Merge pull request #171 from rubys/flavorjones-check-for-html5-alread…

…y-defined

feat: Nokogumbo detects Nokogiri's HTML5 API

.github/workflows/ci.yml

@@ -64,6 +64,10 @@ jobs:
       if: matrix.os == 'ubuntu'
       run: sudo apt-get install -y ragel
 
+    - name: Install ragel
+      if: matrix.os == 'macos'
+      run: brew install ragel
+
     - name: Test Gumbo and gem packaging
       shell: bash
       run: ./scripts/ci-package-test.sh

CHANGELOG.md

@@ -18,6 +18,9 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Support Mageia distros when libxml2/libxslt system libraries are install. #165 (Thank you,
   @pterjan!)
 
+### Addded
+- Forward-looking support for a version of Nokogiri that will provide HTML5 parsing. #171
+
 ### Improved
 - Update extconf.rb to use Nokogiri v1.11's CPPFLAGS for more reliable installation.
 

lib/nokogumbo.rb

@@ -1,17 +1,27 @@
 require 'nokogiri'
-require 'nokogumbo/version'
-require 'nokogumbo/html5'
 
-require 'nokogumbo/nokogumbo'
+if ((defined?(Nokogiri::HTML5) && Nokogiri::HTML5.respond_to?(:parse)) &&
+    (defined?(Nokogiri::Gumbo) && Nokogiri::Gumbo.respond_to?(:parse)) &&
+    !(ENV.key?("NOKOGUMBO_IGNORE_NOKOGIRI_HTML5") && ENV["NOKOGUMBO_IGNORE_NOKOGIRI_HTML5"] != "false"))
+
+  warn "NOTE: nokogumbo: Using Nokogiri::HTML5 provided by Nokogiri. See https://github.com/sparklemotion/nokogiri/issues/2205 for more information."
+
+  ::Nokogumbo = ::Nokogiri::Gumbo
+else
+  require 'nokogumbo/html5'
+  require 'nokogumbo/nokogumbo'
 
-module Nokogumbo
-  # The default maximum number of attributes per element.
-  DEFAULT_MAX_ATTRIBUTES = 400
+  module Nokogumbo
+    # The default maximum number of attributes per element.
+    DEFAULT_MAX_ATTRIBUTES = 400
 
-  # The default maximum number of errors for parsing a document or a fragment.
-  DEFAULT_MAX_ERRORS = 0
+    # The default maximum number of errors for parsing a document or a fragment.
+    DEFAULT_MAX_ERRORS = 0
 
-  # The default maximum depth of the DOM tree produced by parsing a document
-  # or fragment.
-  DEFAULT_MAX_TREE_DEPTH = 400
+    # The default maximum depth of the DOM tree produced by parsing a document
+    # or fragment.
+    DEFAULT_MAX_TREE_DEPTH = 400
+  end
 end
+
+require 'nokogumbo/version'