Merge branch 'master' into 1036-local

Conflicts:
	lib/linguist/heuristics.rb
	lib/linguist/samples.json

lib/linguist/blob_helper.rb

@@ -321,6 +321,11 @@ module Linguist
       language ? language.lexer : Pygments::Lexer.find_by_name('Text only')
     end
 
+    # Internal: Get the TextMate compatible scope for the blob
+    def tm_scope
+      language && language.tm_scope
+    end
+
     # Public: Highlight syntax of blob
     #
     # options - A Hash of options (defaults to {})

lib/linguist/generated.rb

@@ -63,6 +63,7 @@ module Linguist
         generated_jni_header? ||
         composer_lock? ||
         node_modules? ||
+        godeps? ||
         vcr_cassette? ||
         generated_by_zephir?
     end
@@ -231,6 +232,14 @@ module Linguist
       !!name.match(/node_modules\//)
     end
 
+    # Internal: Is the blob part of Godeps/,
+    # which are not meant for humans in pull requests.
+    #
+    # Returns true or false.
+    def godeps?
+      !!name.match(/Godeps\//)
+    end
+
     # Internal: Is the blob a generated php composer lock file?
     #
     # Returns true or false.

lib/linguist/heuristics.rb

@@ -14,27 +14,25 @@ module Linguist
     def self.find_by_heuristics(data, languages)
       if active?
         if languages.all? { |l| ["Objective-C", "C++", "C"].include?(l) }
-          disambiguate_c(data, languages)
+          result = disambiguate_c(data, languages)
         end
         if languages.all? { |l| ["Perl", "Prolog"].include?(l) }
-          disambiguate_pl(data, languages)
+          result = disambiguate_pl(data, languages)
         end
         if languages.all? { |l| ["ECL", "Prolog"].include?(l) }
-          disambiguate_ecl(data, languages)
+          result = disambiguate_ecl(data, languages)
         end
-        if languages.all? { |l| ["TypeScript", "XML"].include?(l) }
-          disambiguate_ts(data, languages)
+        if languages.all? { |l| ["IDL", "Prolog"].include?(l) }
+          result = disambiguate_pro(data, languages)
         end
         if languages.all? { |l| ["Common Lisp", "OpenCL"].include?(l) }
-          disambiguate_cl(data, languages)
-        end
-        if languages.all? { |l| ["Rebol", "R"].include?(l) }
-          disambiguate_r(data, languages)
+          result = disambiguate_cl(data, languages)
         end
+        return result
       end
     end
 
-    # .h extensions are ambigious between C, C++, and Objective-C.
+    # .h extensions are ambiguous between C, C++, and Objective-C.
     # We want to shortcut look for Objective-C _and_ now C++ too!
     #
     # Returns an array of Languages or []
@@ -64,6 +62,16 @@ module Linguist
       matches
     end
 
+    def self.disambiguate_pro(data, languages)
+      matches = []
+      if (data.include?(":-"))
+        matches << Language["Prolog"]
+      else
+        matches << Language["IDL"]
+      end
+      matches
+    end
+
     def self.disambiguate_ts(data, languages)
       matches = []
       if (data.include?("</translation>"))

lib/linguist/language.rb

@@ -135,8 +135,8 @@ module Linguist
         # No shebang. Still more work to do. Try to find it with our heuristics.
         elsif (determined = Heuristics.find_by_heuristics(data, possible_language_names)) && !determined.empty?
           determined.first
-        # Lastly, fall back to the probablistic classifier.
-        elsif classified = Classifier.classify(Samples::DATA, data, possible_language_names ).first
+        # Lastly, fall back to the probabilistic classifier.
+        elsif classified = Classifier.classify(Samples.cache, data, possible_language_names).first
           # Return the actual Language object based of the string language name (i.e., first element of `#classify`)
           Language[classified[0]]
         end
@@ -290,6 +290,16 @@ module Linguist
       @lexer = Pygments::Lexer.find_by_name(attributes[:lexer] || name) ||
         raise(ArgumentError, "#{@name} is missing lexer")
 
+      @tm_scope = attributes[:tm_scope] || begin
+        context = case @type
+                  when :data, :markup, :prose
+                    'text'
+                  when :programming, nil
+                    'source'
+                  end
+        "#{context}.#{@name.downcase}"
+      end
+
       @ace_mode = attributes[:ace_mode]
       @wrap = attributes[:wrap] || false
 
@@ -363,6 +373,11 @@ module Linguist
     # Returns the Lexer
     attr_reader :lexer
 
+    # Public: Get the name of a TextMate-compatible scope
+    #
+    # Returns the scope
+    attr_reader :tm_scope
+
     # Public: Get Ace mode
     #
     # Examples
@@ -510,9 +525,9 @@ module Linguist
     end
   end
 
-  extensions = Samples::DATA['extnames']
-  interpreters = Samples::DATA['interpreters']
-  filenames = Samples::DATA['filenames']
+  extensions = Samples.cache['extnames']
+  interpreters = Samples.cache['interpreters']
+  filenames = Samples.cache['filenames']
   popular = YAML.load_file(File.expand_path("../popular.yml", __FILE__))
 
   languages_yml = File.expand_path("../languages.yml", __FILE__)
@@ -564,6 +579,7 @@ module Linguist
       :type              => options['type'],
       :aliases           => options['aliases'],
       :lexer             => options['lexer'],
+      :tm_scope          => options['tm_scope'],
       :ace_mode          => options['ace_mode'],
       :wrap              => options['wrap'],
       :group_name        => options['group'],

lib/linguist/languages.yml

@@ -83,6 +83,7 @@ ATS:
 ActionScript:
   type: programming
   lexer: ActionScript 3
+  tm_scope: source.actionscript.3
   color: "#e3491a"
   search_term: as3
   aliases:
@@ -119,7 +120,7 @@ ApacheConf:
 
 Apex:
   type: programming
-  lexer: Text only
+  lexer: Java
   extensions:
   - .cls
 
@@ -173,6 +174,7 @@ Assembly:
   - nasm
   extensions:
   - .asm
+  - .a51
 
 Augeas:
   type: programming
@@ -284,8 +286,9 @@ C:
 C#:
   type: programming
   ace_mode: csharp
+  tm_scope: source.cs
   search_term: csharp
-  color: "#5a25a2"
+  color: "#178600"
   aliases:
   - csharp
   extensions:
@@ -411,6 +414,7 @@ Clojure:
 
 CoffeeScript:
   type: programming
+  tm_scope: source.coffee
   ace_mode: coffee
   color: "#244776"
   aliases:
@@ -453,6 +457,7 @@ ColdFusion CFC:
 
 Common Lisp:
   type: programming
+  tm_scope: source.lisp
   color: "#3fb68b"
   aliases:
   - lisp
@@ -648,6 +653,7 @@ Elm:
 Emacs Lisp:
   type: programming
   lexer: Common Lisp
+  tm_scope: source.lisp
   color: "#c065db"
   aliases:
   - elisp
@@ -748,6 +754,7 @@ Forth:
   - .fth
   - .4th
   - .forth
+  - .frt
 
 Frege:
   type: programming
@@ -756,6 +763,14 @@ Frege:
   extensions:
   - .fr
 
+G-code:
+  type: data
+  lexer: Text only
+  extensions:
+  - .g
+  - .gco
+  - .gcode
+
 Game Maker Language:
   type: programming
   color: "#8ad353"
@@ -785,6 +800,12 @@ GAS:
   - .s
   - .S
 
+GDScript:
+  type: programming
+  lexer: Text only
+  extensions:
+  - .gd
+
 GLSL:
   group: C
   type: programming
@@ -877,6 +898,12 @@ Grammatical Framework:
   searchable: true
   color: "#ff0000"
 
+Graph Modeling Language:
+  type: data
+  lexer: Text only
+  extensions:
+  - .gml
+
 Groff:
   extensions:
   - .man
@@ -911,6 +938,7 @@ Groovy Server Pages:
 
 HTML:
   type: markup
+  tm_scope: text.html.basic
   ace_mode: html
   aliases:
   - xhtml
@@ -922,6 +950,7 @@ HTML:
 
 HTML+Django:
   type: markup
+  tm_scope: text.html.django
   group: HTML
   lexer: HTML+Django/Jinja
   extensions:
@@ -930,6 +959,7 @@ HTML+Django:
 
 HTML+ERB:
   type: markup
+  tm_scope: text.html.ruby
   group: HTML
   lexer: RHTML
   aliases:
@@ -940,6 +970,7 @@ HTML+ERB:
 
 HTML+PHP:
   type: markup
+  tm_scope: text.html.php
   group: HTML
   extensions:
   - .phtml
@@ -959,6 +990,8 @@ Haml:
 Handlebars:
   type: markup
   lexer: Handlebars
+  aliases:
+  - hbs
   extensions:
   - .handlebars
   - .hbs
@@ -1075,6 +1108,7 @@ J:
 
 JSON:
   type: data
+  tm_scope: source.json
   group: JavaScript
   ace_mode: json
   searchable: false
@@ -1137,6 +1171,7 @@ Java Server Pages:
 
 JavaScript:
   type: programming
+  tm_scope: source.js
   ace_mode: javascript
   color: "#f1e05a"
   aliases:
@@ -1149,6 +1184,7 @@ JavaScript:
   - .es6
   - .frag
   - .jake
+  - .jsb
   - .jsfl
   - .jsm
   - .jss
@@ -1202,7 +1238,17 @@ LFE:
 LLVM:
   extensions:
   - .ll
-  
+
+LSL:
+  type: programming
+  lexer: LSL
+  ace_mode: lsl
+  extensions:
+  - .lsl
+  interpreters:
+  - lsl
+  color: '#3d9970'
+
 LabVIEW:
   type: programming
   lexer: Text only
@@ -1254,6 +1300,7 @@ Literate Agda:
 
 Literate CoffeeScript:
   type: programming
+  tm_scope: source.litcoffee
   group: CoffeeScript
   lexer: Text only
   ace_mode: markdown
@@ -1310,6 +1357,7 @@ Lua:
   color: "#fa1fa1"
   extensions:
   - .lua
+  - .fcgi
   - .nse
   - .pd_lua
   - .rbxs
@@ -1537,6 +1585,7 @@ ObjDump:
 
 Objective-C:
   type: programming
+  tm_scope: source.objc
   color: "#438eff"
   aliases:
   - obj-c
@@ -1547,6 +1596,7 @@ Objective-C:
 
 Objective-C++:
   type: programming
+  tm_scope: source.objc++
   color: "#4886FC"
   aliases:
   - obj-c++
@@ -1637,12 +1687,14 @@ PAWN:
 
 PHP:
   type: programming
+  tm_scope: text.html.php
   ace_mode: php
   color: "#4F5D95"
   extensions:
   - .php
   - .aw
   - .ctp
+  - .fcgi
   - .module
   - .php3
   - .php4
@@ -1694,6 +1746,7 @@ Pascal:
   - .dfm
   - .dpr
   - .lpr
+  - .pp
 
 Perl:
   type: programming
@@ -1782,11 +1835,13 @@ Processing:
 
 Prolog:
   type: programming
+  lexer: Logtalk
   color: "#74283c"
   extensions:
-  - .prolog
-  - .ecl
   - .pl
+  - .ecl
+  - .pro
+  - .prolog
 
 Propeller Spin:
   type: programming
@@ -1831,6 +1886,8 @@ Python:
   color: "#3581ba"
   extensions:
   - .py
+  - .cgi
+  - .fcgi
   - .gyp
   - .lmi
   - .pyde
@@ -1992,6 +2049,7 @@ Ruby:
   extensions:
   - .rb
   - .builder
+  - .fcgi
   - .gemspec
   - .god
   - .irbrc
@@ -2039,6 +2097,7 @@ SAS:
 
 SCSS:
   type: markup
+  tm_scope: source.scss
   group: CSS
   ace_mode: scss
   extensions:
@@ -2054,6 +2113,7 @@ SQF:
 
 SQL:
   type: data
+  tm_scope: source.sql
   ace_mode: sql
   extensions:
   - .sql
@@ -2078,6 +2138,7 @@ Sage:
 
 Sass:
   type: markup
+  tm_scope: source.sass
   group: CSS
   extensions:
   - .sass
@@ -2140,6 +2201,8 @@ Shell:
   - .sh
   - .bash
   - .bats
+  - .cgi
+  - .fcgi
   - .tmux
   - .zsh
   interpreters:
@@ -2271,6 +2334,9 @@ Tcl:
   - .tcl
   - .adp
   - .tm
+  interpreters:
+  - tclsh
+  - wish
 
 Tcsh:
   type: programming
@@ -2402,6 +2468,7 @@ VimL:
   - .vim
   filenames:
   - .vimrc
+  - _vimrc
   - vimrc
   - gvimrc
 
@@ -2551,6 +2618,7 @@ Xtend:
 
 YAML:
   type: data
+  tm_scope: source.yaml
   aliases:
   - yml
   extensions:

lib/linguist/lazy_blob.rb

@@ -1,8 +1,13 @@
 require 'linguist/blob_helper'
+require 'linguist/language'
 require 'rugged'
 
 module Linguist
   class LazyBlob
+    GIT_ATTR = ['linguist-language', 'linguist-vendored']
+    GIT_ATTR_OPTS = { :priority => [:index], :skip_system => true }
+    GIT_ATTR_FLAGS = Rugged::Repository::Attributes.parse_opts(GIT_ATTR_OPTS)
+
     include BlobHelper
 
     MAX_SIZE = 128 * 1024
@@ -19,6 +24,29 @@ module Linguist
       @mode = mode
     end
 
+    def git_attributes
+      @git_attributes ||= repository.fetch_attributes(
+        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 language
+      return @language if defined?(@language)
+
+      @language = if lang = git_attributes['linguist-language']
+        Language.find_by_name(lang)
+      else
+        super
+      end
+    end
+
     def data
       load_blob!
       @data
@@ -30,6 +58,12 @@ module Linguist
     end
 
     protected
+
+    # Returns true if the attribute is present and not the string "false".
+    def boolean_attribute(attr)
+      attr != "false"
+    end
+
     def load_blob!
       @data, @size = Rugged::Blob.to_buffer(repository, oid, MAX_SIZE) if @data.nil?
     end

lib/linguist/repository.rb

@@ -110,18 +110,37 @@ module Linguist
         if @old_commit_oid == @commit_oid
           @old_stats
         else
-          compute_stats(@old_commit_oid, @commit_oid, @old_stats)
+          compute_stats(@old_commit_oid, @old_stats)
         end
       end
     end
 
-    protected
-    def compute_stats(old_commit_oid, commit_oid, cache = nil)
-      file_map = cache ? cache.dup : {}
-      old_tree = old_commit_oid && Rugged::Commit.lookup(repository, old_commit_oid).tree
-      new_tree = Rugged::Commit.lookup(repository, commit_oid).tree
+    def read_index
+      attr_index = Rugged::Index.new
+      attr_index.read_tree(current_tree)
+      repository.index = attr_index
+    end
 
-      diff = Rugged::Tree.diff(repository, old_tree, new_tree)
+    def current_tree
+      @tree ||= Rugged::Commit.lookup(repository, @commit_oid).tree
+    end
+
+    protected
+
+    def compute_stats(old_commit_oid, cache = nil)
+      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
+      if cache && diff.each_delta.any? { |delta| File.basename(delta.new_file[:path]) == ".gitattributes" }
+        diff = Rugged::Tree.diff(repository, old_tree = nil, current_tree)
+        file_map = {}
+      else
+        file_map = cache ? cache.dup : {}
+      end
 
       diff.each_delta do |delta|
         old = delta.old_file[:path]

lib/linguist/samples.rb

@@ -17,9 +17,11 @@ module Linguist
     PATH = File.expand_path('../samples.json', __FILE__)
 
     # Hash of serialized samples object
-    if File.exist?(PATH)
-      serializer = defined?(JSON) ? JSON : YAML
-      DATA = serializer.load(File.read(PATH))
+    def self.cache
+      @cache ||= begin
+        serializer = defined?(JSON) ? JSON : YAML
+        serializer.load(File.read(PATH))
+      end
     end
 
     # Public: Iterate over each sample.

lib/linguist/vendor.yml

@@ -33,6 +33,9 @@
 # Erlang bundles
 - ^rebar$
 
+# Go dependencies
+- Godeps/_workspace/
+
 # Bootstrap minified css and js
 - (^|/)bootstrap([^.]*)(\.min)?\.(js|css)$
 
@@ -235,3 +238,7 @@
 - octicons.css
 - octicons.min.css
 - sprockets-octicons.scss
+
+# Typesafe Activator
+- (^|/)activator$
+- (^|/)activator\.bat$

lib/linguist/version.rb

@@ -1,3 +1,3 @@
 module Linguist
-  VERSION = "3.1.2"
+  VERSION = "3.4.0"
 end