@ AspectJ:
  color: "#a957b0"
  extensions:
  - .aj
-  tm_scope: none
+  tm_scope: source.aspectj
  ace_mode: text

 Assembly:
@@ -289,6 +291,7 @@ Batchfile:
  - .cmd
  tm_scope: source.dosbatch
  ace_mode: batchfile
+  color: "#C1F12E"

 Befunge:
  type: programming
@@ -303,6 +306,7 @@ Bison:
  extensions:
  - .bison
  ace_mode: text
+  color: "#6A463F"

 BitBake:
  type: programming
@@ -392,6 +396,7 @@ C#:
  - csharp
  extensions:
  - .cs
+  - .cake
  - .cshtml
  - .csx

@@ -470,6 +475,13 @@ CSS:
  extensions:
  - .css

+CSV:
+  type: data
+  ace_mode: text
+  tm_scope: none
+  extensions:
+  - .csv
+
 Cap'n Proto:
  type: programming
  tm_scope: source.capnp
@@ -501,6 +513,13 @@ Chapel:
  - .chpl
  ace_mode: text

+Charity:
+  type: programming
+  extensions:
+    - .ch
+  tm_scope: none
+  ace_mode: text
+
 ChucK:
  type: programming
  extensions:
@@ -532,6 +551,14 @@ Clean:
  tm_scope: none
  ace_mode: text

+Click:
+  type: programming
+  color: "#E4E6F3"
+  extensions:
+  - .click
+  tm_scope: source.click
+  ace_mode: text
+
 Clojure:
  type: programming
  ace_mode: clojure
@@ -560,6 +587,7 @@ CoffeeScript:
  extensions:
  - .coffee
  - ._coffee
+  - .cake
  - .cjsx
  - .cson
  - .iced
@@ -609,6 +637,7 @@ Common Lisp:
  - .lsp
  - .ny
  - .podsl
+  - .sexp
  interpreters:
  - lisp
  - sbcl
@@ -619,7 +648,7 @@ Common Lisp:

 Component Pascal:
  type: programming
-  color: "#b0ce4e"
+  color: "#B0CE4E"
  extensions:
  - .cp
  - .cps
@@ -682,6 +711,7 @@ Cucumber:
  aliases:
  - gherkin
  ace_mode: text
+  color: "#5B2063"

 Cuda:
  type: programming
@@ -690,6 +720,7 @@ Cuda:
  - .cuh
  tm_scope: source.cuda-c++
  ace_mode: c_cpp
+  color: "#3A4E3A"

 Cycript:
  type: programming
@@ -711,7 +742,7 @@ Cython:

 D:
  type: programming
-  color: "#fcd46d"
+  color: "#ba595e"
  extensions:
  - .d
  - .di
@@ -743,6 +774,14 @@ DM:
  tm_scope: source.c++
  ace_mode: c_cpp

+DNS Zone:
+  type: data
+  extensions:
+  - .zone
+  - .arpa
+  tm_scope: text.zone_file
+  ace_mode: text
+
 DTrace:
  type: programming
  aliases:
@@ -774,7 +813,6 @@ Dart:

 Diff:
  type: data
-  color: "#88dddd"
  extensions:
  - .diff
  - .patch
@@ -827,6 +865,14 @@ ECL:
  tm_scope: none
  ace_mode: text

+ECLiPSe:
+  type: programming
+  group: prolog
+  extensions:
+  - .ecl
+  tm_scope: source.prolog.eclipse
+  ace_mode: prolog
+
 Eagle:
  type: markup
  color: "#814C05"
@@ -860,6 +906,8 @@ Elixir:
  ace_mode: elixir
  filenames:
  - mix.lock
+  interpreters:
+  - elixir

 Elm:
  type: programming
@@ -902,6 +950,8 @@ Erlang:
  - .es
  - .escript
  - .hrl
+  - .xrl
+  - .yrl
  filenames:
  - rebar.config
  - rebar.config.lock
@@ -987,6 +1037,7 @@ Formatted:
  type: data
  extensions:
  - .for
+  - .eam.fs
  tm_scope: none
  ace_mode: text

@@ -1004,6 +1055,16 @@ Forth:
  - .fs
  ace_mode: forth

+FreeMarker:
+  type: programming
+  color: "#0050b2"
+  aliases:
+  - ftl
+  extensions:
+  - .ftl
+  tm_scope: text.html.ftl
+  ace_mode: ftl
+
 Frege:
  type: programming
  color: "#00cafe"
@@ -1229,6 +1290,7 @@ Groff:
  - '.8'
  - '.9'
  - .l
+  - .me
  - .ms
  - .n
  - .rno
@@ -1249,6 +1311,8 @@ Groovy:
  - .gvy
  interpreters:
  - groovy
+  filenames:
+  - Jenkinsfile

 Groovy Server Pages:
  type: programming
@@ -1261,6 +1325,24 @@ Groovy Server Pages:
  tm_scope: text.html.jsp
  ace_mode: jsp

+HCL:
+  type: programming
+  extensions:
+    - .hcl
+    - .tf
+  ace_mode: ruby
+  tm_scope: source.ruby
+
+HLSL:
+  type: programming
+  extensions:
+    - .hlsl
+    - .fx
+    - .fxh
+    - .hlsli
+  ace_mode: text
+  tm_scope: none
+
 HTML:
  type: markup
  tm_scope: text.html.basic
@@ -1285,11 +1367,22 @@ HTML+Django:
  - .mustache
  - .jinja
  aliases:
+  - django
  - html+django/jinja
  - html+jinja
  - htmldjango
  ace_mode: django

+HTML+EEX:
+  type: markup
+  tm_scope: text.html.elixir
+  group: HTML
+  aliases:
+  - eex
+  extensions:
+  - .eex
+  ace_mode: text
+
 HTML+ERB:
  type: markup
  tm_scope: text.html.erb
@@ -1299,7 +1392,7 @@ HTML+ERB:
  extensions:
  - .erb
  - .erb.deface
-  ace_mode: html_ruby
+  ace_mode: text

 HTML+PHP:
  type: markup
@@ -1323,6 +1416,7 @@ Hack:
  - .hh
  - .php
  tm_scope: text.html.php
+  color: "#878787"

 Haml:
  group: HTML
@@ -1331,10 +1425,12 @@ Haml:
  - .haml
  - .haml.deface
  ace_mode: haml
+  color: "#ECE2A9"

 Handlebars:
  type: markup
  color: "#01a9d6"
+  group: HTML
  aliases:
  - hbs
  - htmlbars
@@ -1442,7 +1538,7 @@ Inform 7:
  extensions:
  - .ni
  - .i7x
-  tm_scope: source.Inform7
+  tm_scope: source.inform7
  aliases:
  - i7
  - inform7
@@ -1513,7 +1609,9 @@ JSON:
  searchable: false
  extensions:
  - .json
+  - .geojson
  - .lock
+  - .topojson
  filenames:
  - .jshintrc
  - composer.lock
@@ -1541,12 +1639,20 @@ JSONiq:
  - .jq
  tm_scope: source.jq

+JSX:
+  type: programming
+  group: JavaScript
+  extensions:
+  - .jsx
+  tm_scope: source.js.jsx
+  ace_mode: javascript
+
 Jade:
  group: HTML
  type: markup
  extensions:
  - .jade
-  tm_scope: source.jade
+  tm_scope: text.jade
  ace_mode: jade

 Jasmin:
@@ -1591,10 +1697,10 @@ JavaScript:
  - .gs
  - .jake
  - .jsb
+  - .jscad
  - .jsfl
  - .jsm
  - .jss
-  - .jsx
  - .njs
  - .pac
  - .sjs
@@ -1626,6 +1732,18 @@ Julia:
  color: "#a270ba"
  ace_mode: julia

+Jupyter Notebook:
+  type: markup
+  ace_mode: json
+  tm_scope: source.json
+  color: "#DA5B0B"
+  extensions:
+  - .ipynb
+  filenames:
+  - Notebook
+  aliases:
+  - IPython Notebook
+
 KRL:
  type: programming
  color: "#28431f"
@@ -1638,6 +1756,7 @@ KiCad:
  type: programming
  extensions:
  - .sch
+  - .brd
  - .kicad_pcb
  tm_scope: none
  ace_mode: text
@@ -1673,6 +1792,7 @@ LLVM:
  extensions:
  - .ll
  ace_mode: text
+  color: "#185619"

 LOLCODE:
  type: programming
@@ -1687,6 +1807,7 @@ LSL:
  ace_mode: lsl
  extensions:
  - .lsl
+  - .lslp
  interpreters:
  - lsl
  color: '#3d9970'
@@ -1695,8 +1816,8 @@ LabVIEW:
  type: programming
  extensions:
  - .lvproj
-  tm_scope: none
-  ace_mode: text
+  tm_scope: text.xml
+  ace_mode: xml

 Lasso:
  type: programming
@@ -1735,6 +1856,7 @@ Less:
  - .less
  tm_scope: source.css.less
  ace_mode: less
+  color: "#A1D9A1"

 Lex:
  type: programming
@@ -1886,6 +2008,34 @@ M:
  tm_scope: source.lisp
  ace_mode: lisp

+M4:
+  type: programming
+  extensions:
+  - .m4
+  tm_scope: none
+  ace_mode: text
+
+M4Sugar:
+  type: programming
+  group: M4
+  aliases:
+  - autoconf
+  extensions:
+  - .m4
+  filenames:
+  - configure.ac
+  tm_scope: none
+  ace_mode: text
+
+MAXScript:
+  type: programming
+  color: "#00a6a6"
+  extensions:
+  - .ms
+  - .mcr
+  tm_scope: source.maxscript
+  ace_mode: text
+
 MTML:
  type: markup
  color: "#b7e1f4"
@@ -1918,6 +2068,9 @@ Makefile:
  - GNUmakefile
  - Kbuild
  - Makefile
+  - Makefile.am
+  - Makefile.in
+  - Makefile.inc
  - makefile
  interpreters:
  - make
@@ -1943,6 +2096,7 @@ Markdown:
  - .mkdown
  - .ron
  tm_scope: source.gfm
+  color: "#083FA1"

 Mask:
  type: markup
@@ -1959,6 +2113,7 @@ Mathematica:
  - .cdf
  - .m
  - .ma
+  - .mt
  - .nb
  - .nbp
  - .wl
@@ -1970,6 +2125,8 @@ Mathematica:
 Matlab:
  type: programming
  color: "#bb92ac"
+  aliases:
+  - octave
  extensions:
  - .matlab
  - .m
@@ -2003,6 +2160,7 @@ MediaWiki:
  wrap: true
  extensions:
  - .mediawiki
+  - .wiki
  tm_scope: text.html.mediawiki
  ace_mode: text

@@ -2018,6 +2176,14 @@ Mercury:
  tm_scope: source.mercury
  ace_mode: prolog

+Metal:
+  type: programming
+  color: "#8f14e9"
+  extensions:
+  - .metal
+  tm_scope: source.c++
+  ace_mode: c_cpp
+
 MiniD: # Legacy
  type: programming
  searchable: false
@@ -2093,7 +2259,7 @@ Myghty:

 NCL:
  type: programming
-  color: #28431f
+  color: "#28431f"
  extensions:
  - .ncl
  tm_scope: source.ncl
@@ -2170,6 +2336,7 @@ Nginx:
  aliases:
  - nginx configuration file
  ace_mode: text
+  color: "#9469E9"

 Nimrod:
  type: programming
@@ -2228,6 +2395,7 @@ NumPy:
  - .numsc
  tm_scope: none
  ace_mode: text
+  color: "#9C8AF9"

 OCaml:
  type: programming
@@ -2397,6 +2565,7 @@ PHP:
  - .php3
  - .php4
  - .php5
+  - .phps
  - .phpt
  filenames:
  - Phakefile
@@ -2410,8 +2579,10 @@ PLSQL:
  type: programming
  ace_mode: sql
  tm_scope: source.plsql.oracle
+  color: "#dad8d8"
  extensions:
  - .pls
+  - .pck
  - .pkb
  - .pks
  - .plb
@@ -2426,6 +2597,16 @@ PLpgSQL:
  extensions:
  - .sql

+POV-Ray SDL:
+  type: programming
+  aliases:
+  - pov-ray
+  - povray
+  extensions:
+  - .pov
+  - .inc
+  ace_mode: text
+
 Pan:
  type: programming
  color: '#cc0000'
@@ -2476,7 +2657,7 @@ Parrot Internal Representation:

 Pascal:
  type: programming
-  color: "#b0ce4e"
+  color: "#E3F171"
  extensions:
  - .pas
  - .dfm
@@ -2525,9 +2706,16 @@ Perl6:
  - Rexfile
  interpreters:
  - perl6
-  tm_scope: source.perl.6
+  tm_scope: source.perl6fe
  ace_mode: perl

+Pickle:
+  type: data
+  extensions:
+  - .pkl
+  tm_scope: none
+  ace_mode: text
+
 PicoLisp:
  type: programming
  extensions:
@@ -2572,6 +2760,13 @@ PogoScript:
  tm_scope: source.pogoscript
  ace_mode: text

+Pony:
+  type: programming
+  extensions:
+  - .pony
+  tm_scope: source.pony
+  ace_mode: text
+
 PostScript:
  type: markup
  extensions:
@@ -2604,11 +2799,13 @@ Prolog:
  color: "#74283c"
  extensions:
  - .pl
-  - .ecl
  - .pro
  - .prolog
+  - .yap
  interpreters:
  - swipl
+  - yap
+  tm_scope: source.prolog
  ace_mode: prolog

 Propeller Spin:
@@ -2639,7 +2836,7 @@ Public Key:

 Puppet:
  type: programming
-  color: "#332A77"
+  color: "#302B6D"
  extensions:
  - .pp
  filenames:
@@ -2668,7 +2865,7 @@ PureScript:
  color: "#1D222D"
  extensions:
  - .purs
-  tm_scope: source.haskell
+  tm_scope: source.purescript
  ace_mode: haskell

 Python:
@@ -2677,6 +2874,7 @@ Python:
  color: "#3572A5"
  extensions:
  - .py
+  - .bzl
  - .cgi
  - .fcgi
  - .gyp
@@ -2685,10 +2883,12 @@ Python:
  - .pyp
  - .pyt
  - .pyw
+  - .rpy
  - .tac
  - .wsgi
  - .xpy
  filenames:
+  - BUCK
  - BUILD
  - SConscript
  - SConstruct
@@ -2730,7 +2930,7 @@ QMake:

 R:
  type: programming
-  color: "#198ce7"
+  color: "#198CE7"
  aliases:
  - R
  - Rscript
@@ -2746,7 +2946,7 @@ R:
  ace_mode: r

 RAML:
-  type: data
+  type: markup
  ace_mode: yaml
  tm_scope: source.yaml
  color: "#77d9fb"
@@ -2760,6 +2960,7 @@ RDoc:
  extensions:
  - .rdoc
  tm_scope: text.rdoc
+  color: "#8E84BF"

 REALbasic:
  type: programming
@@ -2789,7 +2990,7 @@ RMarkdown:
  ace_mode: markdown
  extensions:
  - .rmd
-  tm_scope: none
+  tm_scope: source.gfm

 Racket:
  type: programming
@@ -2855,6 +3056,16 @@ Redcode:
  tm_scope: none
  ace_mode: text

+Ren'Py:
+  type: programming
+  aliases:
+  - renpy
+  color: "#ff7f7f"
+  extensions:
+  - .rpy
+  tm_scope: source.renpy
+  ace_mode: python
+
 RenderScript:
  type: programming
  extensions:
@@ -2919,6 +3130,7 @@ Ruby:
  - .pryrc
  - Appraisals
  - Berksfile
+  - Brewfile
  - Buildfile
  - Deliverfile
  - Fastfile
@@ -2939,6 +3151,7 @@ Rust:
  color: "#dea584"
  extensions:
  - .rs
+  - .rs.in
  ace_mode: rust

 SAS:
@@ -2956,6 +3169,7 @@ SCSS:
  ace_mode: scss
  extensions:
  - .scss
+  color: "#CF649A"

 SMT:
  type: programming
@@ -3058,11 +3272,12 @@ Sass:
  extensions:
  - .sass
  ace_mode: sass
+  color: "#CF649A"

 Scala:
  type: programming
  ace_mode: scala
-  color: "#7dd3b0"
+  color: "#DC322F"
  extensions:
  - .scala
  - .sbt
@@ -3115,6 +3330,7 @@ Shell:
  color: "#89e051"
  aliases:
  - sh
+  - shell-script
  - bash
  - zsh
  extensions:
@@ -3125,6 +3341,7 @@ Shell:
  - .command
  - .fcgi
  - .ksh
+  - .sh.in
  - .tmux
  - .tool
  - .zsh
@@ -3167,6 +3384,7 @@ Slim:
  color: "#ff8f77"
  extensions:
  - .slim
+  tm_scope: text.slim
  ace_mode: text

 Smali:
@@ -3213,6 +3431,14 @@ Squirrel:
  tm_scope: source.c++
  ace_mode: c_cpp

+Stan:
+  type: programming
+  color: "#b2011d"
+  extensions:
+  - .stan
+  ace_mode: text
+  tm_scope: source.stan
+
 Standard ML:
  type: programming
  color: "#dc566d"
@@ -3250,9 +3476,12 @@ SuperCollider:
  type: programming
  color: "#46390b"
  extensions:
-  - .scd
  - .sc
-  tm_scope: none
+  - .scd
+  interpreters:
+  - sclang
+  - scsynth
+  tm_scope: source.supercollider
  ace_mode: text

 Swift:
@@ -3337,6 +3566,16 @@ Tea:
  tm_scope: source.tea
  ace_mode: text

+Terra:
+  type: programming
+  extensions:
+  - .t
+  color: "#00004c"
+  ace_mode: lua
+  group: Lua
+  interpreters:
+  - lua
+
 Text:
  type: prose
  wrap: true
@@ -3345,7 +3584,20 @@ Text:
  extensions:
  - .txt
  - .fr
+  - .nb
  - .ncl
+  - .no
+  filenames:
+  - COPYING
+  - INSTALL
+  - LICENSE
+  - NEWS
+  - README.me
+  - click.me
+  - delete.me
+  - keep.me
+  - read.me
+  - test.me
  tm_scope: none
  ace_mode: text

@@ -3395,6 +3647,7 @@ TypeScript:
  - ts
  extensions:
  - .ts
+  - .tsx
  tm_scope: source.ts
  ace_mode: typescript

@@ -3410,7 +3663,6 @@ Unified Parallel C:
 Unity3D Asset:
  type: data
  ace_mode: yaml
-  color: "#ab69a1"
  extensions:
  - .anim
  - .asset
@@ -3420,6 +3672,13 @@ Unity3D Asset:
  - .unity
  tm_scope: source.yaml

+Uno:
+  type: programming
+  extensions:
+  - .uno
+  ace_mode: csharp
+  tm_scope: source.cs
+
 UnrealScript:
  type: programming
  color: "#a54c4d"
@@ -3428,6 +3687,17 @@ UnrealScript:
  tm_scope: source.java
  ace_mode: java

+UrWeb:
+  type: programming
+  aliases:
+  - Ur/Web
+  - Ur
+  extensions:
+  - .ur
+  - .urs
+  tm_scope: source.ur
+  ace_mode: text
+
 VCL:
  group: Perl
  type: programming
@@ -3510,6 +3780,14 @@ Volt:
  tm_scope: source.d
  ace_mode: d

+Vue:
+  type: markup
+  color: "#2c3e50"
+  extensions:
+  - .vue
+  tm_scope: text.html.vue
+  ace_mode: html
+
 Web Ontology Language:
  type: markup
  color: "#9cc9dd"
@@ -3525,6 +3803,16 @@ WebIDL:
  tm_scope: source.webidl
  ace_mode: text

+X10:
+  type: programming
+  aliases:
+  - xten
+  ace_mode: text
+  extensions:
+  - .x10
+  color: "#4B6BEF"
+  tm_scope: source.x10
+
 XC:
  type: programming
  color: "#99DA07"
@@ -3547,6 +3835,7 @@ XML:
  - .ccxml
  - .clixml
  - .cproject
+  - .csl
  - .csproj
  - .ct
  - .dita
@@ -3562,6 +3851,7 @@ XML:
  - .iml
  - .ivy
  - .jelly
+  - .jsproj
  - .kml
  - .launch
  - .mdpolicy
@@ -3574,6 +3864,7 @@ XML:
  - .osm
  - .plist
  - .pluginspec
+  - .props
  - .ps1xml
  - .psc1
  - .pt
@@ -3592,8 +3883,10 @@ XML:
  - .tmSnippet
  - .tmTheme
  - .ts
+  - .tsx
  - .ui
  - .urdf
+  - .ux
  - .vbproj
  - .vcxproj
  - .vxml
@@ -3610,6 +3903,7 @@ XML:
  - .xliff
  - .xmi
  - .xml.dist
+  - .xproj
  - .xsd
  - .xul
  - .zcml
@@ -3622,6 +3916,14 @@ XML:
  - Web.config
  - packages.config

+XPages:
+  type: programming
+  extensions:
+  - .xsp-config
+  - .xsp.metadata
+  tm_scope: none
+  ace_mode: xml
+
 XProc:
  type: programming
  extensions:
@@ -3658,6 +3960,7 @@ XSLT:
  - .xsl
  tm_scope: text.xml.xsl
  ace_mode: xml
+  color: "#EB8CEB"

 Xojo:
  type: programming
@@ -3686,9 +3989,19 @@ YAML:
  - .yml
  - .reek
  - .rviz
+  - .sublime-syntax
+  - .syntax
  - .yaml
+  - .yaml-tmlanguage
  ace_mode: yaml

+YANG:
+  type: data
+  extensions:
+  - .yang
+  tm_scope: source.yang
+  ace_mode: text
+
 Yacc:
  type: programming
  extensions:
@@ -3697,6 +4010,7 @@ Yacc:
  - .yy
  tm_scope: source.bison
  ace_mode: text
+  color: "#4B6C4B"

 Zephir:
  type: programming
@@ -3736,7 +4050,6 @@ eC:
 edn:
  type: data
  ace_mode: clojure
-  color: "#db5855"
  extensions:
  - .edn
  tm_scope: source.clojure
@@ -3778,7 +4091,10 @@ reStructuredText:
  extensions:
  - .rst
  - .rest
+  - .rest.txt
+  - .rst.txt
  ace_mode: text
+  color: "#B3BCBC"

 wisp:
  type: programming
@@ -3791,8 +4107,13 @@ wisp:
 xBase:
  type: programming
  color: "#403a40"
+  aliases:
+  - advpl
+  - clipper
+  - foxpro
  extensions:
  - .prg
  - .ch
+  - .prw
  tm_scope: source.harbour
  ace_mode: text
--- a/lib/linguist/lazy_blob.rb
+++ b/lib/linguist/lazy_blob.rb
@@ -4,7 +4,11 @@ require 'rugged'

 module Linguist
  class LazyBlob
-    GIT_ATTR = ['linguist-documentation', 'linguist-language', 'linguist-vendored']
+    GIT_ATTR = ['linguist-documentation',
+                'linguist-language',
+                'linguist-vendored',
+                'linguist-generated']
+
    GIT_ATTR_OPTS = { :priority => [:index], :skip_system => true }
    GIT_ATTR_FLAGS = Rugged::Repository::Attributes.parse_opts(GIT_ATTR_OPTS)

@@ -31,14 +35,6 @@ module Linguist
        name, GIT_ATTR, GIT_ATTR_FLAGS)
    end

-    def vendored?
-      if attr = git_attributes['linguist-vendored']
-        return boolean_attribute(attr)
-      else
-        return super
-      end
-    end
-
    def documentation?
      if attr = git_attributes['linguist-documentation']
        boolean_attribute(attr)
@@ -47,6 +43,22 @@ module Linguist
      end
    end

+    def generated?
+      if attr = git_attributes['linguist-generated']
+        boolean_attribute(attr)
+      else
+        super
+      end
+    end
+
+    def vendored?
+      if attr = git_attributes['linguist-vendored']
+        return boolean_attribute(attr)
+      else
+        super
+      end
+    end
+
    def language
      return @language if defined?(@language)

@@ -67,11 +79,15 @@ module Linguist
      @size
    end

+    def cleanup!
+      @data.clear if @data
+    end
+
    protected

    # Returns true if the attribute is present and not the string "false".
-    def boolean_attribute(attr)
-      attr != "false"
+    def boolean_attribute(attribute)
+      attribute != "false"
    end

    def load_blob!
--- a/lib/linguist/repository.rb
+++ b/lib/linguist/repository.rb
@@ -126,12 +126,13 @@ module Linguist
    end

    protected
+    MAX_TREE_SIZE = 100_000

    def compute_stats(old_commit_oid, cache = nil)
+      return {} if current_tree.count_recursive(MAX_TREE_SIZE) >= MAX_TREE_SIZE
+
      old_tree = old_commit_oid && Rugged::Commit.lookup(repository, old_commit_oid).tree
-
      read_index
-
      diff = Rugged::Tree.diff(repository, old_tree, current_tree)

      # Clear file map and fetch full diff if any .gitattributes files are changed
@@ -157,8 +158,11 @@ module Linguist

          blob = Linguist::LazyBlob.new(repository, delta.new_file[:oid], new, mode.to_s(8))

-          next unless blob.include_in_language_stats?
-          file_map[new] = [blob.language.group.name, blob.size]
+          if blob.include_in_language_stats?
+            file_map[new] = [blob.language.group.name, blob.size]
+          end
+
+          blob.cleanup!
        end
      end

--- a/lib/linguist/strategy/modeline.rb
+++ b/lib/linguist/strategy/modeline.rb
@@ -1,8 +1,19 @@
 module Linguist
  module Strategy
    class Modeline
-      EmacsModeline = /-\*-\s*(?:(?!mode)[\w-]+\s*:\s*(?:[\w+-]+)\s*;?\s*)*(?:mode\s*:)?\s*([\w+-]+)\s*(?:;\s*(?!mode)[\w-]+\s*:\s*[\w+-]+\s*)*;?\s*-\*-/i
-      VimModeline = /vim:\s*set\s*(?:ft|filetype)=(\w+):/i
+      EMACS_MODELINE = /-\*-\s*(?:(?!mode)[\w-]+\s*:\s*(?:[\w+-]+)\s*;?\s*)*(?:mode\s*:)?\s*([\w+-]+)\s*(?:;\s*(?!mode)[\w-]+\s*:\s*[\w+-]+\s*)*;?\s*-\*-/i
+
+      # First form vim modeline
+      # [text]{white}{vi:|vim:|ex:}[white]{options}
+      # ex: 'vim: syntax=ruby'
+      VIM_MODELINE_1 = /(?:vim|vi|ex):\s*(?:ft|filetype|syntax)=(\w+)\s?/i
+
+      # Second form vim modeline (compatible with some versions of Vi)
+      # [text]{white}{vi:|vim:|Vim:|ex:}[white]se[t] {options}:[text]
+      # ex: 'vim set syntax=ruby:'
+      VIM_MODELINE_2 = /(?:vim|vi|Vim|ex):\s*se(?:t)?.*\s(?:ft|filetype|syntax)=(\w+)\s?.*:/i
+
+      MODELINES = [EMACS_MODELINE, VIM_MODELINE_1, VIM_MODELINE_2]

      # Public: Detects language based on Vim and Emacs modelines
      #
@@ -22,7 +33,7 @@ module Linguist
      #
      # Returns a String or nil
      def self.modeline(data)
-        match = data.match(EmacsModeline) || data.match(VimModeline)
+        match = MODELINES.map { |regex| data.match(regex) }.reject(&:nil?).first
        match[1] if match
      end
    end
--- a/lib/linguist/tokenizer.rb
+++ b/lib/linguist/tokenizer.rb
@@ -86,17 +86,17 @@ module Linguist
          if s.peek(1) == "\""
            s.getch
          else
-            s.skip_until(/[^\\]"/)
+            s.skip_until(/(?<!\\)"/)
          end
        elsif s.scan(/'/)
          if s.peek(1) == "'"
            s.getch
          else
-            s.skip_until(/[^\\]'/)
+            s.skip_until(/(?<!\\)'/)
          end

        # Skip number literals
-        elsif s.scan(/(0x)?\d(\d|\.)*/)
+        elsif s.scan(/(0x\h(\h|\.)*|\d(\d|\.)*)([uU][lL]{0,2}|([eE][-+]\d*)?[fFlL]*)/)

        # SGML style brackets
        elsif token = s.scan(/<[^\s<>][^<>]*>/)
--- a/lib/linguist/vendor.yml
+++ b/lib/linguist/vendor.yml
@@ -20,10 +20,17 @@
 - ^deps/
 - ^tools/
 - (^|/)configure$
- (^|/)configure.ac$
 - (^|/)config.guess$
 - (^|/)config.sub$

+# stuff autogenerated by autoconf - still C deps
+- (^|/)aclocal.m4
+- (^|/)libtool.m4
+- (^|/)ltoptions.m4
+- (^|/)ltsugar.m4
+- (^|/)ltversion.m4
+- (^|/)lt~obsolete.m4
+
 # Linters
 - cpplint.py

@@ -78,6 +85,9 @@
 # Haxelib projects often contain a neko bytecode file named run.n
 - run.n$

+# Bootstrap Datepicker
+- bootstrap-datepicker/
+
 ## Commonly Bundled JavaScript frameworks ##

 # jQuery
@@ -88,6 +98,34 @@
 - (^|/)jquery\-ui(\-\d\.\d+(\.\d+)?)?(\.\w+)?\.(js|css)$
 - (^|/)jquery\.(ui|effects)\.([^.]*)\.(js|css)$

+# jQuery Gantt
+- jquery.fn.gantt.js
+
+# jQuery fancyBox
+- jquery.fancybox.(js|css)
+
+# Fuel UX
+- fuelux.js
+
+# jQuery File Upload
+- (^|/)jquery\.fileupload(-\w+)?\.js$
+
+# Slick
+- (^|/)slick\.\w+.js$
+
+# Leaflet plugins
+- (^|/)Leaflet\.Coordinates-\d+\.\d+\.\d+\.src\.js$
+- leaflet.draw-src.js
+- leaflet.draw.css
+- Control.FullScreen.css
+- Control.FullScreen.js
+- leaflet.spin.js
+- wicket-leaflet.js
+
+# Sublime Text workspace files
+- .sublime-project
+- .sublime-workspace
+
 # Prototype
 - (^|/)prototype(.*)\.js$
 - (^|/)effects\.js$
@@ -122,7 +160,7 @@
 - (^|/)Chart\.js$

 # Codemirror
- (^|/)[Cc]ode[Mm]irror/(lib|mode|theme|addon|keymap)
+- (^|/)[Cc]ode[Mm]irror/(\d+\.\d+/)?(lib|mode|theme|addon|keymap|demo)

 # SyntaxHighlighter - http://alexgorbatchev.com/
 - (^|/)shBrush([^.]*)\.js$
@@ -164,6 +202,11 @@

 ## Obj-C ##

+# Xcode
+
+- \.xctemplate/
+- \.imageset/
+
 # Carthage
 - ^Carthage/

@@ -179,6 +222,10 @@
 # Fabric
 - Fabric.framework/

+# git config files
+- gitattributes$
+- gitignore$
+- gitmodules$

 ## Groovy ##

@@ -224,21 +271,9 @@
 # Html5shiv
 - (^|/)html5shiv\.js$

-# Samples folders
- ^[Ss]amples/
-
-# LICENSE, README, git config files
- ^COPYING$
- LICENSE$
- License$
- gitattributes$
- gitignore$
- gitmodules$
- ^README$
- ^readme$
-
 # Test fixtures
 - ^[Tt]ests?/fixtures/
+- ^[Ss]pecs?/fixtures/

 # PhoneGap/Cordova
 - (^|/)cordova([^.]*)\.js$
@@ -274,3 +309,6 @@

 # Android Google APIs
 - (^|/)\.google_apis/
+
+# Jenkins Pipeline
+- ^Jenkinsfile$
--- a/lib/linguist/version.rb
+++ b/lib/linguist/version.rb
@@ -1,3 +1,3 @@
 module Linguist
-  VERSION = "4.5.9"
+  VERSION = "4.8.0"
 end