Added HTML::Mixin#html_open (closes #12).

ronin-rb · Oct 13, 2023 · eecec10 · eecec10
1 parent 802c268
commit eecec10
Show file tree

Hide file tree

Showing 2 changed files with 62 additions and 0 deletions.
diff --git a/lib/ronin/support/web/html/mixin.rb b/lib/ronin/support/web/html/mixin.rb
@@ -53,6 +53,35 @@ def html_parse(html,&block)
             HTML.parse(html,&block)
           end
 
+          #
+          # Opens an HTML file.
+          #
+          # @param [String] path
+          #   The path to the HTML file.
+          #
+          # @yield [doc]
+          #   If a block is given, it will be passed the newly created document
+          #   object.
+          #
+          # @yieldparam [Nokogiri::HTML::Document] doc
+          #   The new HTML document object.
+          #
+          # @return [Nokogiri::HTML::Document]
+          #   The parsed HTML file.
+          #
+          # @example
+          #   doc = HTML.open('index.html')
+          #   # => #<Nokogiri::HTML::Document:...>
+          #
+          # @see http://rubydoc.info/gems/nokogiri/Nokogiri/HTML/Document
+          # @see HTML.open
+          #
+          def html_open(path,&block)
+            HTML.open(path,&block)
+          end
+
+          alias open_html html_open
+
           #
           # Creates a new `Nokogiri::HTML::Builder`.
           #

diff --git a/spec/html/mixin_spec.rb b/spec/html/mixin_spec.rb
@@ -45,6 +45,39 @@
     end
   end
 
+  describe "#html_open" do
+    it "must open and parse the given path, and return a Nokogiri::HTML::Document" do
+      doc = subject.html_open(html_file)
+
+      expect(doc).to be_kind_of(Nokogiri::HTML::Document)
+
+      # XXX: nokogiri's java extensions behave differently from libxml2
+      if RUBY_ENGINE == 'jruby'
+        expect(doc.to_s).to eq(
+          <<~HTML
+            <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
+            #{html}
+          HTML
+        )
+      else
+        expect(doc.to_s).to eq(
+          <<~HTML
+            <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
+            #{html.chomp}
+          HTML
+        )
+      end
+    end
+
+    context "when given a block" do
+      it "must yield the Nokogiri::HTML::Document object" do
+        expect { |b|
+          subject.html_open(html_file,&b)
+        }.to yield_with_args(Nokogiri::HTML::Document)
+      end
+    end
+  end
+
   describe "#html_build" do
     it "must build an HTML document" do
       doc = subject.html_build do |html|