nokogiri 1.11.0.rc3-java → 1.11.0.rc4-java

data/lib/nokogiri/html/document.rb

@@ -1,4 +1,7 @@
 # frozen_string_literal: true
+
+require 'pathname'
+
 module Nokogiri
   module HTML
     class Document < Nokogiri::XML::Document
@@ -161,11 +164,12 @@ module Nokogiri
         # Nokogiri::XML::ParseOptions::RECOVER.  See the constants in
         # Nokogiri::XML::ParseOptions.
         def parse string_or_io, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML
-
           options = Nokogiri::XML::ParseOptions.new(options) if Integer === options
-          # Give the options to the user
+
           yield options if block_given?
 
+          url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+
           if string_or_io.respond_to?(:encoding)
             unless string_or_io.encoding.name == "ASCII-8BIT"
               encoding ||= string_or_io.encoding.name
@@ -173,7 +177,12 @@ module Nokogiri
           end
 
           if string_or_io.respond_to?(:read)
-            url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+            if string_or_io.is_a?(Pathname)
+              # resolve the Pathname to the file and open it as an IO object, see #2110
+              string_or_io = string_or_io.expand_path.open
+              url ||= string_or_io.path
+            end
+
             unless encoding
               # Libxml2's parser has poor support for encoding
               # detection.  First, it does not recognize the HTML5
@@ -252,9 +261,6 @@ module Nokogiri
         end
 
         def self.detect_encoding(chunk)
-          if Nokogiri.jruby? && EncodingReader.is_jruby_without_fix?
-            return EncodingReader.detect_encoding_for_jruby_without_fix(chunk)
-          end
           m = chunk.match(/\A(<\?xml[ \t\r\n]+[^>]*>)/) and
             return Nokogiri.XML(m[1]).encoding
 
@@ -273,26 +279,6 @@ module Nokogiri
           end
         end
 
-        def self.is_jruby_without_fix?
-          JRUBY_VERSION.split('.').join.to_i < 165
-        end
-
-        def self.detect_encoding_for_jruby_without_fix(chunk)
-          m = chunk.match(/\A(<\?xml[ \t\r\n]+[^>]*>)/) and
-            return Nokogiri.XML(m[1]).encoding
-
-          m = chunk.match(/(<meta\s)(.*)(charset\s*=\s*([\w-]+))(.*)/i) and
-            return m[4]
-
-          catch(:encoding_found) {
-            Nokogiri::HTML::SAX::Parser.new(JumpSAXHandler.new(:encoding_found.to_s)).parse(chunk)
-            nil
-          }
-        rescue Nokogiri::SyntaxError, RuntimeError
-          # Ignore parser errors that nokogiri may raise
-          nil
-        end
-
         def initialize(io)
           @io = io
           @firstchunk = nil

data/lib/nokogiri/version.rb

@@ -1,150 +1,3 @@
 # frozen_string_literal: true
-module Nokogiri
-  # The version of Nokogiri you are using
-  VERSION = "1.11.0.rc3"
-
-  class VersionInfo # :nodoc:
-    def jruby?
-      ::JRUBY_VERSION if RUBY_PLATFORM == "java"
-    end
-
-    def engine
-      defined?(RUBY_ENGINE) ? RUBY_ENGINE : "mri"
-    end
-
-    def loaded_libxml_version
-      Gem::Version.new(LIBXML_LOADED_VERSION.
-        scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first.
-        collect(&:to_i).
-        join("."))
-    end
-
-    def compiled_libxml_version
-      Gem::Version.new LIBXML_COMPILED_VERSION
-    end
-
-    def loaded_libxslt_version
-      Gem::Version.new(LIBXSLT_LOADED_VERSION.
-        scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first.
-        collect(&:to_i).
-        join("."))
-    end
-
-    def compiled_libxslt_version
-      Gem::Version.new LIBXSLT_COMPILED_VERSION
-    end
-
-    def libxml2?
-      defined?(LIBXML_COMPILED_VERSION)
-    end
-
-    def libxml2_using_system?
-      !libxml2_using_packaged?
-    end
-
-    def libxml2_using_packaged?
-      NOKOGIRI_USE_PACKAGED_LIBRARIES
-    end
-
-    def warnings
-      warnings = []
-
-      if libxml2?
-        if compiled_libxml_version != loaded_libxml_version
-          warnings << "Nokogiri was built against libxml version #{compiled_libxml_version}, but has dynamically loaded #{loaded_libxml_version}"
-        end
-
-        if compiled_libxslt_version != loaded_libxslt_version
-          warnings << "Nokogiri was built against libxslt version #{compiled_libxslt_version}, but has dynamically loaded #{loaded_libxslt_version}"
-        end
-      end
-
-      warnings
-    end
-
-    def to_hash
-      {}.tap do |vi|
-        vi["warnings"] = []
-        vi["nokogiri"] = Nokogiri::VERSION
-        vi["ruby"] = {}.tap do |ruby|
-          ruby["version"] = ::RUBY_VERSION
-          ruby["platform"] = ::RUBY_PLATFORM
-          ruby["gem_platform"] = ::Gem::Platform.local.to_s
-          ruby["description"] = ::RUBY_DESCRIPTION
-          ruby["engine"] = engine
-          ruby["jruby"] = jruby? if jruby?
-        end
-
-        if libxml2?
-          vi["libxml"] = {}.tap do |libxml|
-            if libxml2_using_packaged?
-              libxml["source"] = "packaged"
-              libxml["patches"] = NOKOGIRI_LIBXML2_PATCHES
-            else
-              libxml["source"] = "system"
-            end
-            libxml["compiled"] = compiled_libxml_version.to_s
-            libxml["loaded"] = loaded_libxml_version.to_s
-          end
-
-          vi["libxslt"] = {}.tap do |libxslt|
-            if libxml2_using_packaged?
-              libxslt["source"] = "packaged"
-              libxslt["patches"] = NOKOGIRI_LIBXSLT_PATCHES
-            else
-              libxslt["source"] = "system"
-            end
-            libxslt["compiled"] = compiled_libxslt_version.to_s
-            libxslt["loaded"] = loaded_libxslt_version.to_s
-          end
-
-          vi["warnings"] = warnings
-        elsif jruby?
-          vi["xerces"] = Nokogiri::XERCES_VERSION
-          vi["nekohtml"] = Nokogiri::NEKO_VERSION
-        end
-      end
-    end
-
-    def to_markdown
-      begin
-        require "psych"
-      rescue LoadError
-      end
-      require "yaml"
-      "# Nokogiri (#{Nokogiri::VERSION})\n" +
-      YAML.dump(to_hash).each_line.map { |line| "    #{line}" }.join
-    end
-
-    # FIXME: maybe switch to singleton?
-    @@instance = new
-    @@instance.warnings.each do |warning|
-      warn "WARNING: #{warning}"
-    end
-    def self.instance; @@instance; end
-  end
-
-  def self.uses_libxml?(requirement = nil) # :nodoc:
-    return false unless VersionInfo.instance.libxml2?
-    return true unless requirement
-    return Gem::Requirement.new(requirement).satisfied_by?(VersionInfo.instance.loaded_libxml_version)
-  end
-
-  def self.jruby? # :nodoc:
-    VersionInfo.instance.jruby?
-  end
-
-  # Ensure constants used in this file are loaded - see #1896
-  if Nokogiri.jruby?
-    require "nokogiri/jruby/dependencies"
-  end
-  begin
-    RUBY_VERSION =~ /(\d+\.\d+)/
-    require "nokogiri/#{$1}/nokogiri"
-  rescue LoadError
-    require "nokogiri/nokogiri"
-  end
-
-  # More complete version information about libxml
-  VERSION_INFO = VersionInfo.instance.to_hash
-end
+require_relative "version/constant"
+require_relative "version/info"

data/lib/nokogiri/version/constant.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+module Nokogiri
+  # The version of Nokogiri you are using
+  VERSION = "1.11.0.rc4"
+end

data/lib/nokogiri/version/info.rb

@@ -0,0 +1,182 @@
+# frozen_string_literal: true
+require "singleton"
+require "shellwords"
+
+module Nokogiri
+  class VersionInfo # :nodoc:
+    include Singleton
+
+    def jruby?
+      ::JRUBY_VERSION if ::RUBY_PLATFORM == "java"
+    end
+
+    def engine
+      defined?(::RUBY_ENGINE) ? ::RUBY_ENGINE : "mri"
+    end
+
+    def loaded_libxml_version
+      Gem::Version.new(Nokogiri::LIBXML_LOADED_VERSION
+        .scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first
+        .collect(&:to_i)
+        .join("."))
+    end
+
+    def compiled_libxml_version
+      Gem::Version.new(Nokogiri::LIBXML_COMPILED_VERSION)
+    end
+
+    def loaded_libxslt_version
+      Gem::Version.new(Nokogiri::LIBXSLT_LOADED_VERSION
+        .scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first
+        .collect(&:to_i)
+        .join("."))
+    end
+
+    def compiled_libxslt_version
+      Gem::Version.new(Nokogiri::LIBXSLT_COMPILED_VERSION)
+    end
+
+    def libxml2?
+      defined?(Nokogiri::LIBXML_COMPILED_VERSION)
+    end
+
+    def libxml2_has_iconv?
+      defined?(Nokogiri::LIBXML_ICONV_ENABLED) && Nokogiri::LIBXML_ICONV_ENABLED
+    end
+
+    def libxml2_using_packaged?
+      libxml2? && Nokogiri::PACKAGED_LIBRARIES
+    end
+
+    def libxml2_using_system?
+      libxml2? && !libxml2_using_packaged?
+    end
+
+    def libxml2_precompiled?
+      libxml2_using_packaged? && Nokogiri::PRECOMPILED_LIBRARIES
+    end
+
+    def warnings
+      warnings = []
+
+      if libxml2?
+        if compiled_libxml_version != loaded_libxml_version
+          warnings << "Nokogiri was built against libxml version #{compiled_libxml_version}, but has dynamically loaded #{loaded_libxml_version}"
+        end
+
+        if compiled_libxslt_version != loaded_libxslt_version
+          warnings << "Nokogiri was built against libxslt version #{compiled_libxslt_version}, but has dynamically loaded #{loaded_libxslt_version}"
+        end
+      end
+
+      warnings
+    end
+
+    def to_hash
+      header_directory = File.expand_path(File.join(File.dirname(__FILE__), "../../../ext/nokogiri"))
+      {}.tap do |vi|
+        vi["warnings"] = []
+        vi["nokogiri"] = {}.tap do |nokogiri|
+          nokogiri["version"] = Nokogiri::VERSION
+
+          unless jruby?
+            cppflags = ["-I#{header_directory.shellescape}"]
+            if libxml2_using_packaged?
+              cppflags << "-I#{File.join(header_directory, "include").shellescape}"
+              cppflags << "-I#{File.join(header_directory, "include/libxml2").shellescape}"
+            end
+            nokogiri["cppflags"] = cppflags
+          end
+        end
+        vi["ruby"] = {}.tap do |ruby|
+          ruby["version"] = ::RUBY_VERSION
+          ruby["platform"] = ::RUBY_PLATFORM
+          ruby["gem_platform"] = ::Gem::Platform.local.to_s
+          ruby["description"] = ::RUBY_DESCRIPTION
+          ruby["engine"] = engine
+          ruby["jruby"] = jruby? if jruby?
+        end
+
+        if libxml2?
+          vi["libxml"] = {}.tap do |libxml|
+            if libxml2_using_packaged?
+              libxml["source"] = "packaged"
+              libxml["precompiled"] = libxml2_precompiled?
+              libxml["patches"] = Nokogiri::LIBXML2_PATCHES
+
+              # this is for nokogumbo and shouldn't be forever
+              libxml["libxml2_path"] = header_directory
+            else
+              libxml["source"] = "system"
+            end
+            libxml["iconv_enabled"] = libxml2_has_iconv?
+            libxml["compiled"] = compiled_libxml_version.to_s
+            libxml["loaded"] = loaded_libxml_version.to_s
+          end
+
+          vi["libxslt"] = {}.tap do |libxslt|
+            if libxml2_using_packaged?
+              libxslt["source"] = "packaged"
+              libxslt["precompiled"] = libxml2_precompiled?
+              libxslt["patches"] = Nokogiri::LIBXSLT_PATCHES
+            else
+              libxslt["source"] = "system"
+            end
+            libxslt["compiled"] = compiled_libxslt_version.to_s
+            libxslt["loaded"] = loaded_libxslt_version.to_s
+          end
+
+          vi["warnings"] = warnings
+        end
+
+        if defined?(Nokogiri::OTHER_LIBRARY_VERSIONS)
+          # see extconf for how this string is assembled: "lib1name:lib1version,lib2name:lib2version"
+          vi["other_libraries"] = Hash[*Nokogiri::OTHER_LIBRARY_VERSIONS.split(/[,:]/)]
+        elsif jruby?
+          vi["other_libraries"] = {}.tap do |ol|
+            ol["xerces"] = Nokogiri::XERCES_VERSION
+            ol["nekohtml"] = Nokogiri::NEKO_VERSION
+          end
+        end
+      end
+    end
+
+    def to_markdown
+      begin
+        require "psych"
+      rescue LoadError
+      end
+      require "yaml"
+      "# Nokogiri (#{Nokogiri::VERSION})\n" +
+      YAML.dump(to_hash).each_line.map { |line| "    #{line}" }.join
+    end
+
+    instance.warnings.each do |warning|
+      warn "WARNING: #{warning}"
+    end
+  end
+
+  def self.uses_libxml?(requirement = nil) # :nodoc:
+    return false unless VersionInfo.instance.libxml2?
+    return true unless requirement
+    Gem::Requirement.new(requirement).satisfied_by?(VersionInfo.instance.loaded_libxml_version)
+  end
+
+  def self.jruby? # :nodoc:
+    VersionInfo.instance.jruby?
+  end
+
+  # Ensure constants used in this file are loaded - see #1896
+  if Nokogiri.jruby?
+    require "nokogiri/jruby/dependencies"
+  end
+  begin
+    ::RUBY_VERSION =~ /(\d+\.\d+)/
+    require "nokogiri/#{Regexp.last_match(1)}/nokogiri"
+  rescue LoadError
+    require "nokogiri/nokogiri"
+  end
+
+  # More complete version information about libxml
+  VERSION_INFO = VersionInfo.instance.to_hash
+end

data/lib/nokogiri/xml/document.rb

@@ -1,4 +1,7 @@
 # frozen_string_literal: true
+
+require 'pathname'
+
 module Nokogiri
   module XML
     ##
@@ -44,9 +47,11 @@ module Nokogiri
       #
       def self.parse string_or_io, url = nil, encoding = nil, options = ParseOptions::DEFAULT_XML
         options = Nokogiri::XML::ParseOptions.new(options) if Integer === options
-        # Give the options to the user
+
         yield options if block_given?
 
+        url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+
         if empty_doc?(string_or_io)
           if options.strict?
             raise Nokogiri::XML::SyntaxError.new("Empty document")
@@ -56,12 +61,17 @@ module Nokogiri
         end
 
         doc = if string_or_io.respond_to?(:read)
-          url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
-          read_io(string_or_io, url, encoding, options.to_i)
-        else
-          # read_memory pukes on empty docs
-          read_memory(string_or_io, url, encoding, options.to_i)
-        end
+                if string_or_io.is_a?(Pathname)
+                  # resolve the Pathname to the file and open it as an IO object, see #2110
+                  string_or_io = string_or_io.expand_path.open
+                  url ||= string_or_io.path
+                end
+
+                read_io(string_or_io, url, encoding, options.to_i)
+              else
+                # read_memory pukes on empty docs
+                read_memory(string_or_io, url, encoding, options.to_i)
+              end
 
         # do xinclude processing
         doc.do_xinclude(options) if options.xinclude?