KevinMidboe/linguist
1
0
Fork 0
mirror of https://github.com/KevinMidboe/linguist.git synced 2025-10-29 17:50:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into blob_in_memory

Browse Source
This commit is contained in:
Paul Chaignon
2015-08-19 18:02:31 +02:00
parent eca10056a8 c7bab11ebe
commit c777f2d388
115 changed files with 6625 additions and 163 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
lib/linguist/documentation.yml
View File

@@ -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/

107
lib/linguist/heuristics.rb
View File

@@ -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
@@ -38,22 +41,22 @@ 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)
@extensions.any? { |ext| filename.downcase.end_with?(ext) }
end
# Internal: Perform the heuristic
@@ -64,7 +67,7 @@ module Linguist
# Common heuristics
ObjectiveCRegex = /^[ \t]*@(interface|class|protocol|property|end|synchronised|selector|implementation)\b/
disambiguate "BitBake", "BlitzBasic" do |data|
disambiguate ".bb" do |data|
if /^\s*; /.match(data) || data.include?("End Function")
Language["BlitzBasic"]
elsif /^\s*(# |include|require)\b/.match(data)
@@ -72,7 +75,7 @@ module Linguist
end
end
disambiguate "C#", "Smalltalk" do |data|
disambiguate ".cs" do |data|
if /![\w\s]+methodsFor: /.match(data)
Language["Smalltalk"]
elsif /^\s*namespace\s*[\w\.]+\s*{/.match(data) || /^\s*\/\//.match(data)
@@ -80,7 +83,7 @@ module Linguist
end
end
disambiguate "Objective-C", "C++", "C" do |data|
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) ||
@@ -89,17 +92,25 @@ module Linguist
end
end
disambiguate "Perl", "Perl6", "Prolog" do |data|
if data.include?("use v6")
disambiguate ".pl" do |data|
if /^(use v6|(my )?class|module)/.match(data)
Language["Perl6"]
elsif data.match(/use strict|use\s+v?5\./)
elsif /use strict|use\s+v?5\./.match(data)
Language["Perl"]
elsif /^[^#]+:-/.match(data)
Language["Prolog"]
end
end
disambiguate "ECL", "Prolog" do |data|
disambiguate ".pm" 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["Prolog"]
elsif data.include?(":=")
@@ -107,7 +118,7 @@ module Linguist
end
end
disambiguate "IDL", "Prolog", "INI", "QMake" do |data|
disambiguate ".pro" do |data|
if /^[^#]+:-/.match(data)
Language["Prolog"]
elsif data.include?("last_client=")
@@ -119,7 +130,7 @@ module Linguist
end
end
disambiguate "GAP", "Scilab" do |data|
disambiguate ".tst" do |data|
if (data.include?("gap> "))
Language["GAP"]
# Heads up - we don't usually write heuristics like this (with no regex match)
@@ -128,7 +139,7 @@ module Linguist
end
end
disambiguate "Common Lisp", "OpenCL", "Cool" do |data|
disambiguate ".cl" do |data|
if /^\s*\((defun|in-package|defpackage) /i.match(data)
Language["Common Lisp"]
elsif /^class/x.match(data)
@@ -138,7 +149,7 @@ module Linguist
end
end
disambiguate "Hack", "PHP" do |data|
disambiguate ".php" do |data|
if data.include?("<?hh")
Language["Hack"]
elsif /<?[^h]/.match(data)
@@ -146,7 +157,7 @@ module Linguist
end
end
disambiguate "Scala", "SuperCollider" do |data|
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)
@@ -154,7 +165,7 @@ module Linguist
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)
@@ -164,7 +175,7 @@ module Linguist
end
end
disambiguate "FORTRAN", "Forth", "Formatted" do |data|
disambiguate ".for", ".f" do |data|
if /^: /.match(data)
Language["Forth"]
elsif /^([c*][^a-z]| (subroutine|program)\s|\s*!)/i.match(data)
@@ -172,7 +183,7 @@ module Linguist
end
end
disambiguate "F#", "Forth", "GLSL", "Filterscript" do |data|
disambiguate ".fs" do |data|
if /^(: |new-device)/.match(data)
Language["Forth"]
elsif /^\s*(#light|import|let|module|namespace|open|type)/.match(data)
@@ -184,7 +195,7 @@ module Linguist
end
end
disambiguate "Limbo", "M", "MUF", "Mathematica", "Matlab", "Mercury", "Objective-C" do |data|
disambiguate ".m" do |data|
if ObjectiveCRegex.match(data)
Language["Objective-C"]
elsif data.include?(":- module")
@@ -202,11 +213,11 @@ module Linguist
end
end
disambiguate "Gosu", "JavaScript" do |data|
disambiguate ".gs" do |data|
Language["Gosu"] if /^uses java\./.match(data)
end
disambiguate "LoomScript", "LiveScript" do |data|
disambiguate ".ls" do |data|
if /^\s*package\s*[\w\.\/\*\s]*\s*{/.match(data)
Language["LoomScript"]
else
@@ -214,7 +225,7 @@ module Linguist
end
end
disambiguate "Common Lisp", "NewLisp" do |data|
disambiguate ".lsp", ".lisp" do |data|
if /^\s*\((defun|in-package|defpackage) /i.match(data)
Language["Common Lisp"]
elsif /^\s*\(define /.match(data)
@@ -222,7 +233,7 @@ module Linguist
end
end
disambiguate "TypeScript", "XML" do |data|
disambiguate ".ts" do |data|
if data.include?("<TS ")
Language["XML"]
else
@@ -230,7 +241,7 @@ module Linguist
end
end
disambiguate "Frege", "Forth", "Text" do |data|
disambiguate ".fr" do |data|
if /^(: |also |new-device|previous )/.match(data)
Language["Forth"]
elsif /^\s*(import|module|package|data|type) /.match(data)
@@ -240,7 +251,7 @@ module Linguist
end
end
disambiguate "PLSQL", "SQLPL", "PLpgSQL", "SQL" do |data|
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"]
@@ -256,7 +267,7 @@ module Linguist
end
end
disambiguate "D", "DTrace", "Makefile" do |data|
disambiguate ".d" do |data|
if /^module /.match(data)
Language["D"]
elsif /^((dtrace:::)?BEGIN|provider |#pragma (D (option|attributes)|ident)\s)/.match(data)
@@ -266,7 +277,7 @@ module Linguist
end
end
disambiguate "OCaml", "Standard ML" do |data|
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)
@@ -274,7 +285,7 @@ module Linguist
end
end
disambiguate "XML", "Modula-2", "Linux Kernel Module", "AMPL" do |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)
@@ -284,7 +295,13 @@ module Linguist
end
end
disambiguate "NL", "NewLisp" do |data|
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
@@ -292,7 +309,7 @@ module Linguist
end
end
disambiguate "Rust", "RenderScript" do |data|
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)
@@ -300,7 +317,7 @@ module Linguist
end
end
disambiguate "Common Lisp", "Lex", "Groff", "PicoLisp" do |data|
disambiguate ".l" do |data|
if /\(def(un|macro)\s/.match(data)
Language["Common Lisp"]
elsif /^(%[%{}]xs|<.*>)/.match(data)
@@ -312,7 +329,7 @@ module Linguist
end
end
disambiguate "Groff", "Nemerle" do |data|
disambiguate ".n" do |data|
if /^[.']/.match(data)
Language["Groff"]
elsif /^(module|namespace|using)\s/.match(data)
@@ -320,12 +337,26 @@ module Linguist
end
end
disambiguate "GAS", "Groff" do |data|
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"]
end
end
end
end

9
lib/linguist/language.rb
View File

@@ -150,7 +150,8 @@ 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 +165,8 @@ 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 +242,8 @@ 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

74
lib/linguist/languages.yml
View File

@@ -214,7 +214,7 @@ AspectJ:
color: "#a957b0"
extensions:
- .aj
tm_scope: none
tm_scope: source.aspectj
ace_mode: text
Assembly:
@@ -501,6 +501,13 @@ Chapel:
- .chpl
ace_mode: text
Charity:
type: programming
extensions:
- .ch
tm_scope: none
ace_mode: text
ChucK:
type: programming
extensions:
@@ -609,6 +616,7 @@ Common Lisp:
- .lsp
- .ny
- .podsl
- .sexp
interpreters:
- lisp
- sbcl
@@ -743,6 +751,14 @@ 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:
@@ -1261,6 +1277,14 @@ Groovy Server Pages:
tm_scope: text.html.jsp
ace_mode: jsp
HCL:
type: programming
extensions:
- .hcl
- .tf
ace_mode: ruby
tm_scope: source.ruby
HTML:
type: markup
tm_scope: text.html.basic
@@ -1349,7 +1373,7 @@ Harbour:
color: "#0e60e3"
extensions:
- .hb
tm_scope: none
tm_scope: source.harbour
ace_mode: text
Haskell:
@@ -1695,8 +1719,8 @@ LabVIEW:
type: programming
extensions:
- .lvproj
tm_scope: none
ace_mode: text
tm_scope: text.xml
ace_mode: xml
Lasso:
type: programming
@@ -1918,6 +1942,7 @@ Makefile:
- GNUmakefile
- Kbuild
- Makefile
- Makefile.inc
- makefile
interpreters:
- make
@@ -2091,6 +2116,14 @@ Myghty:
tm_scope: none
ace_mode: text
NCL:
type: programming
color: #28431f
extensions:
- .ncl
tm_scope: source.ncl
ace_mode: text
NL:
type: data
extensions:
@@ -2630,7 +2663,7 @@ Public Key:
Puppet:
type: programming
color: "#332A77"
color: "#302B6D"
extensions:
- .pp
filenames:
@@ -2659,7 +2692,7 @@ PureScript:
color: "#1D222D"
extensions:
- .purs
tm_scope: source.haskell
tm_scope: source.purescript
ace_mode: haskell
Python:
@@ -2910,6 +2943,7 @@ Ruby:
- .pryrc
- Appraisals
- Berksfile
- Brewfile
- Buildfile
- Deliverfile
- Fastfile
@@ -3158,6 +3192,7 @@ Slim:
color: "#ff8f77"
extensions:
- .slim
tm_scope: text.slim
ace_mode: text
Smali:
@@ -3336,6 +3371,7 @@ Text:
extensions:
- .txt
- .fr
- .ncl
tm_scope: none
ace_mode: text
@@ -3500,6 +3536,14 @@ 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"
@@ -3612,6 +3656,14 @@ XML:
- Web.config
- packages.config
XPages:
type: programming
extensions:
- .xsp-config
- .xsp.metadata
tm_scope: none
ace_mode: xml
XProc:
type: programming
extensions:
@@ -3676,7 +3728,9 @@ YAML:
- .yml
- .reek
- .rviz
- .syntax
- .yaml
- .yaml-tmlanguage
ace_mode: yaml
Yacc:
@@ -3781,7 +3835,13 @@ wisp:
xBase:
type: programming
color: "#403a40"
aliases:
- advpl
- clipper
- foxpro
extensions:
- .prg
tm_scope: none
- .ch
- .prw
tm_scope: source.harbour
ace_mode: text

2
lib/linguist/strategy/modeline.rb
View File

@@ -2,7 +2,7 @@ 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
VimModeline = /vim:\s*set.*\s(?:ft|filetype)=(\w+)\s?.*:/i
# Public: Detects language based on Vim and Emacs modelines
#

2
lib/linguist/tokenizer.rb
View File

@@ -96,7 +96,7 @@ module Linguist
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<>][^<>]*>/)

19
lib/linguist/vendor.yml
View File

@@ -122,7 +122,7 @@
- (^|/)Chart\.js$
# Codemirror
- (^|/)[Cc]ode[Mm]irror/(lib|mode|theme|addon|keymap)
- (^|/)[Cc]ode[Mm]irror/(lib|mode|theme|addon|keymap|demo)
# SyntaxHighlighter - http://alexgorbatchev.com/
- (^|/)shBrush([^.]*)\.js$
@@ -179,6 +179,10 @@
# Fabric
- Fabric.framework/
# git config files
- gitattributes$
- gitignore$
- gitmodules$
## Groovy ##
@@ -224,19 +228,6 @@
# 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/

2
lib/linguist/version.rb
View File

@@ -1,3 +1,3 @@
module Linguist
VERSION = "4.5.8"
VERSION = "4.5.14"
end