KevinMidboe/linguist
1
0
Fork 0
mirror of https://github.com/KevinMidboe/linguist.git synced 2025-12-08 04:18:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into 2603-local

Browse Source
This commit is contained in:
Arfon Smith
2016-03-08 21:53:23 -06:00
parent 26a35ea43d 6b88c5ba86
commit 4637da8c32
265 changed files with 46637 additions and 985 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
lib/linguist/blob.rb Normal file
View File

@@ -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

45
lib/linguist/file_blob.rb
View File

@@ -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

53
lib/linguist/generated.rb
View File

@@ -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

395
lib/linguist/heuristics.rb
View File

@@ -56,7 +56,8 @@ module Linguist
# Internal: Check if this heuristic matches the candidate languages.
def matches?(filename)
@extensions.any? { |ext| filename.downcase.end_with?(ext) }
filename = filename.downcase
@extensions.any? { |ext| filename.end_with?(ext) }
end
# Internal: Perform the heuristic
@@ -65,7 +66,17 @@ 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 ".asc" do |data|
if /^(----[- ]BEGIN|ssh-(rsa|dss)) /.match(data)
Language["Public Key"]
elsif /^[=-]+(\s|\n)|{{[A-Za-z]/.match(data)
Language["AsciiDoc"]
elsif /^(\/\/.+|((import|export)\s+)?(function|int|float|char)\s+((room|repeatedly|on|game)_)?([A-Za-z]+[A-Za-z_0-9]+)\s*[;\(])/.match(data)
Language["AGS Script"]
end
end
disambiguate ".bb" do |data|
if /^\s*; /.match(data) || data.include?("End Function")
@@ -75,67 +86,9 @@ module Linguist
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 ".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):$/.match(data) || /std::\w+/.match(data))
Language["C++"]
end
end
disambiguate ".pl" do |data|
if /^(use v6|(my )?class|module)/.match(data)
Language["Perl6"]
elsif /use strict|use\s+v?5\./.match(data)
Language["Perl"]
elsif /^[^#]+:-/.match(data)
Language["Prolog"]
end
end
disambiguate ".pm", ".t" do |data|
if /^(use v6|(my )?class|module)/.match(data)
Language["Perl6"]
elsif /use strict|use\s+v?5\./.match(data)
Language["Perl"]
end
end
disambiguate ".ecl" do |data|
if /^[^#]+:-/.match(data)
Language["ECLiPSe"]
elsif data.include?(":=")
Language["ECL"]
end
end
disambiguate ".pro" do |data|
if /^[^#]+:-/.match(data)
Language["Prolog"]
elsif data.include?("last_client=")
Language["INI"]
elsif data.include?("HEADERS") && data.include?("SOURCES")
Language["QMake"]
elsif /^\s*function[ \w,]+$/.match(data)
Language["IDL"]
end
end
disambiguate ".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["Scilab"]
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
@@ -149,40 +102,50 @@ module Linguist
end
end
disambiguate ".php" do |data|
if data.include?("<?hh")
Language["Hack"]
elsif /<?[^h]/.match(data)
Language["PHP"]
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 ".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"]
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 ".asc" do |data|
if /^(----[- ]BEGIN|ssh-(rsa|dss)) /.match(data)
Language["Public Key"]
elsif /^[=-]+(\s|\n)|{{[A-Za-z]/.match(data)
Language["AsciiDoc"]
elsif /^(\/\/.+|((import|export)\s+)?(function|int|float|char)\s+((room|repeatedly|on|game)_)?([A-Za-z]+[A-Za-z_0-9]+)\s*[;\(])/.match(data)
Language["AGS Script"]
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 ".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"]
@@ -195,6 +158,47 @@ module Linguist
end
end
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"]
@@ -213,41 +217,131 @@ module Linguist
end
end
disambiguate ".gs" do |data|
Language["Gosu"] if /^uses java\./.match(data)
end
disambiguate ".ls" 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 ".lsp", ".lisp" do |data|
if /^\s*\((defun|in-package|defpackage) /i.match(data)
Language["Common Lisp"]
elsif /^\s*\(define /.match(data)
disambiguate ".mod" do |data|
if data.include?('<!ENTITY ')
Language["XML"]
elsif /MODULE\s\w+\s*;/i.match(data) || /^\s*END \w+;$/i.match(data)
Language["Modula-2"]
else
[Language["Linux Kernel Module"], Language["AMPL"]]
end
end
disambiguate ".ms" do |data|
if /^[.'][a-z][a-z](\s|$)/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 ".nl" do |data|
if /^(b|g)[0-9]+ /.match(data)
Language["NL"]
else
Language["NewLisp"]
end
end
disambiguate ".ts" do |data|
if data.include?("<TS ")
Language["XML"]
else
Language["TypeScript"]
disambiguate ".php" do |data|
if data.include?("<?hh")
Language["Hack"]
elsif /<?[^h]/.match(data)
Language["PHP"]
end
end
disambiguate ".fr" do |data|
if /^(: |also |new-device|previous )/.match(data)
Language["Forth"]
elsif /^\s*(import|module|package|data|type) /.match(data)
Language["Frege"]
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+$/.match(data)
Language["Pod"]
else
Language["Text"]
Language["Perl"]
end
end
disambiguate ".pro" do |data|
if /^[^#]+:-/.match(data)
Language["Prolog"]
elsif data.include?("last_client=")
Language["INI"]
elsif data.include?("HEADERS") && data.include?("SOURCES")
Language["QMake"]
elsif /^\s*function[ \w,]+$/.match(data)
Language["IDL"]
end
end
disambiguate ".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
@@ -267,95 +361,20 @@ module Linguist
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 ".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 ".mod" 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 ".ncl" do |data|
if data.include?("THE_TITLE")
Language["Text"]
end
end
disambiguate ".nl" 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 ".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 ".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 ".n" do |data|
if /^[.']/.match(data)
Language["Groff"]
elsif /^(module|namespace|using)\s/.match(data)
Language["Nemerle"]
end
end
disambiguate ".ms" do |data|
if /^[.'][a-z][a-z](\s|$)/i.match(data)
Language["Groff"]
elsif /((^|\s)move?[. ])|\.(include|globa?l)\s/.match(data)
Language["GAS"]
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 ".r" do |data|
if /\bRebol\b/i.match(data)
Language["Rebol"]
elsif data.include?("<-")
Language["R"]
Language["Scilab"]
end
end
end

253
lib/linguist/languages.yml Normal file → Executable file
View File

@@ -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
@@ -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
@@ -539,6 +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
@@ -567,6 +587,7 @@ CoffeeScript:
extensions:
- .coffee
- ._coffee
- .cake
- .cjsx
- .cson
- .iced
@@ -627,7 +648,7 @@ Common Lisp:
Component Pascal:
type: programming
color: "#b0ce4e"
color: "#B0CE4E"
extensions:
- .cp
- .cps
@@ -690,6 +711,7 @@ Cucumber:
aliases:
- gherkin
ace_mode: text
color: "#5B2063"
Cuda:
type: programming
@@ -698,6 +720,7 @@ Cuda:
- .cuh
tm_scope: source.cuda-c++
ace_mode: c_cpp
color: "#3A4E3A"
Cycript:
type: programming
@@ -719,7 +742,7 @@ Cython:
D:
type: programming
color: "#fcd46d"
color: "#ba595e"
extensions:
- .d
- .di
@@ -790,7 +813,6 @@ Dart:
Diff:
type: data
color: "#88dddd"
extensions:
- .diff
- .patch
@@ -884,6 +906,8 @@ Elixir:
ace_mode: elixir
filenames:
- mix.lock
interpreters:
- elixir
Elm:
type: programming
@@ -926,6 +950,8 @@ Erlang:
- .es
- .escript
- .hrl
- .xrl
- .yrl
filenames:
- rebar.config
- rebar.config.lock
@@ -1011,6 +1037,7 @@ Formatted:
type: data
extensions:
- .for
- .eam.fs
tm_scope: none
ace_mode: text
@@ -1028,6 +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"
@@ -1273,6 +1310,8 @@ Groovy:
- .gvy
interpreters:
- groovy
filenames:
- Jenkinsfile
Groovy Server Pages:
type: programming
@@ -1293,6 +1332,16 @@ HCL:
ace_mode: ruby
tm_scope: source.ruby
HLSL:
type: programming
extensions:
- .fx
- .fxh
- .hlsl
- .hlsli
ace_mode: text
tm_scope: none
HTML:
type: markup
tm_scope: text.html.basic
@@ -1317,11 +1366,22 @@ 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
@@ -1331,7 +1391,7 @@ HTML+ERB:
extensions:
- .erb
- .erb.deface
ace_mode: html_ruby
ace_mode: text
HTML+PHP:
type: markup
@@ -1355,6 +1415,7 @@ Hack:
- .hh
- .php
tm_scope: text.html.php
color: "#878787"
Haml:
group: HTML
@@ -1363,10 +1424,12 @@ Haml:
- .haml
- .haml.deface
ace_mode: haml
color: "#ECE2A9"
Handlebars:
type: markup
color: "#01a9d6"
group: HTML
aliases:
- hbs
- htmlbars
@@ -1474,7 +1537,7 @@ Inform 7:
extensions:
- .ni
- .i7x
tm_scope: source.Inform7
tm_scope: source.inform7
aliases:
- i7
- inform7
@@ -1545,7 +1608,9 @@ JSON:
searchable: false
extensions:
- .json
- .geojson
- .lock
- .topojson
filenames:
- .jshintrc
- composer.lock
@@ -1573,12 +1638,20 @@ 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:
@@ -1623,10 +1696,10 @@ JavaScript:
- .gs
- .jake
- .jsb
- .jscad
- .jsfl
- .jsm
- .jss
- .jsx
- .njs
- .pac
- .sjs
@@ -1658,6 +1731,18 @@ 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"
@@ -1670,6 +1755,7 @@ KiCad:
type: programming
extensions:
- .sch
- .brd
- .kicad_pcb
tm_scope: none
ace_mode: text
@@ -1705,6 +1791,7 @@ LLVM:
extensions:
- .ll
ace_mode: text
color: "#185619"
LOLCODE:
type: programming
@@ -1719,6 +1806,7 @@ LSL:
ace_mode: lsl
extensions:
- .lsl
- .lslp
interpreters:
- lsl
color: '#3d9970'
@@ -1767,6 +1855,7 @@ Less:
- .less
tm_scope: source.css.less
ace_mode: less
color: "#A1D9A1"
Lex:
type: programming
@@ -1935,6 +2024,14 @@ M4Sugar:
filenames:
- configure.ac
tm_scope: none
MAXScript:
type: programming
color: "#00a6a6"
extensions:
- .ms
- .mcr
tm_scope: source.maxscript
ace_mode: text
MTML:
@@ -1969,6 +2066,8 @@ Makefile:
- GNUmakefile
- Kbuild
- Makefile
- Makefile.am
- Makefile.in
- Makefile.inc
- makefile
interpreters:
@@ -1995,6 +2094,7 @@ Markdown:
- .mkdown
- .ron
tm_scope: source.gfm
color: "#083FA1"
Mask:
type: markup
@@ -2011,6 +2111,7 @@ Mathematica:
- .cdf
- .m
- .ma
- .mt
- .nb
- .nbp
- .wl
@@ -2022,6 +2123,8 @@ Mathematica:
Matlab:
type: programming
color: "#bb92ac"
aliases:
- octave
extensions:
- .matlab
- .m
@@ -2055,6 +2158,7 @@ MediaWiki:
wrap: true
extensions:
- .mediawiki
- .wiki
tm_scope: text.html.mediawiki
ace_mode: text
@@ -2070,6 +2174,14 @@ 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
@@ -2145,7 +2257,7 @@ Myghty:
NCL:
type: programming
color: #28431f
color: "#28431f"
extensions:
- .ncl
tm_scope: source.ncl
@@ -2222,6 +2334,7 @@ Nginx:
aliases:
- nginx configuration file
ace_mode: text
color: "#9469E9"
Nimrod:
type: programming
@@ -2280,6 +2393,7 @@ NumPy:
- .numsc
tm_scope: none
ace_mode: text
color: "#9C8AF9"
OCaml:
type: programming
@@ -2448,6 +2562,7 @@ PHP:
- .php3
- .php4
- .php5
- .phps
- .phpt
filenames:
- Phakefile
@@ -2461,8 +2576,10 @@ PLSQL:
type: programming
ace_mode: sql
tm_scope: source.plsql.oracle
color: "#dad8d8"
extensions:
- .pls
- .pck
- .pkb
- .pks
- .plb
@@ -2477,6 +2594,16 @@ PLpgSQL:
extensions:
- .sql
POV-Ray SDL:
type: programming
aliases:
- pov-ray
- povray
extensions:
- .pov
- .inc
ace_mode: text
Pan:
type: programming
color: '#cc0000'
@@ -2527,7 +2654,7 @@ Parrot Internal Representation:
Pascal:
type: programming
color: "#b0ce4e"
color: "#E3F171"
extensions:
- .pas
- .dfm
@@ -2576,9 +2703,16 @@ 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:
@@ -2623,6 +2757,13 @@ PogoScript:
tm_scope: source.pogoscript
ace_mode: text
Pony:
type: programming
extensions:
- .pony
tm_scope: source.pony
ace_mode: text
PostScript:
type: markup
extensions:
@@ -2657,8 +2798,10 @@ Prolog:
- .pl
- .pro
- .prolog
- .yap
interpreters:
- swipl
- yap
tm_scope: source.prolog
ace_mode: prolog
@@ -2728,6 +2871,7 @@ Python:
color: "#3572A5"
extensions:
- .py
- .bzl
- .cgi
- .fcgi
- .gyp
@@ -2736,10 +2880,12 @@ Python:
- .pyp
- .pyt
- .pyw
- .rpy
- .tac
- .wsgi
- .xpy
filenames:
- BUCK
- BUILD
- SConscript
- SConstruct
@@ -2781,7 +2927,7 @@ QMake:
R:
type: programming
color: "#198ce7"
color: "#198CE7"
aliases:
- R
- Rscript
@@ -2811,6 +2957,7 @@ RDoc:
extensions:
- .rdoc
tm_scope: text.rdoc
color: "#8E84BF"
REALbasic:
type: programming
@@ -2906,6 +3053,16 @@ 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:
@@ -2991,6 +3148,7 @@ Rust:
color: "#dea584"
extensions:
- .rs
- .rs.in
ace_mode: rust
SAS:
@@ -3008,6 +3166,7 @@ SCSS:
ace_mode: scss
extensions:
- .scss
color: "#CF649A"
SMT:
type: programming
@@ -3110,11 +3269,12 @@ Sass:
extensions:
- .sass
ace_mode: sass
color: "#CF649A"
Scala:
type: programming
ace_mode: scala
color: "#7dd3b0"
color: "#DC322F"
extensions:
- .scala
- .sbt
@@ -3167,6 +3327,7 @@ Shell:
color: "#89e051"
aliases:
- sh
- shell-script
- bash
- zsh
extensions:
@@ -3177,6 +3338,7 @@ Shell:
- .command
- .fcgi
- .ksh
- .sh.in
- .tmux
- .tool
- .zsh
@@ -3266,6 +3428,14 @@ 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"
@@ -3303,9 +3473,12 @@ SuperCollider:
type: programming
color: "#46390b"
extensions:
- .scd
- .sc
tm_scope: none
- .scd
interpreters:
- sclang
- scsynth
tm_scope: source.supercollider
ace_mode: text
Swift:
@@ -3448,6 +3621,7 @@ TypeScript:
- ts
extensions:
- .ts
- .tsx
tm_scope: source.ts
ace_mode: typescript
@@ -3463,7 +3637,6 @@ Unified Parallel C:
Unity3D Asset:
type: data
ace_mode: yaml
color: "#ab69a1"
extensions:
- .anim
- .asset
@@ -3473,6 +3646,13 @@ Unity3D Asset:
- .unity
tm_scope: source.yaml
Uno:
type: programming
extensions:
- .uno
ace_mode: csharp
tm_scope: source.cs
UnrealScript:
type: programming
color: "#a54c4d"
@@ -3481,6 +3661,17 @@ 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
@@ -3586,6 +3777,16 @@ 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"
@@ -3608,6 +3809,7 @@ XML:
- .ccxml
- .clixml
- .cproject
- .csl
- .csproj
- .ct
- .dita
@@ -3623,6 +3825,7 @@ XML:
- .iml
- .ivy
- .jelly
- .jsproj
- .kml
- .launch
- .mdpolicy
@@ -3653,8 +3856,10 @@ XML:
- .tmSnippet
- .tmTheme
- .ts
- .tsx
- .ui
- .urdf
- .ux
- .vbproj
- .vcxproj
- .vxml
@@ -3671,6 +3876,7 @@ XML:
- .xliff
- .xmi
- .xml.dist
- .xproj
- .xsd
- .xul
- .zcml
@@ -3727,6 +3933,7 @@ XSLT:
- .xsl
tm_scope: text.xml.xsl
ace_mode: xml
color: "#EB8CEB"
Xojo:
type: programming
@@ -3760,6 +3967,13 @@ YAML:
- .yaml-tmlanguage
ace_mode: yaml
YANG:
type: data
extensions:
- .yang
tm_scope: source.yang
ace_mode: text
Yacc:
type: programming
extensions:
@@ -3768,6 +3982,7 @@ Yacc:
- .yy
tm_scope: source.bison
ace_mode: text
color: "#4B6C4B"
Zephir:
type: programming
@@ -3807,7 +4022,6 @@ eC:
edn:
type: data
ace_mode: clojure
color: "#db5855"
extensions:
- .edn
tm_scope: source.clojure
@@ -3849,7 +4063,10 @@ reStructuredText:
extensions:
- .rst
- .rest
- .rest.txt
- .rst.txt
ace_mode: text
color: "#B3BCBC"
wisp:
type: programming

38
lib/linguist/lazy_blob.rb
View File

@@ -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!

12
lib/linguist/repository.rb
View File

@@ -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

17
lib/linguist/strategy/modeline.rb
View File

@@ -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+)\s?.*:/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

4
lib/linguist/tokenizer.rb
View File

@@ -86,13 +86,13 @@ 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

42
lib/linguist/vendor.yml
View File

@@ -85,6 +85,9 @@
# Haxelib projects often contain a neko bytecode file named run.n
- run.n$
# Bootstrap Datepicker
- bootstrap-datepicker/
## Commonly Bundled JavaScript frameworks ##
# jQuery
@@ -95,6 +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$
@@ -129,7 +160,7 @@
- (^|/)Chart\.js$
# Codemirror
- (^|/)[Cc]ode[Mm]irror/(lib|mode|theme|addon|keymap|demo)
- (^|/)[Cc]ode[Mm]irror/(\d+\.\d+/)?(lib|mode|theme|addon|keymap|demo)
# SyntaxHighlighter - http://alexgorbatchev.com/
- (^|/)shBrush([^.]*)\.js$
@@ -171,6 +202,11 @@
## Obj-C ##
# Xcode
- \.xctemplate/
- \.imageset/
# Carthage
- ^Carthage/
@@ -237,6 +273,7 @@
# Test fixtures
- ^[Tt]ests?/fixtures/
- ^[Ss]pecs?/fixtures/
# PhoneGap/Cordova
- (^|/)cordova([^.]*)\.js$
@@ -272,3 +309,6 @@
# Android Google APIs
- (^|/)\.google_apis/
# Jenkins Pipeline
- ^Jenkinsfile$

2
lib/linguist/version.rb
View File

@@ -1,3 +1,3 @@
module Linguist
VERSION = "4.5.14"
VERSION = "4.7.6"
end