Skip to content

Commit

Permalink
Showing 9 changed files with 1,417 additions and 40 deletions.
1 change: 1 addition & 0 deletions lib/ruby/truffle/mri/unicode_normalize.rb
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
require_relative '../stdlib/unicode_normalize'
1 change: 1 addition & 0 deletions lib/ruby/truffle/mri/unicode_normalize/normalize.rb
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
require_relative '../../stdlib/unicode_normalize/' + File.basename(__FILE__)
1 change: 1 addition & 0 deletions lib/ruby/truffle/mri/unicode_normalize/tables.rb
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
require_relative '../../stdlib/unicode_normalize/' + File.basename(__FILE__)
77 changes: 77 additions & 0 deletions lib/ruby/truffle/stdlib/unicode_normalize.rb
Original file line number Diff line number Diff line change
@@ -0,0 +1,77 @@
# coding: utf-8

# Copyright Ayumu Nojima (野島 歩) and Martin J. Dürst (duerst@it.aoyama.ac.jp)

# additions to class String for Unicode normalization
class String
# === Unicode Normalization
#
# :call-seq:
# str.unicode_normalize(form=:nfc)
#
# Returns a normalized form of +str+, using Unicode normalizations
# NFC, NFD, NFKC, or NFKD. The normalization form used is determined
# by +form+, which is any of the four values :nfc, :nfd, :nfkc, or :nfkd.
# The default is :nfc.
#
# If the string is not in a Unicode Encoding, then an Exception is raised.
# In this context, 'Unicode Encoding' means any of UTF-8, UTF-16BE/LE,
# and UTF-32BE/LE, as well as GB18030, UCS_2BE, and UCS_4BE. Anything
# else than UTF-8 is implemented by converting to UTF-8,
# which makes it slower than UTF-8.
#
# _Examples_
#
# "a\u0300".unicode_normalize #=> 'à' (same as "\u00E0")
# "a\u0300".unicode_normalize(:nfc) #=> 'à' (same as "\u00E0")
# "\u00E0".unicode_normalize(:nfd) #=> 'à' (same as "a\u0300")
# "\xE0".force_encoding('ISO-8859-1').unicode_normalize(:nfd)
# #=> Encoding::CompatibilityError raised
#
def unicode_normalize(form = :nfc)
require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize
## The following line can be uncommented to avoid repeated checking for
## UnicodeNormalize. However, tests didn't show any noticeable speedup
## when doing this. This comment also applies to the commented out lines
## in String#unicode_normalize! and String#unicode_normalized?.
# String.send(:define_method, :unicode_normalize, ->(form = :nfc) { UnicodeNormalize.normalize(self, form) } )
UnicodeNormalize.normalize(self, form)
end

# :call-seq:
# str.unicode_normalize!(form=:nfc)
#
# Destructive version of String#unicode_normalize, doing Unicode
# normalization in place.
#
def unicode_normalize!(form = :nfc)
require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize
# String.send(:define_method, :unicode_normalize!, ->(form = :nfc) { replace(unicode_normalize(form)) } )
replace(unicode_normalize(form))
end

# :call-seq:
# str.unicode_normalized?(form=:nfc)
#
# Checks whether +str+ is in Unicode normalization form +form+,
# which is any of the four values :nfc, :nfd, :nfkc, or :nfkd.
# The default is :nfc.
#
# If the string is not in a Unicode Encoding, then an Exception is raised.
# For details, see String#unicode_normalize.
#
# _Examples_
#
# "a\u0300".unicode_normalized? #=> false
# "a\u0300".unicode_normalized?(:nfd) #=> true
# "\u00E0".unicode_normalized? #=> true
# "\u00E0".unicode_normalized?(:nfd) #=> false
# "\xE0".force_encoding('ISO-8859-1').unicode_normalized?
# #=> Encoding::CompatibilityError raised
#
def unicode_normalized?(form = :nfc)
require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize
# String.send(:define_method, :unicode_normalized?, ->(form = :nfc) { UnicodeNormalize.normalized?(self, form) } )
UnicodeNormalize.normalized?(self, form)
end
end
168 changes: 168 additions & 0 deletions lib/ruby/truffle/stdlib/unicode_normalize/normalize.rb
Original file line number Diff line number Diff line change
@@ -0,0 +1,168 @@
# coding: utf-8

# Copyright Ayumu Nojima (野島 歩) and Martin J. Dürst (duerst@it.aoyama.ac.jp)

require 'unicode_normalize/tables.rb'


module UnicodeNormalize
## Constant for max hash capacity to avoid DoS attack
MAX_HASH_LENGTH = 18000 # enough for all test cases, otherwise tests get slow

## Regular Expressions and Hash Constants
REGEXP_D = Regexp.compile(REGEXP_D_STRING, Regexp::EXTENDED)
REGEXP_C = Regexp.compile(REGEXP_C_STRING, Regexp::EXTENDED)
REGEXP_K = Regexp.compile(REGEXP_K_STRING, Regexp::EXTENDED)
NF_HASH_D = Hash.new do |hash, key|
hash.shift if hash.length>MAX_HASH_LENGTH # prevent DoS attack
hash[key] = nfd_one(key)
end
NF_HASH_C = Hash.new do |hash, key|
hash.shift if hash.length>MAX_HASH_LENGTH # prevent DoS attack
hash[key] = nfc_one(key)
end
NF_HASH_K = Hash.new do |hash, key|
hash.shift if hash.length>MAX_HASH_LENGTH # prevent DoS attack
hash[key] = nfkd_one(key)
end

## Constants For Hangul
# for details such as the meaning of the identifiers below, please see
# http://www.unicode.org/versions/Unicode7.0.0/ch03.pdf, pp. 144/145
SBASE = 0xAC00
LBASE = 0x1100
VBASE = 0x1161
TBASE = 0x11A7
LCOUNT = 19
VCOUNT = 21
TCOUNT = 28
NCOUNT = VCOUNT * TCOUNT
SCOUNT = LCOUNT * NCOUNT

# Unicode-based encodings (except UTF-8)
UNICODE_ENCODINGS = [Encoding::UTF_16BE, Encoding::UTF_16LE, Encoding::UTF_32BE, Encoding::UTF_32LE,
Encoding::GB18030, Encoding::UCS_2BE, Encoding::UCS_4BE]

## Hangul Algorithm
def self.hangul_decomp_one(target)
syllable_index = target.ord - SBASE
return target if syllable_index < 0 || syllable_index >= SCOUNT
l = LBASE + syllable_index / NCOUNT
v = VBASE + (syllable_index % NCOUNT) / TCOUNT
t = TBASE + syllable_index % TCOUNT
(t==TBASE ? [l, v] : [l, v, t]).pack('U*') + target[1..-1]
end

def self.hangul_comp_one(string)
length = string.length
if length>1 and 0 <= (lead =string[0].ord-LBASE) and lead < LCOUNT and
0 <= (vowel=string[1].ord-VBASE) and vowel < VCOUNT
lead_vowel = SBASE + (lead * VCOUNT + vowel) * TCOUNT
if length>2 and 0 <= (trail=string[2].ord-TBASE) and trail < TCOUNT
(lead_vowel + trail).chr(Encoding::UTF_8) + string[3..-1]
else
lead_vowel.chr(Encoding::UTF_8) + string[2..-1]
end
else
string
end
end

## Canonical Ordering
def self.canonical_ordering_one(string)
sorting = string.each_char.collect { |c| [c, CLASS_TABLE[c]] }
(sorting.length-2).downto(0) do |i| # almost, but not exactly bubble sort
(0..i).each do |j|
later_class = sorting[j+1].last
if 0<later_class and later_class<sorting[j].last
sorting[j], sorting[j+1] = sorting[j+1], sorting[j]
end
end
end
return sorting.collect(&:first).join('')
end

## Normalization Forms for Patterns (not whole Strings)
def self.nfd_one(string)
string = string.chars.map! {|c| DECOMPOSITION_TABLE[c] || c}.join('')
canonical_ordering_one(hangul_decomp_one(string))
end

def self.nfkd_one(string)
string.chars.map! {|c| KOMPATIBLE_TABLE[c] || c}.join('')
end

def self.nfc_one(string)
nfd_string = nfd_one string
start = nfd_string[0]
last_class = CLASS_TABLE[start]-1
accents = ''
nfd_string[1..-1].each_char do |accent|
accent_class = CLASS_TABLE[accent]
if last_class<accent_class and composite = COMPOSITION_TABLE[start+accent]
start = composite
else
accents << accent
last_class = accent_class
end
end
hangul_comp_one(start+accents)
end

def self.normalize(string, form = :nfc)
encoding = string.encoding
case encoding
when Encoding::UTF_8
case form
when :nfc then
string.gsub REGEXP_C, NF_HASH_C
when :nfd then
string.gsub REGEXP_D, NF_HASH_D
when :nfkc then
string.gsub(REGEXP_K, NF_HASH_K).gsub REGEXP_C, NF_HASH_C
when :nfkd then
string.gsub(REGEXP_K, NF_HASH_K).gsub REGEXP_D, NF_HASH_D
else
raise ArgumentError, "Invalid normalization form #{form}."
end
when Encoding::US_ASCII
string
when *UNICODE_ENCODINGS
normalize(string.encode(Encoding::UTF_8), form).encode(encoding)
else
raise Encoding::CompatibilityError, "Unicode Normalization not appropriate for #{encoding}"
end
end

def self.normalized?(string, form = :nfc)
encoding = string.encoding
case encoding
when Encoding::UTF_8
case form
when :nfc then
string.scan REGEXP_C do |match|
return false if NF_HASH_C[match] != match
end
true
when :nfd then
string.scan REGEXP_D do |match|
return false if NF_HASH_D[match] != match
end
true
when :nfkc then
normalized?(string, :nfc) and string !~ REGEXP_K
when :nfkd then
normalized?(string, :nfd) and string !~ REGEXP_K
else
raise ArgumentError, "Invalid normalization form #{form}."
end
when Encoding::US_ASCII
true
when *UNICODE_ENCODINGS
normalized? string.encode(Encoding::UTF_8), form
else
raise Encoding::CompatibilityError, "Unicode Normalization not appropriate for #{encoding}"
end
end

end # module
1,163 changes: 1,163 additions & 0 deletions lib/ruby/truffle/stdlib/unicode_normalize/tables.rb

Large diffs are not rendered by default.

17 changes: 0 additions & 17 deletions spec/truffle/tags/core/string/unicode_normalize_tags.txt

This file was deleted.

23 changes: 0 additions & 23 deletions spec/truffle/tags/core/string/unicode_normalized_tags.txt

This file was deleted.

6 changes: 6 additions & 0 deletions truffle/src/main/ruby/core/post.rb
Original file line number Diff line number Diff line change
@@ -93,3 +93,9 @@ def include(*modules)
end

end

# TODO BJF 29-Jun-2016 Normally loaded by a prelude file, review where to load this
begin
require 'unicode_normalize'
rescue LoadError
end

0 comments on commit d3ae421

Please sign in to comment.