From 45b0cdf8114d361018dc4f49d29e7919e57ddf15 Mon Sep 17 00:00:00 2001
From: Yorick Peterse <yorickpeterse@gmail.com>
Date: Mon, 28 Apr 2014 23:40:36 +0200
Subject: [PATCH] Track element name nesting in the pull parser.

Tracking the names of nested elements makes it a lot easier to do contextual
pull parsing. Without this it's impossible to know what context the parser is
in at a given moment.

For memory reasons the parser currently only tracks the element names. In the
future it might perhaps also track extra information to make parsing easier.
---
 lib/oga/xml/parser.y                          | 16 +++++++-
 lib/oga/xml/pull_parser.rb                    | 37 +++++++++++++++++--
 .../xml/pull_parser/element_nesting_spec.rb   | 23 ++++++++++++
 3 files changed, 71 insertions(+), 5 deletions(-)
 create mode 100644 spec/oga/xml/pull_parser/element_nesting_spec.rb

diff --git a/lib/oga/xml/parser.y b/lib/oga/xml/parser.y
index 0492912..66687f5 100644
--- a/lib/oga/xml/parser.y
+++ b/lib/oga/xml/parser.y
@@ -103,7 +103,11 @@ rule
   element
     : element_start expressions T_ELEM_END
       {
-        on_element_children(val[0], val[1] ? val[1].flatten : [])
+        if val[0]
+          on_element_children(val[0], val[1] ? val[1].flatten : [])
+        end
+
+        after_element(val[0])
       }
     ;
 
@@ -149,6 +153,8 @@ end
   def initialize(data, options = {})
     @data  = data
     @lexer = Lexer.new(data, options)
+
+    reset
   end
 
   ##
@@ -341,6 +347,14 @@ Unexpected #{name} with value #{value.inspect} on line #{@line}:
     return element
   end
 
+  ##
+  # @param [Oga::XML::Element]
+  # @return [Oga::XML::Element]
+  #
+  def after_element(element)
+    return element
+  end
+
   ##
   # @param [Array] pairs
   # @return [Hash]
diff --git a/lib/oga/xml/pull_parser.rb b/lib/oga/xml/pull_parser.rb
index 457625e..7bd55e3 100644
--- a/lib/oga/xml/pull_parser.rb
+++ b/lib/oga/xml/pull_parser.rb
@@ -19,7 +19,13 @@ module Oga
     # This parses yields proper XML instances such as {Oga::XML::Element}.
     # Doctypes and XML declarations are ignored by this parser.
     #
+    # @!attribute [r] nesting
+    #  Array containing the names of the currently nested elements.
+    #  @return [Array]
+    #
     class PullParser < Parser
+      attr_reader :nesting
+
       ##
       # @return [Array]
       #
@@ -36,8 +42,7 @@ module Oga
       BLOCK_CALLBACKS = [
         :on_cdata,
         :on_comment,
-        :on_text,
-        :on_element
+        :on_text
       ]
 
       ##
@@ -46,7 +51,8 @@ module Oga
       def reset
         super
 
-        @block = nil
+        @block   = nil
+        @nesting = []
       end
 
       ##
@@ -68,7 +74,7 @@ module Oga
       # JRuby.
       DISABLED_CALLBACKS.each do |method|
         eval <<-EOF, nil, __FILE__, __LINE__ + 1
-        def #{method}(*_)
+        def #{method}(*args)
           return
         end
         EOF
@@ -78,9 +84,32 @@ module Oga
         eval <<-EOF, nil, __FILE__, __LINE__ + 1
         def #{method}(*args)
           @block.call(super)
+          return
         end
         EOF
       end
+
+      ##
+      # @see Oga::XML::Parser#on_element
+      #
+      def on_element(*args)
+        element = super
+
+        nesting << element.name
+
+        @block.call(element)
+
+        return
+      end
+
+      ##
+      # @see Oga::XML::Parser#on_element_children
+      #
+      def after_element(*args)
+        nesting.pop
+
+        return
+      end
     end # PullParser
   end # XML
 end # Oga
diff --git a/spec/oga/xml/pull_parser/element_nesting_spec.rb b/spec/oga/xml/pull_parser/element_nesting_spec.rb
new file mode 100644
index 0000000..d08e2b2
--- /dev/null
+++ b/spec/oga/xml/pull_parser/element_nesting_spec.rb
@@ -0,0 +1,23 @@
+require 'spec_helper'
+
+describe Oga::XML::PullParser do
+  context 'tracking element nesting' do
+    before do
+      @parser = described_class.new('<a><b></b></a>')
+    end
+
+    example 'set the nesting for the outer element' do
+      @parser.parse do |node|
+        @parser.nesting.should == %w{a} if node.name == 'a'
+
+        @parser.nesting.should == %w{a b} if node.name == 'b'
+      end
+    end
+
+    example 'pop element names after leaving an element' do
+      @parser.nesting.should_receive(:pop).twice
+
+      @parser.parse { |node| }
+    end
+  end
+end