uri-whatwg_parser 0.3.1 → 0.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 340c89cc9b0c80699547d7daa0a33c3c081250ebee1dd1178a36ee288a484a22
-  data.tar.gz: 961034f93393490c76ad015aea1524b27b5955ef38112892d50b2a1e15a3ac50
+  metadata.gz: 21b70ae0fff882053d71cb5449b278d4e7b32f16933433f9425903f83c12305d
+  data.tar.gz: a51ae685b0cb8a7db1adb9f05f656be83e60e0da62c6eca0438872ca046debd8
 SHA512:
-  metadata.gz: 987525847dbe6f04d4a2a4c3c8e338d57ebd964d4312528bc16e5e515ba7460922c3fedf08a4151227dbd1fbca11be4af38db6db53e1d741992460fd89c86bd0
-  data.tar.gz: 12eae3fe44a8136aa66a55d1aa7e12b172980ae96470f890e34e721fae4c1f3f23155ff1a2e1db305ce2f239cdf239a7db43fe5feb153eb11f3f230c09fd77a4
+  metadata.gz: 905506fad119c654dc873129f8854431c9260b880de07b154b8ae45bd88b7bc4787a186236ae68ccf21af840c51a22bb84558a5651bb17e7070b166a222a96e0
+  data.tar.gz: 5e5cc0503cf20d6f5eb7fad89bdb60336516c8a83df9fe64715b84af77dcf50caf721ca7f36b31926a6a261ae6f824868a0d13f5892ee3b49b2a34a210a30078

data/CHANGELOG.md

@@ -1,3 +1,14 @@
+## Unreleased
+
+## 0.3.3
+
+* Fix parsing an IPv4 address embedded in an IPv6 address with leading zeroes and non-decimal numbers
+* Fix parsing an IPv6 address to reject a trailing `:`
+
+## 0.3.2
+
+* Fix domain-to-ASCII process on host parsing to follow the latest spec
+
 ## 0.3.1
 
 * Fix a bug of parsing Windows drive letter

data/README.md

@@ -1,7 +1,7 @@
 # URI::WhatwgParser
 Ruby implementation of the [WHATWG URL Living Standard](https://url.spec.whatwg.org/).
 
-The latest revision that this package implements of the standard is [14 April 2026](https://url.spec.whatwg.org/commit-snapshots/b11d73b8caefe90403afe19210db05acba897722/)
+The latest revision that this package implements of the standard is [18 August 2026](https://url.spec.whatwg.org/commit-snapshots/55d6699373ba68a16ec182f34222a74ed8bc3dac/)
 
 ## Installation
 

data/lib/uri/whatwg_parser/host_parser.rb

@@ -77,6 +77,7 @@ class URI::WhatwgParser
       if chars[i] == ":"
         raise ParseError, "invalid IPv6 format" unless chars[i + 1] == ":"
         i += 2
+        piece_index += 1
         compress = piece_index
       end
 
@@ -86,6 +87,7 @@ class URI::WhatwgParser
         if chars[i] == ":"
           raise ParseError, "invalid IPv6 format" if compress
           i += 1
+          piece_index += 1
           compress = piece_index
           next
         end
@@ -99,18 +101,9 @@ class URI::WhatwgParser
         end
 
         if chars[i] == "."
-          # IPv4-mapped address must be valid and complete, no trailing dot
-          ipv4_piece = chars[i - length, chars.length - (i - length)].join
-          parts = ipv4_piece.split(".")
-          if parts.length != 4 || parts.any? { |p| p.empty? } || ipv4_piece.end_with?(".")
-            raise ParseError, "invalid IPv6 format"
-          end
+          raise ParseError, "invalid IPv6 format" if length == 0 || piece_index > 6
 
-          ipv4 = parse_ipv4(ipv4_piece)
-          address[piece_index] = (ipv4 >> 16) & 0xFFFF
-          address[piece_index + 1] = ipv4 & 0xFFFF
-          piece_index += 2
-          i = chars.length
+          piece_index = parse_embedded_ipv4(chars, i - length, address, piece_index)
           break
         end
 
@@ -119,21 +112,16 @@ class URI::WhatwgParser
         address[piece_index] = value
         piece_index += 1
 
-        if i < chars.length
-          if chars[i] == ":"
-            i += 1
-          elsif chars[i] != nil
-            raise ParseError, "invalid IPv6 format"
-          end
+        if chars[i] == ":"
+          i += 1
+          raise ParseError, "invalid IPv6 format" unless chars[i]
+        elsif chars[i]
+          raise ParseError, "invalid IPv6 format"
         end
       end
 
       if compress
-        swaps = piece_index - compress
-        (0...swaps).each do |j|
-          address[7 - j] = address[compress + swaps - 1 - j]
-          address[compress + swaps - 1 - j] = 0
-        end
+        expand_ipv6_compression(address, compress, piece_index)
       elsif piece_index != 8
         raise ParseError, "invalid IPv6 format"
       end
@@ -141,6 +129,52 @@ class URI::WhatwgParser
       compress_ipv6(address)
     end
 
+    def parse_embedded_ipv4(chars, i, address, piece_index)
+      numbers_seen = 0
+
+      while i < chars.length
+        if numbers_seen > 0
+          raise ParseError, "invalid IPv6 format" unless chars[i] == "." && numbers_seen < 4
+          i += 1
+        end
+
+        raise ParseError, "invalid IPv6 format" unless chars[i]&.match?(/[0-9]/)
+
+        ipv4_piece = nil
+        while chars[i]&.match?(/[0-9]/)
+          number = chars[i].to_i
+          if ipv4_piece.nil?
+            ipv4_piece = number
+          elsif ipv4_piece == 0
+            raise ParseError, "invalid IPv6 format"
+          else
+            ipv4_piece = ipv4_piece * 10 + number
+          end
+          raise ParseError, "invalid IPv6 format" if ipv4_piece > 255
+          i += 1
+        end
+
+        address[piece_index] = address[piece_index] * 256 + ipv4_piece
+        numbers_seen += 1
+        piece_index += 1 if numbers_seen == 2 || numbers_seen == 4
+      end
+
+      raise ParseError, "invalid IPv6 format" if numbers_seen != 4
+
+      piece_index
+    end
+
+    def expand_ipv6_compression(address, compress, piece_index)
+      swaps = piece_index - compress
+      piece_index = 7
+
+      while piece_index != 0 && swaps > 0
+        address[piece_index], address[compress + swaps - 1] = address[compress + swaps - 1], address[piece_index]
+        piece_index -= 1
+        swaps -= 1
+      end
+    end
+
     def compress_ipv6(address)
       # Find the longest run of zeros for '::' compression
       best_base = nil
@@ -259,7 +293,13 @@ class URI::WhatwgParser
         return domain
       end
 
-      ascii_domain = URI::IDNA.whatwg_to_ascii(domain.force_encoding(Encoding::UTF_8), be_strict: false)
+      begin
+        ascii_domain = URI::IDNA.whatwg_to_ascii(domain.force_encoding(Encoding::UTF_8), be_strict: false)
+      rescue URI::IDNA::Error, Encoding::CompatibilityError, ArgumentError
+        raise ParseError, "invalid host value" unless domain.ascii_only?
+
+        ascii_domain = domain.downcase
+      end
 
       raise ParseError, "including invalid value in host" if include_forbidden_domain_code_point?(ascii_domain)
       raise ParseError, "host can't be empty" if ascii_domain.empty?

data/lib/uri/whatwg_parser/version.rb

@@ -2,6 +2,6 @@
 
 module URI
   class WhatwgParser
-    VERSION = "0.3.1"
+    VERSION = "0.3.3"
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: uri-whatwg_parser
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.3
 platform: ruby
 authors:
 - Yuji Yaginuma
@@ -73,7 +73,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.10
+rubygems_version: 4.0.16
 specification_version: 4
 summary: Ruby implementation of the WHATWG URL Living Standard
 test_files: []