addressable 2.8.1 → 2.8.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3ddda72232f6aef9f6f4311c2855f1b3fea9acc80f51c4e5e90bd23820b0d74e
-  data.tar.gz: 88f208cb2d73dec64663e6e3c1710dc08b188402937038fefce6a2d47debc97d
+  metadata.gz: 62b62e368b9dffbec4c9aaf879257067698b83daff1cc13542352da419f3f88e
+  data.tar.gz: 7f3231aecc91ef0c3087ff2325d78e9a1f4c9f05a752f2c35c8f0bb740fd5996
 SHA512:
-  metadata.gz: 4e1b0c83fc2f2e54cba6804f3840c8bb29f0d73259979d5bf678ebd5d32257b91585d9b71a780321427835d10f98b9c07969267878f0b032c53d031c30202a4c
-  data.tar.gz: 35abc3652c8ff92032411e4daad7133b7c4649aff4a1a25fe622cf1c9b661d56f096a1084392e053788baea79f551a9d4a448d16c9e61e23bf4d9692a6728bf3
+  metadata.gz: 670e1aeddc5a4f62ad39040007cef7371107237b2e4064755355dfbc43cfc62a956f78358c95793a9f55cd325b7b46ce532003bdd9711b1bc3c79b35e0d3cbf6
+  data.tar.gz: d7aed8469d926b0bcdd19202593f8bfa162d9dafb38dabca632ccee8af3fd26920dfbce3f18b674d583f5aaa2e7c51717ccf46fab3a1a4fb4a8dfbd74a80da16

data/CHANGELOG.md

@@ -1,20 +1,59 @@
-# Addressable 2.8.1
+# Addressable 2.8.7 <a name="v2.8.7">
+- Allow `public_suffix` 6 ([#535])
+
+[#535]: https://github.com/sporkmonger/addressable/pull/535
+
+# Addressable 2.8.6 <a name="v2.8.6">
+- Memoize regexps for common character classes ([#524])
+
+[#524]: https://github.com/sporkmonger/addressable/pull/524
+
+# Addressable 2.8.5 <a name="v2.8.5">
+- Fix thread safety issue with encoding tables ([#515])
+- Define URI::NONE as a module to avoid serialization issues ([#509])
+- Fix YAML serialization ([#508])
+
+[#508]: https://github.com/sporkmonger/addressable/pull/508
+[#509]: https://github.com/sporkmonger/addressable/pull/509
+[#515]: https://github.com/sporkmonger/addressable/pull/515
+
+# Addressable 2.8.4 <a name="v2.8.4">
+- Restore `Addressable::IDNA.unicode_normalize_kc` as a deprecated method ([#504])
+
+[#504]: https://github.com/sporkmonger/addressable/pull/504
+
+# Addressable 2.8.3 <a name="v2.8.3">
+- Fix template expand level 2 hash support for non-string objects ([#499], [#498])
+
+[#499]: https://github.com/sporkmonger/addressable/pull/499
+[#498]: https://github.com/sporkmonger/addressable/pull/498
+
+# Addressable 2.8.2 <a name="v2.8.2">
+- Improve cache hits and JIT friendliness ([#486](https://github.com/sporkmonger/addressable/pull/486))
+- Improve code style and test coverage ([#482](https://github.com/sporkmonger/addressable/pull/482))
+- Ensure reset of deferred validation ([#481](https://github.com/sporkmonger/addressable/pull/481))
+- Resolve normalization differences between `IDNA::Native` and `IDNA::Pure` ([#408](https://github.com/sporkmonger/addressable/issues/408), [#492])
+- Remove redundant colon in `Addressable::URI::CharacterClasses::AUTHORITY` regex ([#438](https://github.com/sporkmonger/addressable/pull/438)) (accidentally reverted by [#449] merge but [added back](https://github.com/sporkmonger/addressable/pull/492#discussion_r1105125280) in [#492])
+
+[#492]: https://github.com/sporkmonger/addressable/pull/492
+
+# Addressable 2.8.1 <a name="v2.8.1">
 - refactor `Addressable::URI.normalize_path` to address linter offenses ([#430](https://github.com/sporkmonger/addressable/pull/430))
-- remove redundant colon in `Addressable::URI::CharacterClasses::AUTHORITY` regex ([#438](https://github.com/sporkmonger/addressable/pull/438))
 - update gemspec to reflect supported Ruby versions ([#466], [#464], [#463])
 - compatibility w/ public_suffix 5.x ([#466], [#465], [#460])
 - fixes "invalid byte sequence in UTF-8" exception when unencoding URLs containing non UTF-8 characters ([#459](https://github.com/sporkmonger/addressable/pull/459))
-- `Ractor` compatibility ([#449](https://github.com/sporkmonger/addressable/pull/449))
+- `Ractor` compatibility ([#449])
 - use the whole string instead of a single line for template match ([#431](https://github.com/sporkmonger/addressable/pull/431))
 - force UTF-8 encoding only if needed ([#341](https://github.com/sporkmonger/addressable/pull/341))
 
+[#449]: https://github.com/sporkmonger/addressable/pull/449
 [#460]: https://github.com/sporkmonger/addressable/pull/460
 [#463]: https://github.com/sporkmonger/addressable/pull/463
 [#464]: https://github.com/sporkmonger/addressable/pull/464
 [#465]: https://github.com/sporkmonger/addressable/pull/465
 [#466]: https://github.com/sporkmonger/addressable/pull/466
 
-# Addressable 2.8.0
+# Addressable 2.8.0 <a name="v2.8.0">
 - fixes ReDoS vulnerability in Addressable::Template#match
 - no longer replaces `+` with spaces in queries for non-http(s) schemes
 - fixed encoding ipv6 literals
@@ -26,14 +65,14 @@
 - performance improvements
 - switch CI/CD to GitHub Actions
 
-# Addressable 2.7.0
+# Addressable 2.7.0 <a name="v2.7.0">
 - added `:compacted` flag to `normalized_query`
 - `heuristic_parse` handles `mailto:` more intuitively
 - dropped explicit support for JRuby 9.0.5.0
 - compatibility w/ public_suffix 4.x
 - performance improvements
 
-# Addressable 2.6.0
+# Addressable 2.6.0 <a name="v2.6.0">
 - added `tld=` method to allow assignment to the public suffix
 - most `heuristic_parse` patterns are now case-insensitive
 - `heuristic_parse` handles more `file://` URI variations
@@ -44,17 +83,17 @@
 - minor performance improvements in regexps
 - fixes to eliminate warnings
 
-# Addressable 2.5.2
+# Addressable 2.5.2 <a name="v2.5.2">
 - better support for frozen string literals
 - fixed bug w/ uppercase characters in scheme
 - IDNA errors w/ emoji URLs
 - compatibility w/ public_suffix 3.x
 
-# Addressable 2.5.1
+# Addressable 2.5.1 <a name="v2.5.1">
 - allow unicode normalization to be disabled for URI Template expansion
 - removed duplicate test
 
-# Addressable 2.5.0
+# Addressable 2.5.0 <a name="v2.5.0">
 - dropping support for Ruby 1.9
 - adding support for Ruby 2.4 preview
 - add support for public suffixes and tld; first runtime dependency
@@ -68,7 +107,7 @@
 - host parts longer than 63 bytes will be ignored and not passed to libidn
 - normalized values always encoded as UTF-8
 
-# Addressable 2.4.0
+# Addressable 2.4.0 <a name="v2.4.0">
 - support for 1.8.x dropped
 - double quotes in a host now raises an error
 - newlines in host will no longer get unescaped during normalization
@@ -80,17 +119,17 @@
 - fixed minor bug where an exception would be thrown for a missing ACE suffix
 - better partial expansion of URI templates
 
-# Addressable 2.3.8
+# Addressable 2.3.8 <a name="v2.3.8">
 - fix warnings
 - update dependency gems
 - support for 1.8.x officially deprecated
 
-# Addressable 2.3.7
+# Addressable 2.3.7 <a name="v2.3.7">
 - fix scenario in which invalid URIs don't get an exception until inspected
 - handle hostnames with two adjacent periods correctly
 - upgrade of RSpec
 
-# Addressable 2.3.6
+# Addressable 2.3.6 <a name="v2.3.6">
 - normalization drops empty query string
 - better handling in template extract for missing values
 - template modifier for `'?'` now treated as optional
@@ -99,19 +138,19 @@
 - added `:sorted` option to normalization of query strings
 - fixed issue with normalization of hosts given in `'example.com.'` form
 
-# Addressable 2.3.5
+# Addressable 2.3.5 <a name="v2.3.5">
 - added Addressable::URI#empty? method
 - Addressable::URI#hostname methods now strip square brackets from IPv6 hosts
 - compatibility with Net::HTTP in Ruby 2.0.0
 - Addressable::URI#route_from should always give relative URIs
 
-# Addressable 2.3.4
+# Addressable 2.3.4 <a name="v2.3.4">
 - fixed issue with encoding altering its inputs
 - query string normalization now leaves ';' characters alone
 - FakeFS is detected before attempting to load unicode tables
 - additional testing to ensure frozen objects don't cause problems
 
-# Addressable 2.3.3
+# Addressable 2.3.3 <a name="v2.3.3">
 - fixed issue with converting common primitives during template expansion
 - fixed port encoding issue
 - removed a few warnings
@@ -120,59 +159,59 @@
 - no template match should now result in nil instead of an empty MatchData
 - added license information to gemspec
 
-# Addressable 2.3.2
+# Addressable 2.3.2 <a name="v2.3.2">
 - added Addressable::URI#default_port method
 - fixed issue with Marshalling Unicode data on Windows
 - improved heuristic parsing to better handle IPv4 addresses
 
-# Addressable 2.3.1
+# Addressable 2.3.1 <a name="v2.3.1">
 - fixed missing unicode data file
 
-# Addressable 2.3.0
+# Addressable 2.3.0 <a name="v2.3.0">
 - updated Addressable::Template to use RFC 6570, level 4
 - fixed compatibility problems with some versions of Ruby
 - moved unicode tables into a data file for performance reasons
 - removing support for multiple query value notations
 
-# Addressable 2.2.8
+# Addressable 2.2.8 <a name="v2.2.8">
 - fixed issues with dot segment removal code
 - form encoding can now handle multiple values per key
 - updated development environment
 
-# Addressable 2.2.7
+# Addressable 2.2.7 <a name="v2.2.7">
 - fixed issues related to Addressable::URI#query_values=
 - the Addressable::URI.parse method is now polymorphic
 
-# Addressable 2.2.6
+# Addressable 2.2.6 <a name="v2.2.6">
 - changed the way ambiguous paths are handled
 - fixed bug with frozen URIs
 - https supported in heuristic parsing
 
-# Addressable 2.2.5
+# Addressable 2.2.5 <a name="v2.2.5">
 - 'parsing' a pre-parsed URI object is now a dup operation
 - introduced conditional support for libidn
 - fixed normalization issue on ampersands in query strings
 - added additional tests around handling of query strings
 
-# Addressable 2.2.4
+# Addressable 2.2.4 <a name="v2.2.4">
 - added origin support from draft-ietf-websec-origin-00
 - resolved issue with attempting to navigate below root
 - fixed bug with string splitting in query strings
 
-# Addressable 2.2.3
+# Addressable 2.2.3 <a name="v2.2.3">
 - added :flat_array notation for query strings
 
-# Addressable 2.2.2
+# Addressable 2.2.2 <a name="v2.2.2">
 - fixed issue with percent escaping of '+' character in query strings
 
-# Addressable 2.2.1
+# Addressable 2.2.1 <a name="v2.2.1">
 - added support for application/x-www-form-urlencoded.
 
-# Addressable 2.2.0
+# Addressable 2.2.0 <a name="v2.2.0">
 - added site methods
 - improved documentation
 
-# Addressable 2.1.2
+# Addressable 2.1.2 <a name="v2.1.2">
 - added HTTP request URI methods
 - better handling of Windows file paths
 - validation_deferred boolean replaced with defer_validation block
@@ -180,14 +219,14 @@
 - fixed issue with constructing URIs with relative paths
 - fixed warnings
 
-# Addressable 2.1.1
+# Addressable 2.1.1 <a name="v2.1.1">
 - more type checking changes
 - fixed issue with unicode normalization
 - added method to find template defaults
 - symbolic keys are now allowed in template mappings
 - numeric values and symbolic values are now allowed in template mappings
 
-# Addressable 2.1.0
+# Addressable 2.1.0 <a name="v2.1.0">
 - refactored URI template support out into its own class
 - removed extract method due to being useless and unreliable
 - removed Addressable::URI.expand_template
@@ -201,15 +240,15 @@
 - worked around issue with freezing URIs
 - improved specs
 
-# Addressable 2.0.2
+# Addressable 2.0.2 <a name="v2.0.2">
 - fixed issue with URI template expansion
 - fixed issue with percent escaping characters 0-15
 
-# Addressable 2.0.1
+# Addressable 2.0.1 <a name="v2.0.1">
 - fixed issue with query string assignment
 - fixed issue with improperly encoded components
 
-# Addressable 2.0.0
+# Addressable 2.0.0 <a name="v2.0.0">
 - the initialize method now takes an options hash as its only parameter
 - added query_values method to URI class
 - completely replaced IDNA implementation with pure Ruby
@@ -224,20 +263,20 @@
 - updated URI Template code to match v 03 of the draft spec
 - added a bunch of new specifications
 
-# Addressable 1.0.4
+# Addressable 1.0.4 <a name="v1.0.4">
 - switched to using RSpec's pending system for specs that rely on IDN
 - fixed issue with creating URIs with paths that are not prefixed with '/'
 
-# Addressable 1.0.3
+# Addressable 1.0.3 <a name="v1.0.3">
 - implemented a hash method
 
-# Addressable 1.0.2
+# Addressable 1.0.2 <a name="v1.0.2">
 - fixed minor bug with the extract_mapping method
 
-# Addressable 1.0.1
+# Addressable 1.0.1 <a name="v1.0.1">
 - fixed minor bug with the extract_mapping method
 
-# Addressable 1.0.0
+# Addressable 1.0.0 <a name="v1.0.0">
 - heuristic parse method added
 - parsing is slightly more strict
 - replaced to_h with to_hash
@@ -246,16 +285,16 @@
 - improved heckle rake task
 - no surviving heckle mutations
 
-# Addressable 0.1.2
+# Addressable 0.1.2 <a name="v0.1.2">
 - improved normalization
 - fixed bug in joining algorithm
 - updated specifications
 
-# Addressable 0.1.1
+# Addressable 0.1.1 <a name="v0.1.1">
 - updated documentation
 - added URI Template variable extraction
 
-# Addressable 0.1.0
+# Addressable 0.1.0 <a name="v0.1.0">
 - initial release
 - implementation based on RFC 3986, 3987
 - support for IRIs via libidn

data/Gemfile

@@ -5,6 +5,7 @@ source 'https://rubygems.org'
 gemspec
 
 group :test do
+  gem 'bigdecimal' if RUBY_VERSION > '2.4'
   gem 'rspec', '~> 3.8'
   gem 'rspec-its', '~> 1.3'
 end

data/Rakefile

@@ -20,15 +20,18 @@ additionally provides extensive support for IRIs and URI templates.
 TEXT
 
 PKG_FILES = FileList[
-    "lib/**/*", "spec/**/*", "vendor/**/*", "data/**/*",
-    "tasks/**/*",
-    "[A-Z]*", "Rakefile"
-].exclude(/pkg/).exclude(/database\.yml/).
-  exclude(/Gemfile\.lock/).exclude(/[_\.]git$/)
+    "data/**/*",
+    "lib/**/*.rb",
+    "spec/**/*.rb",
+    "tasks/**/*.rake",
+    "addressable.gemspec",
+    "CHANGELOG.md",
+    "Gemfile",
+    "LICENSE.txt",
+    "README.md",
+    "Rakefile",
+]
 
 task :default => "spec"
 
-WINDOWS = (RUBY_PLATFORM =~ /mswin|win32|mingw|bccwin|cygwin/) rescue false
-SUDO = WINDOWS ? '' : ('sudo' unless ENV['SUDOLESS'])
-
 Dir['tasks/**/*.rake'].each { |rake| load rake }

data/addressable.gemspec

@@ -0,0 +1,28 @@
+# -*- encoding: utf-8 -*-
+# stub: addressable 2.8.7 ruby lib
+
+Gem::Specification.new do |s|
+  s.name = "addressable".freeze
+  s.version = "2.8.7".freeze
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 0".freeze) if s.respond_to? :required_rubygems_version=
+  s.metadata = { "changelog_uri" => "https://github.com/sporkmonger/addressable/blob/main/CHANGELOG.md#v2.8.7" } if s.respond_to? :metadata=
+  s.require_paths = ["lib".freeze]
+  s.authors = ["Bob Aman".freeze]
+  s.date = "2024-06-21"
+  s.description = "Addressable is an alternative implementation to the URI implementation that is\npart of Ruby's standard library. It is flexible, offers heuristic parsing, and\nadditionally provides extensive support for IRIs and URI templates.\n".freeze
+  s.email = "bob@sporkmonger.com".freeze
+  s.extra_rdoc_files = ["README.md".freeze]
+  s.files = ["CHANGELOG.md".freeze, "Gemfile".freeze, "LICENSE.txt".freeze, "README.md".freeze, "Rakefile".freeze, "addressable.gemspec".freeze, "data/unicode.data".freeze, "lib/addressable.rb".freeze, "lib/addressable/idna.rb".freeze, "lib/addressable/idna/native.rb".freeze, "lib/addressable/idna/pure.rb".freeze, "lib/addressable/template.rb".freeze, "lib/addressable/uri.rb".freeze, "lib/addressable/version.rb".freeze, "spec/addressable/idna_spec.rb".freeze, "spec/addressable/net_http_compat_spec.rb".freeze, "spec/addressable/security_spec.rb".freeze, "spec/addressable/template_spec.rb".freeze, "spec/addressable/uri_spec.rb".freeze, "spec/spec_helper.rb".freeze, "tasks/clobber.rake".freeze, "tasks/gem.rake".freeze, "tasks/git.rake".freeze, "tasks/metrics.rake".freeze, "tasks/profile.rake".freeze, "tasks/rspec.rake".freeze, "tasks/yard.rake".freeze]
+  s.homepage = "https://github.com/sporkmonger/addressable".freeze
+  s.licenses = ["Apache-2.0".freeze]
+  s.rdoc_options = ["--main".freeze, "README.md".freeze]
+  s.required_ruby_version = Gem::Requirement.new(">= 2.2".freeze)
+  s.rubygems_version = "3.5.11".freeze
+  s.summary = "URI Implementation".freeze
+
+  s.specification_version = 4
+
+  s.add_runtime_dependency(%q<public_suffix>.freeze, [">= 2.0.2".freeze, "< 7.0".freeze])
+  s.add_development_dependency(%q<bundler>.freeze, [">= 1.0".freeze, "< 3.0".freeze])
+end

data/lib/addressable/idna/native.rb

@@ -29,8 +29,14 @@ module Addressable
        IDN::Punycode.decode(value.to_s)
      end
 
-    def self.unicode_normalize_kc(value)
-      IDN::Stringprep.nfkc_normalize(value.to_s)
+    class << self
+      # @deprecated Use {String#unicode_normalize(:nfkc)} instead
+      def unicode_normalize_kc(value)
+        value.to_s.unicode_normalize(:nfkc)
+      end
+
+      extend Gem::Deprecate
+      deprecate :unicode_normalize_kc, "String#unicode_normalize(:nfkc)", 2023, 4
     end
 
     def self.to_ascii(value)

data/lib/addressable/idna/pure.rb

@@ -66,7 +66,7 @@ module Addressable
     # domain name as described in RFC 3490.
     def self.to_ascii(input)
       input = input.to_s unless input.is_a?(String)
-      input = input.dup
+      input = input.dup.force_encoding(Encoding::UTF_8).unicode_normalize(:nfkc)
       if input.respond_to?(:force_encoding)
         input.force_encoding(Encoding::ASCII_8BIT)
       end
@@ -77,7 +77,7 @@ module Addressable
             part.force_encoding(Encoding::ASCII_8BIT)
           end
           if part =~ UTF8_REGEX && part =~ UTF8_REGEX_MULTIBYTE
-            ACE_PREFIX + punycode_encode(unicode_normalize_kc(part))
+            ACE_PREFIX + punycode_encode(part)
           else
             part
           end
@@ -112,13 +112,14 @@ module Addressable
       output
     end
 
-    # Unicode normalization form KC.
-    def self.unicode_normalize_kc(input)
-      input = input.to_s unless input.is_a?(String)
-      unpacked = input.unpack("U*")
-      unpacked =
-        unicode_compose(unicode_sort_canonical(unicode_decompose(unpacked)))
-      return unpacked.pack("U*")
+    class << self
+      # @deprecated Use {String#unicode_normalize(:nfkc)} instead
+      def unicode_normalize_kc(value)
+        value.to_s.unicode_normalize(:nfkc)
+      end
+
+      extend Gem::Deprecate
+      deprecate :unicode_normalize_kc, "String#unicode_normalize(:nfkc)", 2023, 4
     end
 
     ##
@@ -136,164 +137,6 @@ module Addressable
     end
     private_class_method :unicode_downcase
 
-    def self.unicode_compose(unpacked)
-      unpacked_result = []
-      length = unpacked.length
-
-      return unpacked if length == 0
-
-      starter = unpacked[0]
-      starter_cc = lookup_unicode_combining_class(starter)
-      starter_cc = 256 if starter_cc != 0
-      for i in 1...length
-        ch = unpacked[i]
-
-        if (starter_cc == 0 &&
-            (composite = unicode_compose_pair(starter, ch)) != nil)
-          starter = composite
-        else
-          unpacked_result << starter
-          starter = ch
-        end
-      end
-      unpacked_result << starter
-      return unpacked_result
-    end
-    private_class_method :unicode_compose
-
-    def self.unicode_compose_pair(ch_one, ch_two)
-      if ch_one >= HANGUL_LBASE && ch_one < HANGUL_LBASE + HANGUL_LCOUNT &&
-          ch_two >= HANGUL_VBASE && ch_two < HANGUL_VBASE + HANGUL_VCOUNT
-        # Hangul L + V
-        return HANGUL_SBASE + (
-          (ch_one - HANGUL_LBASE) * HANGUL_VCOUNT + (ch_two - HANGUL_VBASE)
-        ) * HANGUL_TCOUNT
-      elsif ch_one >= HANGUL_SBASE &&
-          ch_one < HANGUL_SBASE + HANGUL_SCOUNT &&
-          (ch_one - HANGUL_SBASE) % HANGUL_TCOUNT == 0 &&
-          ch_two >= HANGUL_TBASE && ch_two < HANGUL_TBASE + HANGUL_TCOUNT
-           # Hangul LV + T
-        return ch_one + (ch_two - HANGUL_TBASE)
-      end
-
-      p = []
-
-      ucs4_to_utf8(ch_one, p)
-      ucs4_to_utf8(ch_two, p)
-
-      return lookup_unicode_composition(p)
-    end
-    private_class_method :unicode_compose_pair
-
-    def self.ucs4_to_utf8(char, buffer)
-      if char < 128
-        buffer << char
-      elsif char < 2048
-        buffer << (char >> 6 | 192)
-        buffer << (char & 63 | 128)
-      elsif char < 0x10000
-        buffer << (char >> 12 | 224)
-        buffer << (char >> 6 & 63 | 128)
-        buffer << (char & 63 | 128)
-      elsif char < 0x200000
-        buffer << (char >> 18 | 240)
-        buffer << (char >> 12 & 63 | 128)
-        buffer << (char >> 6 & 63 | 128)
-        buffer << (char & 63 | 128)
-      elsif char < 0x4000000
-        buffer << (char >> 24 | 248)
-        buffer << (char >> 18 & 63 | 128)
-        buffer << (char >> 12 & 63 | 128)
-        buffer << (char >> 6 & 63 | 128)
-        buffer << (char & 63 | 128)
-      elsif char < 0x80000000
-        buffer << (char >> 30 | 252)
-        buffer << (char >> 24 & 63 | 128)
-        buffer << (char >> 18 & 63 | 128)
-        buffer << (char >> 12 & 63 | 128)
-        buffer << (char >> 6 & 63 | 128)
-        buffer << (char & 63 | 128)
-      end
-    end
-    private_class_method :ucs4_to_utf8
-
-    def self.unicode_sort_canonical(unpacked)
-      unpacked = unpacked.dup
-      i = 1
-      length = unpacked.length
-
-      return unpacked if length < 2
-
-      while i < length
-        last = unpacked[i-1]
-        ch = unpacked[i]
-        last_cc = lookup_unicode_combining_class(last)
-        cc = lookup_unicode_combining_class(ch)
-        if cc != 0 && last_cc != 0 && last_cc > cc
-          unpacked[i] = last
-          unpacked[i-1] = ch
-          i -= 1 if i > 1
-        else
-          i += 1
-        end
-      end
-      return unpacked
-    end
-    private_class_method :unicode_sort_canonical
-
-    def self.unicode_decompose(unpacked)
-      unpacked_result = []
-      for cp in unpacked
-        if cp >= HANGUL_SBASE && cp < HANGUL_SBASE + HANGUL_SCOUNT
-          l, v, t = unicode_decompose_hangul(cp)
-          unpacked_result << l
-          unpacked_result << v if v
-          unpacked_result << t if t
-        else
-          dc = lookup_unicode_compatibility(cp)
-          unless dc
-            unpacked_result << cp
-          else
-            unpacked_result.concat(unicode_decompose(dc.unpack("U*")))
-          end
-        end
-      end
-      return unpacked_result
-    end
-    private_class_method :unicode_decompose
-
-    def self.unicode_decompose_hangul(codepoint)
-      sindex = codepoint - HANGUL_SBASE;
-      if sindex < 0 || sindex >= HANGUL_SCOUNT
-        l = codepoint
-        v = t = nil
-        return l, v, t
-      end
-      l = HANGUL_LBASE + sindex / HANGUL_NCOUNT
-      v = HANGUL_VBASE + (sindex % HANGUL_NCOUNT) / HANGUL_TCOUNT
-      t = HANGUL_TBASE + sindex % HANGUL_TCOUNT
-      if t == HANGUL_TBASE
-        t = nil
-      end
-      return l, v, t
-    end
-    private_class_method :unicode_decompose_hangul
-
-    def self.lookup_unicode_combining_class(codepoint)
-      codepoint_data = UNICODE_DATA[codepoint]
-      (codepoint_data ?
-        (codepoint_data[UNICODE_DATA_COMBINING_CLASS] || 0) :
-        0)
-    end
-    private_class_method :lookup_unicode_combining_class
-
-    def self.lookup_unicode_compatibility(codepoint)
-      codepoint_data = UNICODE_DATA[codepoint]
-      (codepoint_data ?
-        codepoint_data[UNICODE_DATA_COMPATIBILITY] : nil)
-    end
-    private_class_method :lookup_unicode_compatibility
-
     def self.lookup_unicode_lowercase(codepoint)
       codepoint_data = UNICODE_DATA[codepoint]
       (codepoint_data ?
@@ -302,21 +145,6 @@ module Addressable
     end
     private_class_method :lookup_unicode_lowercase
 
-    def self.lookup_unicode_composition(unpacked)
-      return COMPOSITION_TABLE[unpacked]
-    end
-    private_class_method :lookup_unicode_composition
-
-    HANGUL_SBASE =  0xac00
-    HANGUL_LBASE =  0x1100
-    HANGUL_LCOUNT = 19
-    HANGUL_VBASE =  0x1161
-    HANGUL_VCOUNT = 21
-    HANGUL_TBASE =  0x11a7
-    HANGUL_TCOUNT = 28
-    HANGUL_NCOUNT = HANGUL_VCOUNT * HANGUL_TCOUNT # 588
-    HANGUL_SCOUNT = HANGUL_LCOUNT * HANGUL_NCOUNT # 11172
-
     UNICODE_DATA_COMBINING_CLASS = 0
     UNICODE_DATA_EXCLUSION = 1
     UNICODE_DATA_CANONICAL = 2

data/lib/addressable/template.rb

@@ -157,7 +157,7 @@ module Addressable
       #   the ::MatchData#[] behavior.
       #
       # @param [#to_int, nil] len
-      #   If provided, an array of values will be returend with the given
+      #   If provided, an array of values will be returned with the given
       #   parameter used as length.
       #
       # @return [Array, String, nil]
@@ -892,25 +892,24 @@ module Addressable
     # operator.
     #
     # @param [Hash, Array, String] value
-    #   Normalizes keys and values with IDNA#unicode_normalize_kc
+    #   Normalizes unicode keys and values with String#unicode_normalize (NFC)
     #
     # @return [Hash, Array, String] The normalized values
     def normalize_value(value)
-      unless value.is_a?(Hash)
-        value = value.respond_to?(:to_ary) ? value.to_ary : value.to_str
-      end
-
       # Handle unicode normalization
-      if value.kind_of?(Array)
-        value.map! { |val| Addressable::IDNA.unicode_normalize_kc(val) }
+      if value.respond_to?(:to_ary)
+        value.to_ary.map! { |val| normalize_value(val) }
       elsif value.kind_of?(Hash)
         value = value.inject({}) { |acc, (k, v)|
-          acc[Addressable::IDNA.unicode_normalize_kc(k)] =
-            Addressable::IDNA.unicode_normalize_kc(v)
+          acc[normalize_value(k)] = normalize_value(v)
           acc
         }
       else
-        value = Addressable::IDNA.unicode_normalize_kc(value)
+        value = value.to_s if !value.kind_of?(String)
+        if value.encoding != Encoding::UTF_8
+          value = value.dup.force_encoding(Encoding::UTF_8)
+        end
+        value = value.unicode_normalize(:nfc)
       end
       value
     end