Merge branch 'master' of github.com:github/linguist

Conflicts: lib/linguist/languages.yml
2025-12-07 20:08:48 +00:00 · 2013-03-18 01:36:19 +02:00
parent cdde73f5ee 4444b6daa1
commit f76d64f9aa
85 changed files with 36450 additions and 3916 deletions
--- a/lib/linguist/blob_helper.rb
+++ b/lib/linguist/blob_helper.rb
@@ -58,6 +58,15 @@ module Linguist
      _mime_type ? _mime_type.binary? : false
    end

+    # Internal: Is the blob binary according to its mime type,
+    # overriding it if we have better data from the languages.yml
+    # database.
+    #
+    # Return true or false
+    def likely_binary?
+       binary_mime_type? and not Language.find_by_filename(name)
+    end
+
    # Public: Get the Content-Type header value
    #
    # This value is used when serving raw blobs.
@@ -139,7 +148,14 @@ module Linguist
    #
    # Return true or false
    def image?
-      ['.png', '.jpg', '.jpeg', '.gif'].include?(extname)
+      ['.png', '.jpg', '.jpeg', '.gif'].include?(extname.downcase)
+    end
+
+    # Public: Is the blob a supported 3D model format?
+    #
+    # Return true or false
+    def solid?
+      extname.downcase == '.stl'
    end

    MEGABYTE = 1024 * 1024
@@ -251,7 +267,7 @@ module Linguist

    # Public: Is the blob a generated file?
    #
-    # Generated source code is supressed in diffs and is ignored by
+    # Generated source code is suppressed in diffs and is ignored by
    # language statistics.
    #
    # May load Blob#data
@@ -266,7 +282,7 @@ module Linguist
    # Excluded:
    # - Files over 0.1MB
    # - Non-text files
-    # - Langauges marked as not searchable
+    # - Languages marked as not searchable
    # - Generated source files
    #
    # Please add additional test coverage to
--- a/lib/linguist/classifier.rb
+++ b/lib/linguist/classifier.rb
@@ -40,7 +40,7 @@ module Linguist

    # Public: Guess language of data.
    #
-    # db        - Hash of classifer tokens database.
+    # db        - Hash of classifier tokens database.
    # data      - Array of tokens or String data to analyze.
    # languages - Array of language name Strings to restrict to.
    #
@@ -85,7 +85,7 @@ module Linguist
      scores.sort { |a, b| b[1] <=> a[1] }.map { |score| [score[0], score[1]] }
    end

-    # Internal: Probably of set of tokens in a language occuring - P(D | C)
+    # Internal: Probably of set of tokens in a language occurring - P(D | C)
    #
    # tokens   - Array of String tokens.
    # language - Language to check.
@@ -97,7 +97,7 @@ module Linguist
      end
    end

-    # Internal: Probably of token in language occuring - P(F | C)
+    # Internal: Probably of token in language occurring - P(F | C)
    #
    # token    - String token.
    # language - Language to check.
@@ -111,7 +111,7 @@ module Linguist
      end
    end

-    # Internal: Probably of a language occuring - P(C)
+    # Internal: Probably of a language occurring - P(C)
    #
    # language - Language to check.
    #
--- a/lib/linguist/generated.rb
+++ b/lib/linguist/generated.rb
@@ -43,7 +43,7 @@ module Linguist

    # Internal: Is the blob a generated file?
    #
-    # Generated source code is supressed in diffs and is ignored by
+    # Generated source code is suppressed in diffs and is ignored by
    # language statistics.
    #
    # Please add additional test coverage to
@@ -86,7 +86,7 @@ module Linguist

    # Internal: Is the blob of JS generated by CoffeeScript?
    #
-    # CoffeScript is meant to output JS that would be difficult to
+    # CoffeeScript is meant to output JS that would be difficult to
    # tell if it was generated or not. Look for a number of patterns
    # output by the CS compiler.
    #
--- a/lib/linguist/language.rb
+++ b/lib/linguist/language.rb
@@ -73,7 +73,7 @@ module Linguist
    #
    # Returns Language or nil.
    def self.detect(name, data, mode = nil)
-      # A bit of an elegant hack. If the file is exectable but extensionless,
+      # A bit of an elegant hack. If the file is executable but extensionless,
      # append a "magic" extension so it can be classified with other
      # languages that have shebang scripts.
      if File.extname(name).empty? && mode && (mode.to_i(8) & 05) == 05
@@ -329,7 +329,7 @@ module Linguist

    # Deprecated: Get primary extension
    #
-    # Defaults to the first extension but can be overriden
+    # Defaults to the first extension but can be overridden
    # in the languages.yml.
    #
    # The primary extension can not be nil. Tests should verify this.
--- a/lib/linguist/languages.yml
+++ b/lib/linguist/languages.yml
@@ -1,12 +1,10 @@
 # Defines all Languages known to GitHub.
 #
 # All languages have an associated lexer for syntax highlighting. It
-# defaults to name.downcase, which covers most cases. Make sure the
-# lexer exists in lexers.yml. This is a list of available in our
-# version of pygments.
+# defaults to name.downcase, which covers most cases.
 #
 # type              - Either data, programming, markup, or nil
-# lexer             - An explicit lexer String (defaults to name.downcase)
+# lexer             - An explicit lexer String (defaults to name)
 # aliases           - An Array of additional aliases (implicitly
 #                     includes name.downcase)
 # ace_mode          - A String name of Ace Mode (if available)
@@ -26,6 +24,11 @@
 #
 # Please keep this list alphabetized.

+ABAP:
+  type: programming
+  lexer: ABAP
+  primary_extension: .abap
+
 ASP:
  type: programming
  color: "#6a40fd"
@@ -109,6 +112,15 @@ AutoHotkey:
  - ahk
  primary_extension: .ahk

+Awk:
+  type: programming
+  lexer: Awk
+  primary_extension: .awk
+  extensions:
+  - .gawk
+  - .mawk
+  - .nawk
+
 Batchfile:
  type: programming
  group: Shell
@@ -186,6 +198,11 @@ C2hs Haskell:
  - c2hs
  primary_extension: .chs

+CLIPS:
+  type: programming
+  lexer: Text only
+  primary_extension: .clp
+
 CMake:
  primary_extension: .cmake
  extensions:
@@ -224,6 +241,8 @@ CoffeeScript:
  primary_extension: .coffee
  extensions:
  - ._coffee
+  - .cson
+  - .iced
  filenames:
  - Cakefile

@@ -246,6 +265,7 @@ Common Lisp:
  - lisp
  primary_extension: .lisp
  extensions:
+  - .asd
  - .lsp
  - .ny

@@ -285,6 +305,13 @@ D-ObjDump:
  lexer: d-objdump
  primary_extension: .d-objdump

+DOT:
+  type: programming
+  lexer: Text only
+  primary_extension: .dot
+  extensions:
+  - .gv
+
 Darcs Patch:
  search_term: dpatch
  aliases:
@@ -302,6 +329,7 @@ Delphi:
  color: "#b0ce4e"
  primary_extension: .pas
  extensions:
+  - .dfm
  - .lpr

 DCPU-16 ASM:
@@ -429,8 +457,7 @@ Forth:
  color: "#341708"
  lexer: Text only
  extensions:
-  - .forth
-  - .fth
+  - .4th

 GAS:
  type: programming
@@ -521,7 +548,9 @@ HTML+ERB:
  - erb
  primary_extension: .erb
  extensions:
+  - .erb.deface
  - .html.erb
+  - .html.erb.deface

 HTML+PHP:
  type: markup
@@ -536,6 +565,9 @@ Haml:
  group: HTML
  type: markup
  primary_extension: .haml
+  extensions:
+  - .haml.deface
+  - .html.haml.deface

 Handlebars:
  type: markup
@@ -598,8 +630,6 @@ Java:
  ace_mode: java
  color: "#b07219"
  primary_extension: .java
-  extensions:
-  - .pde

 Java Server Pages:
  group: Java
@@ -651,11 +681,6 @@ Lasso:
  ace_mode: lasso
  color: "#2584c3"
  primary_extension: .lasso
-  extensions:
-  - .inc
-  - .las
-  - .lasso9
-  - .ldml

 Less:
  type: markup
@@ -670,6 +695,17 @@ LilyPond:
  extensions:
  - .ily

+Literate CoffeeScript:
+  type: programming
+  group: CoffeeScript
+  lexer: Text only
+  ace_mode: markdown
+  wrap: true
+  search_term: litcoffee
+  aliases:
+  - litcoffee
+  primary_extension: .litcoffee
+
 Literate Haskell:
  type: programming
  group: Haskell
@@ -690,6 +726,14 @@ LiveScript:
  filenames:
  - Slakefile

+Logos:
+  type: programming
+  primary_extension: .xm
+  extensions:
+  - .x
+  - .xi
+  - .xmi
+
 Logtalk:
  type: programming
  primary_extension: .lgt
@@ -701,7 +745,6 @@ Lua:
  primary_extension: .lua
  extensions:
  - .nse
-  - .pd_lua

 Makefile:
  aliases:
@@ -746,6 +789,9 @@ Max:
  - maxmsp
  search_term: max/msp
  primary_extension: .mxt
+  extensions:
+  - .maxhelp
+  - .maxpat

 MiniD: # Legacy
  searchable: false
@@ -762,6 +808,11 @@ Mirah:
  - .mir
  - .mirah

+Monkey:
+  type: programming
+  lexer: Monkey
+  primary_extension: .monkey
+
 Moocode:
  lexer: MOOCode
  primary_extension: .moo
@@ -773,6 +824,9 @@ MoonScript:
 Myghty:
  primary_extension: .myt

+NSIS:
+  primary_extension: .nsi
+
 Nemerle:
  type: programming
  color: "#0d3c6e"
@@ -813,6 +867,7 @@ OCaml:
  color: "#3be133"
  primary_extension: .ml
  extensions:
+  - .eliomi
  - .mli
  - .mll
  - .mly
@@ -846,8 +901,6 @@ Omgrofl:
  primary_extension: .omgrofl
  color: "#cabbff"
  lexer: Text only
-  extensions:
-  - .omgrofl

 Opa:
  type: programming
@@ -858,6 +911,8 @@ OpenCL:
  group: C
  lexer: C
  primary_extension: .cl
+  extensions:
+  - .opencl

 OpenEdge ABL:
  type: programming
@@ -918,6 +973,20 @@ Perl:
  - .pod
  - .psgi

+Pike:
+  type: programming
+  color: "#066ab2"
+  lexer: C
+  primary_extension: .pike
+  extensions:
+  - .pmod
+
+PogoScript:
+  type: programming
+  color: "#d80074"
+  lexer: Text only
+  primary_extension: .pogo
+
 PowerShell:
  type: programming
  ace_mode: powershell
@@ -925,6 +994,12 @@ PowerShell:
  - posh
  primary_extension: .ps1

+Processing:
+  type: programming
+  lexer: Java
+  color: "#2779ab"
+  primary_extension: .pde
+  
 Prolog:
  type: programming
  color: "#74283c"
@@ -953,6 +1028,7 @@ Python:
  color: "#3581ba"
  primary_extension: .py
  extensions:
+  - .gyp
  - .pyw
  - .wsgi
  - .xpy
@@ -986,6 +1062,12 @@ Racket:
  - .rktd
  - .rktl

+Ragel in Ruby Host:
+  type: programming
+  lexer: Ragel in Ruby Host
+  color: "#ff9c2e"
+  primary_extension: .rl
+
 Raw token data:
  search_term: raw
  aliases:
@@ -1004,6 +1086,13 @@ Rebol:
 Redcode:
  primary_extension: .cw

+Rouge:
+  type: programming
+  lexer: Clojure
+  ace_mode: clojure
+  color: "#cc0088"
+  primary_extension: .rg
+
 Ruby:
  type: programming
  ace_mode: ruby
@@ -1037,7 +1126,6 @@ Ruby:
 Rust:
  type: programming
  color: "#dea584"
-  lexer: Text only
  primary_extension: .rs

 SCSS:
@@ -1097,6 +1185,8 @@ Shell:
  - bash
  - zsh
  primary_extension: .sh
+  extensions:
+  - .tmux

 Smalltalk:
  type: programming
@@ -1119,6 +1209,15 @@ SuperCollider:
  lexer: Text only
  primary_extension: .sc

+TOML:
+  type: data
+  primary_extension: .toml
+
+TXL:
+  type: programming
+  lexer: Text only
+  primary_extension: .txl
+
 Tcl:
  type: programming
  color: "#e4cc98"
@@ -1170,6 +1269,14 @@ Twig:
  lexer: HTML+Django/Jinja
  primary_extension: .twig

+TypeScript:
+  type: programming
+  color: "#31859c"
+  lexer: Text only
+  aliases:
+  - ts
+  primary_extension: .ts
+
 VHDL:
  type: programming
  lexer: vhdl
@@ -1188,6 +1295,10 @@ Verilog:
  lexer: verilog
  color: "#848bf3"
  primary_extension: .v
+  extensions:
+  - .sv
+  - .svh
+  - .vh

 VimL:
  type: programming
@@ -1217,21 +1328,29 @@ XML:
  aliases:
  - rss
  - xsd
-  - xsl
  - wsdl
  primary_extension: .xml
  extensions:
  - .ccxml
+  - .dita
+  - .ditamap
+  - .ditaval
  - .glade
  - .grxml
  - .kml
  - .mxml
  - .plist
+  - .pt
  - .rdf
  - .rss
  - .scxml
  - .svg
  - .ui
+  - .tmCommand
+  - .tmLanguage
+  - .tmPreferences
+  - .tmSnippet
+  - .tml
  - .vxml
  - .wsdl
  - .wxi
@@ -1241,12 +1360,19 @@ XML:
  - .xlf
  - .xliff
  - .xsd
-  - .xsl
  - .xul
+  - .zcml
  filenames:
  - .classpath
  - .project

+XProc:
+  type: programming
+  lexer: XML
+  primary_extension: .xpl
+  extensions:
+  - .xproc
+
 XQuery:
  type: programming
  color: "#2700e2"
@@ -1260,9 +1386,16 @@ XS:
  primary_extension: .xs

 XSLT:
-  type: markup
-  group: XML
+  type: programming
+  aliases:
+  - xsl
  primary_extension: .xslt
+  extensions:
+    - .xsl
+
+Xtend:
+  type: programming
+  primary_extension: .xtend

 YAML:
  type: markup
@@ -1279,6 +1412,19 @@ eC:
  extensions:
  - .eh

+edn:
+  type: data
+  lexer: Clojure
+  ace_mode: clojure
+  color: "#db5855"
+  primary_extension: .edn
+
+fish:
+  type: programming
+  group: Shell
+  lexer: Text only
+  primary_extension: .fish
+
 mupad:
  lexer: MuPAD
  primary_extension: .mu
--- a/lib/linguist/md5.rb
+++ b/lib/linguist/md5.rb
@@ -4,7 +4,7 @@ module Linguist
  module MD5
    # Public: Create deep nested digest of value object.
    #
-    # Useful for object comparsion.
+    # Useful for object comparison.
    #
    # obj - Object to digest.
    #
--- a/lib/linguist/repository.rb
+++ b/lib/linguist/repository.rb
@@ -67,8 +67,8 @@ module Linguist
      return if @computed_stats

      @enum.each do |blob|
-        # Skip binary file extensions
-        next if blob.binary_mime_type?
+        # Skip files that are likely binary
+        next if blob.likely_binary?

        # Skip vendored or generated blobs
        next if blob.vendored? || blob.generated? || blob.language.nil?
--- a/lib/linguist/samples.json
+++ b/lib/linguist/samples.json
--- a/lib/linguist/vendor.yml
+++ b/lib/linguist/vendor.yml
@@ -16,11 +16,15 @@
 #  https://github.com/joyent/node
 - ^deps/
 - ^tools/
+- (^|/)configure$
+- (^|/)configure.ac$
+- (^|/)config.guess$
+- (^|/)config.sub$

-# Node depedencies
+# Node dependencies
 - node_modules/

-# Vendored depedencies
+# Vendored dependencies
 - vendor/

 # Debian packaging
@@ -63,8 +67,16 @@
 # MathJax
 - (^|/)MathJax/

+# SyntaxHighlighter - http://alexgorbatchev.com/
+- (^|/)shBrush([^.]*)\.js$
+- (^|/)shCore\.js$
+- (^|/)shLegacy\.js$
+
 ## Python ##

+# django
+- (^|/)admin_media/
+
 # Fabric
 - ^fabfile\.py$

@@ -96,3 +108,6 @@

 # Samples folders
 - ^[Ss]amples/
+
+# Test fixtures
+- ^[Tt]est/fixtures/