uri-whatwg_parser 0.1.3 → 0.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d94a0f3f96af7620d7592bb1c0b215b52bfcb7e19c2be83177407a67ae8eaa79
-  data.tar.gz: 14cd125a49fc9b61eb5f70621cfc8685e0cb9b309f048f5528e6c6a37d389eaf
+  metadata.gz: 2061d819a10365a97f389d9cb70325c87b75c390c4145897ed640c3e4cc85f70
+  data.tar.gz: 905ec3bdd174039fc91ce433f84a079fe12ff93f9ac999f420b08760b3fcee53
 SHA512:
-  metadata.gz: 31e07e1a13f149e0188cd7701e2e0828fa66b078004b30bbe55ae1a34989cb68cf4190a0283d6969b718f90712a1cbf61f1efee095ab5847ea6771caf3936661
-  data.tar.gz: 3e251062ddd86971a0829f83fb27f0b1da05016544b95516f859ce66cb6bd35d61abea75dd1bf693d663de47853c82e7ff326d771f4924a3827fe38c8f5a3df4
+  metadata.gz: fc4a0d83fb46e959feb1f7990fa1f119428764447c6d4d455ec5287d5823b19eec845ef4b8912ceeffd6233266b3142859a837c7a26e47e20315d64bb5660006
+  data.tar.gz: bd36808311ae2330cd897ba824219d82cff0497d75531441b3679def9ff5ea1969297ccfcbb8cd02b83909a2f7a5c0de7f0d505bb79e1117494544dd3d4b3cfa

data/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 0.1.4
+
+* Support passing `encoding` option
+
 ## 0.1.3
 
 * Improve performance of parsing

data/README.md

@@ -17,6 +17,8 @@ gem install uri-whatwg_parser
 This gem is compatible with [`uri`](https://github.com/ruby/uri) gem and automatically switches parser's behavior. So users don't need to set up.
 
 ```ruby
+require "uri/whatwg_parser"
+
 URI.parse("http://日本語.jp")
 # => #<URI::HTTP http://xn--wgv71a119e.jp>
 ```
@@ -31,7 +33,6 @@ To install this gem onto your local machine, run `bundle exec rake install`. To
 
 * Support state override
 * Support validations
-* Support encodings other than UTF-8
 
 ## Contributing
 

data/lib/uri/whatwg_parser/parser_helper.rb

@@ -18,11 +18,14 @@ class URI::WhatwgParser
       ASCII_DIGIT.include?(c)
     end
 
-    def percent_encode(c, encode_set)
-      if encode_set.include?(c) || c.ord > 0x7e
-        return c.unpack("C*").map { |b| sprintf("%%%02X", b) }.join
-      end
-      c
+    def percent_encode(c, encode_set, encoding = Encoding::UTF_8)
+      return c unless encode_set.include?(c) || c.ord > 0x7e
+
+      # For ASCII single-byte characters
+      return "%%%02X" % c.ord if c.bytesize == 1
+
+      bytes = c.encoding == encoding ? c.bytes : c.encode(encoding).bytes
+      bytes.map { |b| "%%%02X" % b }.join
     end
   end
 end

data/lib/uri/whatwg_parser/version.rb

@@ -2,6 +2,6 @@
 
 module URI
   class WhatwgParser
-    VERSION = "0.1.3"
+    VERSION = "0.1.4"
   end
 end

data/lib/uri/whatwg_parser.rb

@@ -34,20 +34,21 @@ module URI
       {}
     end
 
-    def parse(uri, base = nil) # :nodoc:
+    def parse(uri, base = nil, encoding = Encoding::UTF_8) # :nodoc:
       reset
-      URI.for(*self.split(uri, base))
+      URI.for(*self.split(uri, base, encoding))
     end
 
-    def split(uri, base = nil) # :nodoc:
+    def split(uri, base = nil, encoding = Encoding::UTF_8) # :nodoc:
       @base = nil
       if base != nil
-        ary = split(base)
+        ary = split(base, nil, encoding)
         @base = { scheme: ary[0], userinfo: ary[1], host: ary[2], port: ary[3], registry: ary[4], path: ary[5], opaque: ary[6], query: ary[7], fragment: ary[8]}
         @base_paths = @paths
         reset
       end
 
+      @encoding = encoding
       @uri = uri.dup
       @uri.sub!(/\A[\u0000-\u0020]*/, "")
       @uri.sub!(/[\u0000-\u0020]*\z/, "")
@@ -100,7 +101,7 @@ module URI
 
     def scheme_start_state(c)
       if ascii_alpha?(c)
-        @buffer += c.downcase
+        @buffer << c.downcase
         @state = :scheme_state
       else
         @pos -= 1
@@ -110,7 +111,7 @@ module URI
 
     def scheme_state(c)
       if ascii_alphanumerica?(c) || ["+", "-", "."].include?(c)
-        @buffer += c.downcase
+        @buffer << c.downcase
       elsif c == ":"
         @parse_result[:scheme] = @buffer
         @buffer = +""
@@ -129,7 +130,7 @@ module URI
           @state = :opaque_path_state
         end
       else
-        @buffer = +""
+        @buffer.clear
         @pos -= 1
         @state = :no_scheme_state
       end
@@ -241,7 +242,7 @@ module URI
             next
           end
 
-          encoded_char = percent_encode(char, USERINFO_PERCENT_ENCODE_SET)
+          encoded_char = percent_encode(char, USERINFO_PERCENT_ENCODE_SET, @encoding)
 
           if @password_token_seen
             @password = @password.to_s + encoded_char
@@ -250,12 +251,12 @@ module URI
           end
         end
 
-        @buffer = +""
+        @buffer.clear
       elsif c.nil? || ["/", "?", "#"].include?(c) || (special_url? && c == "\\")
         raise ParseError, "host is missing" if @at_sign_seen && @buffer.empty?
 
         @pos -= (@buffer.size + 1)
-        @buffer = +""
+        @buffer.clear
         @state = :host_state
       else
         @buffer << c
@@ -267,7 +268,7 @@ module URI
         raise ParseError, "host is missing" if @buffer.empty?
 
         @parse_result[:host] = @host_parser.parse(@buffer, !special_url?)
-        @buffer = +""
+        @buffer.clear
         @state = :port_state
       elsif c.nil? || ["/", "?", "#"].include?(c) || (special_url? && c == "\\")
         @pos -= 1
@@ -275,19 +276,19 @@ module URI
           raise ParseError, "host is missing"
         else
           @parse_result[:host] = @host_parser.parse(@buffer, !special_url?)
-          @buffer = +""
+          @buffer.clear
           @state = :path_start_state
         end
       else
         @inside_brackets = true if c == "["
         @inside_brackets = false if c == "]"
-        @buffer += c
+        @buffer << c
       end
     end
 
     def port_state(c)
       if ascii_digit?(c)
-        @buffer += c
+        @buffer << c
       elsif c.nil? || ["/", "?", "#"].include?(c) || (special_url? && c == "\\")
         unless @buffer.empty?
           begin
@@ -298,7 +299,7 @@ module URI
             raise ParseError, "port is invalid value"
           end
 
-          @buffer = +""
+          @buffer.clear
         end
 
         @state = :path_start_state
@@ -346,7 +347,7 @@ module URI
         if !@base.nil? && @base[:scheme] == "file"
           @parse_result[:host] = @base[:host]
           if !starts_with_windows_drive_letter?(rest) && @base_paths && normalized_windows_drive_letter?(@base_paths[0])
-            @paths[0] += @base_paths[0]
+            @paths[0] << @base_paths[0]
           end
         end
         @state = :path_state
@@ -369,12 +370,12 @@ module URI
             @parse_result[:host] = host
           end
 
-          @buffer = +""
+          @buffer.clear
           @state = :path_start_state
         end
       end
 
-      @buffer += c unless c.nil?
+      @buffer << c unless c.nil?
     end
 
     def path_start_state(c)
@@ -420,7 +421,7 @@ module URI
           @state = :fragment_state
         end
       else
-        @buffer << percent_encode(c, PATH_PERCENT_ENCODE_SET)
+        @buffer << percent_encode(c, PATH_PERCENT_ENCODE_SET, @encoding)
       end
     end
 
@@ -438,15 +439,19 @@ module URI
           @parse_result[:path] = @parse_result[:path].to_s + " "
         end
       elsif !c.nil?
-        @parse_result[:path] = @parse_result[:path].to_s + percent_encode(c, C0_CONTROL_PERCENT_ENCODE_SET)
+        @parse_result[:path] = @parse_result[:path].to_s + percent_encode(c, C0_CONTROL_PERCENT_ENCODE_SET, @encoding)
       end
     end
 
     def query_state(c)
+      if @encoding != Encoding::UTF_8 && (!special_url? || %w[ws wss].include?(@parse_result[:scheme]))
+        @encoding = Encoding::UTF_8
+      end
+
       if c.nil? || c == "#"
         query_percent_encode_set = special_url? ? SPECIAL_QUERY_PERCENT_ENCODE_SET : QUERY_PERCENT_ENCODE_SET
-        @parse_result[:query] = @buffer.chars.map { |c| percent_encode(c, query_percent_encode_set) }.join
-        @buffer = +""
+        @parse_result[:query] = @buffer.chars.map { |c| percent_encode(c, query_percent_encode_set, @encoding) }.join
+        @buffer.clear
         @state = :fragment_state if c == "#"
       elsif !c.nil?
         @buffer << c
@@ -455,7 +460,7 @@ module URI
 
     def fragment_state(c)
       return if c.nil?
-      @parse_result[:fragment] = @parse_result[:fragment].to_s + percent_encode(c, FRAGMENT_PERCENT_ENCODE_SET)
+      @parse_result[:fragment] = @parse_result[:fragment].to_s + percent_encode(c, FRAGMENT_PERCENT_ENCODE_SET, @encoding)
     end
 
     def windows_drive_letter?(str)

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: uri-whatwg_parser
 version: !ruby/object:Gem::Version
-  version: 0.1.3
+  version: 0.1.4
 platform: ruby
 authors:
 - Yuji Yaginuma
 bindir: exe
 cert_chain: []
-date: 2025-05-25 00:00:00.000000000 Z
+date: 2025-06-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: uri
@@ -15,14 +15,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 1.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 1.0.0
 - !ruby/object:Gem::Dependency
   name: uri-idna
   requirement: !ruby/object:Gem::Requirement