RubyGems - nokolexbor - Versions diffs - 0.2.5 → 0.3.1 - Mend

nokolexbor 0.2.5 → 0.3.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (54) hide show

checksums.yaml +4 -4
data/ext/nokolexbor/CMakeLists.txt +7 -4
data/ext/nokolexbor/config.h.cmake.in +2 -0
data/ext/nokolexbor/extconf.rb +47 -25
data/ext/nokolexbor/libxml/SAX2.h +4 -4
data/ext/nokolexbor/libxml/chvalid.h +21 -21
data/ext/nokolexbor/libxml/dict.h +13 -13
data/ext/nokolexbor/libxml/globals.h +202 -202
data/ext/nokolexbor/libxml/hash.h +25 -25
data/ext/nokolexbor/libxml/parser.h +5 -5
data/ext/nokolexbor/libxml/parserInternals.h +4 -4
data/ext/nokolexbor/libxml/pattern.h +14 -14
data/ext/nokolexbor/libxml/threads.h +15 -15
data/ext/nokolexbor/libxml/tree.h +5 -5
data/ext/nokolexbor/libxml/xmlerror.h +5 -5
data/ext/nokolexbor/libxml/xmlmemory.h +16 -16
data/ext/nokolexbor/libxml/xmlstring.h +30 -30
data/ext/nokolexbor/libxml/xpath.h +43 -43
data/ext/nokolexbor/libxml/xpathInternals.h +128 -128
data/ext/nokolexbor/memory.c +6 -6
data/ext/nokolexbor/nl_cdata.c +44 -0
data/ext/nokolexbor/nl_comment.c +44 -0
data/ext/nokolexbor/nl_document.c +23 -9
data/ext/nokolexbor/nl_node.c +191 -178
data/ext/nokolexbor/nl_node_set.c +38 -73
data/ext/nokolexbor/nl_text.c +44 -0
data/ext/nokolexbor/nl_xpath_context.c +33 -42
data/ext/nokolexbor/nokolexbor.c +7 -3
data/ext/nokolexbor/nokolexbor.h +9 -7
data/ext/nokolexbor/private/buf.h +1 -1
data/ext/nokolexbor/private/error.h +3 -3
data/ext/nokolexbor/xml_SAX2.c +8 -8
data/ext/nokolexbor/xml_buf.c +19 -19
data/ext/nokolexbor/xml_chvalid.c +25 -25
data/ext/nokolexbor/xml_dict.c +69 -69
data/ext/nokolexbor/xml_encoding.c +2 -2
data/ext/nokolexbor/xml_error.c +51 -51
data/ext/nokolexbor/xml_globals.c +329 -329
data/ext/nokolexbor/xml_hash.c +131 -131
data/ext/nokolexbor/xml_memory.c +25 -25
data/ext/nokolexbor/xml_parser.c +3 -3
data/ext/nokolexbor/xml_parserInternals.c +15 -15
data/ext/nokolexbor/xml_pattern.c +103 -103
data/ext/nokolexbor/xml_string.c +93 -93
data/ext/nokolexbor/xml_threads.c +61 -61
data/ext/nokolexbor/xml_tree.c +12 -12
data/ext/nokolexbor/xml_xpath.c +1194 -1203
data/lib/nokolexbor/document.rb +92 -1
data/lib/nokolexbor/node.rb +64 -0
data/lib/nokolexbor/node_set.rb +6 -5
data/lib/nokolexbor/version.rb +1 -1
data/lib/nokolexbor.rb +21 -1
data/patches/0001-lexbor-support-text-pseudo-element.patch +1 -1
metadata +7 -4

data/lib/nokolexbor/document.rb CHANGED Viewed

@@ -1,6 +1,97 @@
 # frozen_string_literal: true
 module Nokolexbor
-  class Document < Node
+  class Document < Nokolexbor::Node
+    def create_element(name, *contents_or_attrs, &block)
+      elm = Nokolexbor::Element.new(name, self, &block)
+      contents_or_attrs.each do |arg|
+        case arg
+        when Hash
+          arg.each do |k, v|
+            elm[k.to_s] = v.to_s
+          end
+        else
+          elm.content = arg
+        end
+      end
+      elm
+    end
+    # Create a Text Node with +string+
+    def create_text_node(string, &block)
+      Nokolexbor::Text.new(string.to_s, self, &block)
+    end
+    # Create a CDATA Node containing +string+
+    def create_cdata(string, &block)
+      Nokolexbor::CDATA.new(string.to_s, self, &block)
+    end
+    # Create a Comment Node containing +string+
+    def create_comment(string, &block)
+      Nokolexbor::Comment.new(string.to_s, self, &block)
+    end
+    # A reference to +self+
+    def document
+      self
+    end
+    def meta_encoding
+      if (meta = at_css("meta[charset]"))
+        meta[:charset]
+      elsif (meta = meta_content_type)
+        meta["content"][/charset\s*=\s*([\w-]+)/i, 1]
+      end
+    end
+    def meta_encoding=(encoding)
+      if (meta = meta_content_type)
+        meta["content"] = format("text/html; charset=%s", encoding)
+        encoding
+      elsif (meta = at_css("meta[charset]"))
+        meta["charset"] = encoding
+      else
+        meta = Nokolexbor::Node.new("meta", self)
+        meta["charset"] = encoding
+        if (head = at_css("head"))
+          head.prepend_child(meta)
+        else
+          set_metadata_element(meta)
+        end
+        encoding
+      end
+    end
+    def meta_content_type
+      xpath("//meta[@http-equiv and boolean(@content)]").find do |node|
+        node["http-equiv"] =~ /\AContent-Type\z/i
+      end
+    end
+    private :meta_content_type
+    def set_metadata_element(element)
+      if (head = at_css("head"))
+        head << element
+      elsif (html = at_css("html"))
+        head = html.prepend_child(Nokolexbor::Node.new("head", self))
+        head.prepend_child(element)
+      elsif (first = children.find do |node|
+               case node
+               when Nokolexbor::Node
+                 true
+               end
+             end)
+        # We reach here only if the underlying document model
+        # allows <html>/<head> elements to be omitted and does not
+        # automatically supply them.
+        first.add_previous_sibling(element)
+      else
+        html = add_child(Nokolexbor::Node.new("html", self))
+        head = html.add_child(Nokolexbor::Node.new("head", self))
+        head.prepend_child(element)
+      end
+    end
   end
 end

data/lib/nokolexbor/node.rb CHANGED Viewed

@@ -45,6 +45,10 @@ module Nokolexbor
       type == ELEMENT_NODE
     end
+    def document?
+      is_a?(Nokolexbor::Document)
+    end
     def ancestors(selector = nil)
       return NodeSet.new(@document) unless respond_to?(:parent)
       return NodeSet.new(@document) unless parent
@@ -87,6 +91,56 @@ module Nokolexbor
       self
     end
+    def add_previous_sibling(node_or_tags)
+      raise ArgumentError,
+        "A document may not have multiple root nodes." if parent&.document? && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
+      add_sibling(:previous, node_or_tags)
+    end
+    def add_next_sibling(node_or_tags)
+      raise ArgumentError,
+        "A document may not have multiple root nodes." if parent&.document? && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
+      add_sibling(:next, node_or_tags)
+    end
+    def before(node_or_tags)
+      add_previous_sibling(node_or_tags)
+      self
+    end
+    def after(node_or_tags)
+      add_next_sibling(node_or_tags)
+      self
+    end
+    alias_method :next_sibling, :next
+    alias_method :previous_sibling, :previous
+    alias_method :next=, :add_next_sibling
+    alias_method :previous=, :add_previous_sibling
+    def <<(node_or_tags)
+      add_child(node_or_tags)
+      self
+    end
+    def prepend_child(node)
+      if (first = children.first)
+        # Mimic the error add_child would raise.
+        raise "Document already has a root node" if document? && !(node.comment? || node.processing_instruction?)
+        first.add_sibling(:previous, node)
+      else
+        add_child(node)
+      end
+    end
+    def traverse(&block)
+      children.each { |j| j.traverse(&block) }
+      yield(self)
+    end
     def matches?(selector)
       ancestors.last.css(selector).any? { |node| node == self }
     end
@@ -118,6 +172,10 @@ module Nokolexbor
       end
     end
+    def parent=(parent_node)
+      parent_node.add_child(self)
+    end
     def each
       attributes.each do |name, node|
         yield [name, node.value]
@@ -233,6 +291,12 @@ module Nokolexbor
       end
     end
+    def write_to(io, *options)
+      io.write(to_html(*options))
+    end
+    alias_method :write_html_to, :write_to
     private
     def xpath_internal(node, paths, handler, ns, binds)

data/lib/nokolexbor/node_set.rb CHANGED Viewed

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 module Nokolexbor
-  class NodeSet < Node
+  class NodeSet < Nokolexbor::Node
     include Enumerable
     def self.new(document, list = [])
@@ -54,16 +54,17 @@ module Nokolexbor
     alias_method :inner_text, :content
     alias_method :to_str, :content
-    def inner_html
-      self.map(&:inner_html).join
+    def inner_html(*args)
+      self.map { |n| n.inner_html(*args) }.join
     end
-    def outer_html
-      self.map(&:outer_html).join
+    def outer_html(*args)
+      self.map { |n| n.outer_html(*args) }.join
     end
     alias_method :to_s, :outer_html
     alias_method :to_html, :outer_html
+    alias_method :serialize, :outer_html
     def remove
       self.each(&:remove)

data/lib/nokolexbor/version.rb CHANGED Viewed

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 module Nokolexbor
-  VERSION = '0.2.5'
+  VERSION = '0.3.1'
 end

data/lib/nokolexbor.rb CHANGED Viewed

@@ -1,6 +1,26 @@
 # frozen_string_literal: true
-require 'nokolexbor/nokolexbor'
+begin
+  # pre-compiled extension by rake-compiler is located inside lib/nokolexbor/<ruby_version>/
+  RUBY_VERSION =~ /(\d+\.\d+)/
+  require "nokolexbor/#{Regexp.last_match(1)}/nokolexbor"
+rescue LoadError => e
+  if /GLIBC/.match?(e.message)
+    warn(<<~EOM)
+      ERROR: It looks like you're trying to use Nokolexbor as a precompiled native gem on a system
+             with an unsupported version of glibc.
+        #{e.message}
+        If that's the case, then please install Nokolexbor via the `ruby` platform gem:
+            gem install nokolexbor --platform=ruby
+        or:
+            bundle config set force_ruby_platform true
+    EOM
+    raise e
+  end
+  require 'nokolexbor/nokolexbor'
+end
 require 'nokolexbor/version'
 require 'nokolexbor/node'
 require 'nokolexbor/document'

data/patches/0001-lexbor-support-text-pseudo-element.patch CHANGED Viewed

@@ -31,7 +31,7 @@ index 398d1bb..b35bfab 100644
      {"after", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_AFTER], 5, 0},
      {"first-letter", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_FIRST_LETTER], 12, 0},
 -    {NULL, NULL, 0, 0},
-+    {"text", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_TEXT], 4, 0},
++    {"text", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_TEXT], 4, 0},
      {NULL, NULL, 0, 0},
      {"grammar-error", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_GRAMMAR_ERROR], 13, 0},
      {"before", (void *) &lxb_css_selectors_pseudo_data_pseudo_element[LXB_CSS_SELECTOR_PSEUDO_ELEMENT_BEFORE], 6, 0},

metadata CHANGED Viewed

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nokolexbor
 version: !ruby/object:Gem::Version
-  version: 0.2.5
+  version: 0.3.1
 platform: ruby
 authors:
 - Yicheng Zhou
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-12-20 00:00:00.000000000 Z
+date: 2022-12-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler
@@ -38,7 +38,7 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '5.0'
-description: Nokolexbor is a high performance HTML5 parser, with support for both
+description: Nokolexbor is a high-performance HTML5 parser, with support for both
   CSS selectors and XPath. It's API is designed to be compatible with Nokogiri.
 email: zyc9012@gmail.com
 executables: []
@@ -79,9 +79,12 @@ files:
 - ext/nokolexbor/libxml/xpathInternals.h
 - ext/nokolexbor/libxml/xpointer.h
 - ext/nokolexbor/memory.c
+- ext/nokolexbor/nl_cdata.c
+- ext/nokolexbor/nl_comment.c
 - ext/nokolexbor/nl_document.c
 - ext/nokolexbor/nl_node.c
 - ext/nokolexbor/nl_node_set.c
+- ext/nokolexbor/nl_text.c
 - ext/nokolexbor/nl_xpath_context.c
 - ext/nokolexbor/nokolexbor.c
 - ext/nokolexbor/nokolexbor.h
@@ -555,5 +558,5 @@ requirements: []
 rubygems_version: 3.0.3.1
 signing_key:
 specification_version: 4
-summary: High performance HTML5 parser, with support for both CSS selectors and XPath.
+summary: High-performance HTML5 parser, with support for both CSS selectors and XPath.
 test_files: []