puppet 4.0.0 → 4.9.1

data/spec/unit/functions/lookup_spec.rb

@@ -1,364 +1,1444 @@
 #! /usr/bin/env ruby
 require 'spec_helper'
 require 'puppet_spec/compiler'
+require 'puppet_spec/files'
 require 'puppet/pops'
+require 'deep_merge/core'
 
-describe "when performing lookup" do
+describe "The lookup function" do
   include PuppetSpec::Compiler
+  include PuppetSpec::Files
 
-  # Assembles code that includes the *abc* class and compiles it into a catalog. This class will use the global
-  # variable $args to perform a lookup and assign the result to $abc::result. Unless the $block is set to
-  # the string 'no_block_present', it will be passed as a lambda to the lookup. The assembled code will declare
-  # a notify resource with a name that is formed by interpolating the result into a format string.
-  #
-  # The method performs the folloging steps.
-  #
-  # - Build the code that:
-  #    - sets the $args variable from _lookup_args_
-  #    - sets the $block parameter to the given block or the string 'no_block_present'
-  #    - includes the abc class
-  #    - assigns the $abc::result to $r
-  #    - interpolates a string using _fmt_ (which is assumed to use $r)
-  #    - declares a notify resource from the interpolated string
-  # - Compile the code into a catalog
-  # - Return the name of all Notify resources in that catalog
-  #
-  # @param fmt [String] The puppet interpolated string used when creating the notify title
-  # @param *args [String] splat of args that will be concatenated to form the puppet args sent to lookup
-  # @return [Array<String>] List of names of Notify resources in the resulting catalog
-  #
-  def assemble_and_compile(fmt, *lookup_args)
-    assemble_and_compile_with_block(fmt, "'no_block_present'", *lookup_args)
-  end
+  context 'with an environment' do
+    let(:env_name) { 'spec' }
+    let(:code_dir_files) { {} }
+    let(:code_dir) { tmpdir('code') }
+    let(:environment_files) do
+      {
+        env_name => {
+          'modules' => {},
+          'hiera.yaml' => <<-YAML.unindent,
+          ---
+          version: 5
+          hierarchy:
+            - name: "Common"
+              data_hash: yaml_data
+              path: "common.yaml"
+        YAML
+        'data' => {
+          'common.yaml' => <<-YAML.unindent
+            ---
+            a: value a
+            mod_a::a: value mod_a::a (from environment)
+            mod_a::hash_a:
+              a: value mod_a::hash_a.a (from environment)
+            mod_a::hash_b:
+              a: value mod_a::hash_b.a (from environment)
+            hash_b:
+              hash_ba:
+                bab: value hash_b.hash_ba.bab (from environment)
+            hash_c:
+              hash_ca:
+                caa: value hash_c.hash_ca.caa (from environment)
+            lookup_options:
+              mod_a::hash_b:
+                merge: hash
+              hash_c:
+                merge: hash
+            YAML
+          }
+        }
+      }
+    end
 
-  def assemble_and_compile_with_block(fmt, block, *lookup_args)
-    Puppet[:code] = <<-END.gsub(/^ {6}/, '')
-      $args = [#{lookup_args.join(',')}]
-      $block = #{block}
-      include abc
-      $r = if $abc::result == undef { 'no_value' } else { $abc::result }
-      notify { \"#{fmt}\": }
-    END
-    compiler.compile().resources.map(&:ref).select { |r| r.start_with?('Notify[') }.map { |r| r[7..-2] }
-  end
+    let(:logs) { [] }
+    let(:notices) { logs.select { |log| log.level == :notice }.map { |log| log.message } }
+    let(:warnings) { logs.select { |log| log.level == :warning }.map { |log| log.message } }
+    let(:debugs) { logs.select { |log| log.level == :debug }.map { |log| log.message } }
+    let(:env) { Puppet::Node::Environment.create(env_name.to_sym, [File.join(populated_env_dir, env_name, 'modules')]) }
+    let(:environments) { Puppet::Environments::Directories.new(populated_env_dir, []) }
+    let(:node) { Puppet::Node.new('test_lookup', :environment => env) }
+    let(:compiler) { Puppet::Parser::Compiler.new(node) }
+    let(:lookup_func) { Puppet.lookup(:loaders).puppet_system_loader.load(:function, 'lookup') }
+    let(:defaults) {
+      {
+        'mod_a::xd' => 'value mod_a::xd (from default)',
+        'mod_a::xd_found' => 'value mod_a::xd_found (from default)',
+        'scope_xd' => 'value scope_xd (from default)'
+      }}
+    let(:overrides) {
+      {
+        'mod_a::xo' => 'value mod_a::xo (from override)',
+        'scope_xo' => 'value scope_xo (from override)'
+      }}
+    let(:invocation_with_explain) { Puppet::Pops::Lookup::Invocation.new(compiler.topscope, {}, {}, true) }
+    let(:explanation) { invocation_with_explain.explainer.explain }
 
-  # There is a fully configured 'production' environment in fixtures at this location
-  let(:environmentpath) { File.join(my_fixture_dir, 'environments') }
-  let(:node) { Puppet::Node.new("testnode", :facts => Puppet::Node::Facts.new("facts", {}), :environment => 'production') }
-  let(:compiler) { Puppet::Parser::Compiler.new(node) }
-
-  around(:each) do |example|
-    # Initialize settings to get a full compile as close as possible to a real
-    # environment load
-    Puppet.settings.initialize_global_settings
-
-    # Initialize loaders based on the environmentpath. It does not work to
-    # just set the setting environmentpath for some reason - this achieves the same:
-    # - first a loader is created, loading directory environments from the fixture (there is
-    # one environment, 'sample', which will be loaded since the node references this
-    # environment by name).
-    # - secondly, the created env loader is set as 'environments' in the puppet context.
-    #
-    environments = Puppet::Environments::Directories.new(environmentpath, [])
-    Puppet.override(:environments => environments) do
-      example.run
+    let(:populated_code_dir) do
+      dir_contained_in(code_dir, code_dir_files)
+      code_dir
     end
-  end
 
-  context 'using normal parameters' do
-    it 'can lookup value provided by the environment' do
-      resources = assemble_and_compile('${r}', "'abc::a'")
-      expect(resources).to include('env_a')
+    let(:env_dir) do
+      d = File.join(populated_code_dir, 'environments')
+      Dir.mkdir(d)
+      d
     end
 
-    it 'can lookup value provided by the module' do
-      resources = assemble_and_compile('${r}', "'abc::b'")
-      expect(resources).to include('module_b')
+    let(:populated_env_dir) do
+      dir_contained_in(env_dir, environment_files)
+      env_dir
     end
 
-    it 'can lookup value provided in global scope' do
-      Hiera.any_instance.expects(:lookup).with('abc::a', any_parameters).returns('global_a')
-      resources = assemble_and_compile('${r}', "'abc::a'")
-      expect(resources).to include('global_a')
+    before(:each) do
+      Puppet.settings[:codedir] = code_dir
+      Puppet.push_context(:environments => environments, :current_environment => env)
     end
 
-    it 'will stop at first found name when several names are provided' do
-      resources = assemble_and_compile('${r}', "['abc::b', 'abc::a']")
-      expect(resources).to include('module_b')
+    after(:each) do
+      Puppet.pop_context
+      if Object.const_defined?(:Hiera)
+        Hiera.send(:remove_instance_variable, :@config) if Hiera.instance_variable_defined?(:@config)
+        Hiera.send(:remove_instance_variable, :@logger) if Hiera.instance_variable_defined?(:@logger)
+        if Hiera.const_defined?(:Config)
+          Hiera::Config.send(:remove_instance_variable, :@config) if Hiera::Config.instance_variable_defined?(:@config)
+        end
+        if Hiera.const_defined?(:Backend)
+          Hiera::Backend.clear!
+        end
+      end
     end
 
-    it 'can lookup value provided by the module that is overriden by environment' do
-      resources = assemble_and_compile('${r}', "'abc::c'")
-      expect(resources).to include('env_c')
+    def collect_notices(code, explain = false, &block)
+      Puppet[:code] = code
+      Puppet::Util::Log.with_destination(Puppet::Test::LogCollector.new(logs)) do
+        scope = compiler.topscope
+        scope['environment'] = env_name
+        scope['domain'] = 'example.com'
+        scope['scope_scalar'] = 'scope scalar value'
+        scope['scope_hash'] = { 'a' => 'scope hash a', 'b' => 'scope hash b' }
+        if explain
+          begin
+            invocation_with_explain.lookup('dummy', nil) do
+              if block_given?
+                compiler.compile { |catalog| block.call(compiler.topscope); catalog }
+              else
+                compiler.compile
+              end
+            end
+          rescue Puppet::DataBinding::LookupError => e
+            invocation_with_explain.report_text { e.message }
+          end
+        else
+          if block_given?
+            compiler.compile { |catalog| block.call(compiler.topscope); catalog }
+          else
+            compiler.compile
+          end
+        end
+      end
+      nil
     end
 
-    it "can 'unique' merge values provided by both the module and the environment" do
-      resources = assemble_and_compile('${r[0]}_${r[1]}', "'abc::c'", 'Array[String]', "'unique'")
-      expect(resources).to include('env_c_module_c')
+    def lookup(key, options = {}, explain = false)
+      nc_opts = options.empty? ? '' : ", #{Puppet::Pops::Types::TypeFormatter.string(options)}"
+      keys = key.is_a?(Array) ? key : [key]
+      collect_notices(keys.map { |k| "notice(String(lookup('#{k}'#{nc_opts}), '%p'))" }.join("\n"), explain)
+      if explain
+        explanation
+      else
+        result = notices.map { |n| Puppet::Pops::Types::TypeParser.singleton.parse_literal(n) }
+        key.is_a?(Array) ? result : result[0]
+      end
     end
 
-    it "can 'hash' merge values provided by the environment only" do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::d'", 'Hash[String,String]', "'hash'")
-      expect(resources).to include('env_d1_env_d2_env_d3')
+    def explain(key, options = {})
+      lookup(key, options, true)[1]
+      explanation
     end
 
-    it "can 'hash' merge values provided by both the environment and the module" do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "'hash'")
-      expect(resources).to include('env_e1_module_e2_env_e3')
+    it 'finds data in the environment' do
+      expect(lookup('a')).to eql('value a')
     end
 
-    it "can 'hash' merge values provided by global, environment, and module" do
-      Hiera.any_instance.expects(:lookup).with('abc::e', any_parameters).returns({ 'k1' => 'global_e1' })
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "'hash'")
-      expect(resources).to include('global_e1_module_e2_env_e3')
-    end
+    context 'that has no lookup configured' do
+      let(:environment_files) do
+        {
+          env_name => {
+            'modules' => {},
+            'data' => {
+              'common.yaml' => <<-YAML.unindent
+              ---
+              a: value a
+            YAML
+            }
+          }
+        }
+      end
 
-    it "can pass merge parameter in the form of a hash with a 'strategy=>unique'" do
-      resources = assemble_and_compile('${r[0]}_${r[1]}', "'abc::c'", 'Array[String]', "{strategy => 'unique'}")
-      expect(resources).to include('env_c_module_c')
-    end
+      it 'does not find data in the environment' do
+        expect { lookup('a') }.to raise_error(Puppet::DataBinding::LookupError, /did not find a value for the name 'a'/)
+      end
 
-    it "can pass merge parameter in the form of a hash with 'strategy=>hash'" do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "{strategy => 'hash'}")
-      expect(resources).to include('env_e1_module_e2_env_e3')
-    end
+      context "but an environment.conf with 'environment_data_provider=hiera'" do
+        let(:environment_files_1) do
+          DeepMerge.deep_merge!(environment_files, 'environment.conf' => "environment_data_provider=hiera\n")
+        end
 
-    it "can pass merge parameter in the form of a hash with a 'strategy=>deep'" do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "{strategy => 'deep'}")
-      expect(resources).to include('env_e1_module_e2_env_e3')
-    end
+        let(:populated_env_dir) do
+          dir_contained_in(env_dir, DeepMerge.deep_merge!(environment_files, env_name => environment_files_1))
+          env_dir
+        end
 
-    it "will fail unless merge in the form of a hash contains a 'strategy'" do
-      expect do
-        assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "{merge_key => 'hash'}")
-      end.to raise_error(Puppet::ParseError, /hash given as 'merge' must contain the name of a strategy/)
-    end
+        it 'finds data in the environment and reports deprecation warning for environment.conf' do
+          expect(lookup('a')).to eql('value a')
+          expect(warnings).to include(/Defining environment_data_provider='hiera' in environment.conf is deprecated. A 'hiera.yaml' file should be used instead/)
+        end
 
-    it 'will raise an exception when value is not found for single key and no default is provided' do
-      expect do
-        assemble_and_compile('${r}', "'abc::x'")
-      end.to raise_error(Puppet::ParseError, /did not find a value for the name 'abc::x'/)
-    end
+        context 'and a hiera.yaml file' do
+          let(:environment_files_2) { DeepMerge.deep_merge!(environment_files_1,'hiera.yaml' => <<-YAML.unindent) }
+            ---
+            version: 4
+            hierarchy:
+              - name: common
+                backend: yaml
+            YAML
 
-    it 'can lookup an undef value' do
-      resources = assemble_and_compile('${r}', "'abc::n'")
-      expect(resources).to include('no_value')
-    end
+          let(:populated_env_dir) do
+            dir_contained_in(env_dir, DeepMerge.deep_merge!(environment_files, env_name => environment_files_2))
+            env_dir
+          end
 
-    it 'will not replace an undef value with a given default' do
-      resources = assemble_and_compile('${r}', "'abc::n'", 'undef', 'undef', '"default_n"')
-      expect(resources).to include('no_value')
-    end
+          it 'finds data in the environment and reports deprecation warnings for both environment.conf and hiera.yaml' do
+            expect(lookup('a')).to eql('value a')
+            expect(warnings).to include(/Defining environment_data_provider='hiera' in environment.conf is deprecated/)
+            expect(warnings).to include(/Use of 'hiera.yaml' version 4 is deprecated. It should be converted to version 5/)
+          end
+        end
+      end
 
-    it 'will not accept a succesful lookup of an undef value when the type rejects it' do
-      expect do
-        assemble_and_compile('${r}', "'abc::n'", 'String')
-      end.to raise_error(Puppet::ParseError, /found value has wrong type/)
-    end
+      context "but an environment.conf with 'environment_data_provider=function'" do
+        let(:environment_files) do
+          {
+            env_name => {
+              'environment.conf' => "environment_data_provider=function\n",
+              'functions' => {
+                'environment' => { 'data.pp' => <<-PUPPET.unindent }
+                    function environment::data() {
+                      { 'a' => 'value a' }
+                    }
+                    PUPPET
+              }
+            }
+          }
+        end
 
-    it 'will raise an exception when value is not found for array key and no default is provided' do
-      expect do
-        assemble_and_compile('${r}', "['abc::x', 'abc::y']")
-      end.to raise_error(Puppet::ParseError, /did not find a value for any of the names \['abc::x', 'abc::y'\]/)
+        it 'finds data in the environment and reports deprecation warning for environment.conf' do
+          expect(lookup('a')).to eql('value a')
+          expect(warnings).to include(/Defining environment_data_provider='function' in environment.conf is deprecated. A 'hiera.yaml' file should be used instead/)
+          expect(warnings).to include(/Using of legacy data provider function 'environment::data'. Please convert to a 'data_hash' function/)
+        end
+      end
     end
 
-    it 'can lookup and deep merge shallow values provided by the environment only' do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::d'", 'Hash[String,String]', "'deep'")
-      expect(resources).to include('env_d1_env_d2_env_d3')
-    end
+    context 'that has interpolated paths configured' do
+      let(:environment_files) do
+        {
+          env_name => {
+            'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: "Varying"
+                  data_hash: yaml_data
+                  path: "x%{::var}.yaml"
+              YAML
+            'modules' => {},
+            'data' => {
+              'x.yaml' => <<-YAML.unindent,
+                y: value y from x
+              YAML
+              'x_d.yaml' => <<-YAML.unindent
+                y: value y from x_d
+              YAML
+            }
+          }
+        }
+      end
 
-    it 'can lookup and deep merge shallow values provided by both the module and the environment' do
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "'deep'")
-      expect(resources).to include('env_e1_module_e2_env_e3')
+      it 'reloads the configuration if interpolated values change' do
+        Puppet[:log_level] = 'debug'
+        collect_notices("notice('success')") do |scope|
+          expect(lookup_func.call(scope, 'y')).to eql('value y from x')
+          scope['var'] = '_d'
+          expect(lookup_func.call(scope, 'y')).to eql('value y from x_d')
+        end
+        expect(notices).to eql(['success'])
+        expect(debugs.any? { |m| m =~ /Hiera configuration recreated due to change of scope variables used in interpolation expressions/ }).to be_truthy
+      end
     end
 
-    it 'can lookup and deep merge deep values provided by global, environment, and module' do
-      Hiera.any_instance.expects(:lookup).with('abc::f', any_parameters).returns({ 'k1' => { 's1' => 'global_f11' }, 'k2' => { 's3' => 'global_f23' }})
-      resources = assemble_and_compile('${r[k1][s1]}_${r[k1][s2]}_${r[k1][s3]}_${r[k2][s1]}_${r[k2][s2]}_${r[k2][s3]}', "'abc::f'", 'Hash[String,Hash[String,String]]', "'deep'")
-      expect(resources).to include('global_f11_env_f12_module_f13_env_f21_module_f22_global_f23')
-    end
+    context 'that uses reserved option' do
+      let(:environment_files) do
+        {
+          env_name => {
+            'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: "Illegal"
+                  options:
+                    #{opt_spec}
+                  data_hash: yaml_data
+              YAML
+            'data' => {
+              'foo.yaml' => "a: The value a\n"
+            }
+          }
+        }
+      end
 
-    it 'will propagate resolution_type :array to Hiera when merge == \'unique\''  do
-      Hiera.any_instance.expects(:lookup).with('abc::c', anything, anything, anything, :array).returns(['global_c'])
-      resources = assemble_and_compile('${r[0]}_${r[1]}_${r[2]}', "'abc::c'", 'Array[String]', "'unique'")
-      expect(resources).to include('global_c_env_c_module_c')
+      context 'path' do
+        let(:opt_spec) { 'path: data/foo.yaml' }
+
+        it 'fails and reports the reserved option key' do
+          expect { lookup('a') }.to raise_error do |e|
+            expect(e.message).to match(/Option key 'path' used in hierarchy 'Illegal' is reserved by Puppet/)
+          end
+        end
+      end
+
+      context 'uri' do
+        let(:opt_spec) { 'uri: file:///data/foo.yaml' }
+
+        it 'fails and reports the reserved option key' do
+          expect { lookup('a') }.to raise_error do |e|
+            expect(e.message).to match(/Option key 'uri' used in hierarchy 'Illegal' is reserved by Puppet/)
+          end
+        end
+      end
     end
 
-    it 'will propagate a Hash resolution_type with :behavior => :native to Hiera when merge == \'hash\''  do
-      Hiera.any_instance.expects(:lookup).with('abc::e', anything, anything, anything, { :behavior => :native }).returns({ 'k1' => 'global_e1' })
-      resources = assemble_and_compile('${r[k1]}_${r[k2]}_${r[k3]}', "'abc::e'", 'Hash[String,String]', "{strategy => 'hash'}")
-      expect(resources).to include('global_e1_module_e2_env_e3')
+
+    context 'with lookup_options configured using patterns' do
+      let(:mod_common) {
+        <<-YAML.unindent
+          mod::hash_a:
+            aa:
+              aaa: aaa (from module)
+            ab:
+              aba: aba (from module)
+          mod::hash_b:
+            ba:
+              baa: baa (from module)
+            bb:
+              bba: bba (from module)
+          lookup_options:
+            '^mod::ha.*_a':
+              merge: deep
+            '^mod::ha.*_b':
+              merge: deep
+        YAML
+      }
+
+      let(:mod_base) do
+        {
+          'hiera.yaml' => <<-YAML.unindent,
+            version: 5
+            YAML
+          'data' => {
+            'common.yaml' => mod_common
+          }
+        }
+      end
+
+      let(:environment_files) do
+        {
+          env_name => {
+            'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: X
+                  paths:
+                  - first.yaml
+                  - second.yaml
+              YAML
+            'data' => {
+              'first.yaml' => <<-YAML.unindent,
+                a:
+                  aa:
+                    aaa: a.aa.aaa
+                b:
+                  ba:
+                    baa: b.ba.baa
+                  bb:
+                    bba: b.bb.bba
+                c:
+                  ca:
+                    caa: c.ca.caa
+                mod::hash_a:
+                  aa:
+                    aab: aab (from environment)
+                  ab:
+                    aba: aba (from environment)
+                    abb: abb (from environment)
+                mod::hash_b:
+                  ba:
+                    bab: bab (from environment)
+                  bc:
+                    bca: bca (from environment)
+                lookup_options:
+                  b:
+                    merge: hash
+                  '^[^b]$':
+                     merge: deep
+                  '^c':
+                     merge: first
+                  '^b':
+                     merge: first
+                  '^mod::ha.*_b':
+                    merge: hash
+                YAML
+              'second.yaml' => <<-YAML.unindent,
+                a:
+                  aa:
+                    aab: a.aa.aab
+                b:
+                  ba:
+                    bab: b.ba.bab
+                  bb:
+                    bbb: b.bb.bbb
+                c:
+                  ca:
+                    cab: c.ca.cab
+                YAML
+            },
+            'modules' => {
+              'mod' => mod_base
+            }
+          }
+        }
+      end
+
+      it 'finds lookup_options that matches a pattern' do
+        expect(lookup('a')).to eql({'aa' => { 'aaa' => 'a.aa.aaa', 'aab' => 'a.aa.aab' }})
+      end
+
+      it 'gives a direct key match higher priority than a matching pattern' do
+        expect(lookup('b')).to eql({'ba' => { 'baa' => 'b.ba.baa' }, 'bb' => { 'bba'=>'b.bb.bba' }})
+      end
+
+      it 'uses the first matching pattern' do
+        expect(lookup('c')).to eql({'ca' => { 'caa' => 'c.ca.caa', 'cab' => 'c.ca.cab' }})
+      end
+
+      it 'uses lookup_option found by pattern from module' do
+        expect(lookup('mod::hash_a')).to eql({
+          'aa' => {
+            'aaa' => 'aaa (from module)',
+            'aab' => 'aab (from environment)'
+          },
+          'ab' => {
+            'aba' => 'aba (from environment)',
+            'abb' => 'abb (from environment)'
+          }
+        })
+      end
+
+      it 'merges lookup_options found by pattern in environment and module (environment wins)' do
+        expect(lookup('mod::hash_b')).to eql({
+          'ba' => {
+            'bab' => 'bab (from environment)'
+          },
+          'bb' => {
+            'bba' => 'bba (from module)'
+          },
+          'bc' => {
+            'bca' => 'bca (from environment)'
+          }
+        })
+      end
+
+      context 'and patterns in module are not limited to module keys' do
+        let(:mod_common) {
+          <<-YAML.unindent
+          mod::hash_a:
+            aa:
+              aaa: aaa (from module)
+            ab:
+              aba: aba (from module)
+          lookup_options:
+            '^.*_a':
+              merge: deep
+          YAML
+        }
+
+        it 'fails with error' do
+          expect { lookup('mod::a') }.to raise_error(Puppet::DataBinding::LookupError, /all lookup_options patterns must match a key starting with module name/)
+        end
+      end
     end
 
-    it 'will propagate a Hash resolution_type with :behavior => :deeper to Hiera when merge == \'deep\''  do
-      Hiera.any_instance.expects(:lookup).with('abc::f', anything, anything, anything, { :behavior => :deeper }).returns({ 'k1' => { 's1' => 'global_f11' }, 'k2' => { 's3' => 'global_f23' }})
-      resources = assemble_and_compile('${r[k1][s1]}_${r[k1][s2]}_${r[k1][s3]}_${r[k2][s1]}_${r[k2][s2]}_${r[k2][s3]}', "'abc::f'", 'Hash[String,Hash[String,String]]', "'deep'")
-      expect(resources).to include('global_f11_env_f12_module_f13_env_f21_module_f22_global_f23')
+    context 'and a global Hiera v4 configuration' do
+      let(:code_dir_files) do
+        {
+          'hiera.yaml' => <<-YAML.unindent,
+            ---
+            version: 4
+        YAML
+        }
+      end
+
+      before(:each) do
+        # Need to set here since spec_helper defines these settings in its "before each"
+        Puppet.settings[:codedir] = populated_code_dir
+        Puppet.settings[:hiera_config] = File.join(code_dir, 'hiera.yaml')
+      end
+
+      it 'raises an error' do
+        expect { lookup('a') }.to raise_error(Puppet::Error, /hiera configuration version 4 cannot be used in the global layer/)
+      end
     end
 
-    it 'will propagate a Hash resolution_type with symbolic deep merge options to Hiera'  do
-      Hiera.any_instance.expects(:lookup).with('abc::f', anything, anything, anything, { :behavior => :deeper, :knockout_prefix => '--' }).returns({ 'k1' => { 's1' => 'global_f11' }, 'k2' => { 's3' => 'global_f23' }})
-      resources = assemble_and_compile('${r[k1][s1]}_${r[k1][s2]}_${r[k1][s3]}_${r[k2][s1]}_${r[k2][s2]}_${r[k2][s3]}', "'abc::f'", 'Hash[String,Hash[String,String]]', "{ 'strategy' => 'deep', 'knockout_prefix' => '--' }")
-      expect(resources).to include('global_f11_env_f12_module_f13_env_f21_module_f22_global_f23')
+    context 'and an environment Hiera v3 configuration' do
+      let(:environment_files) do
+        {
+          env_name => {
+            'hiera.yaml' => <<-YAML.unindent,
+              ---
+              :backends: yaml
+          YAML
+          }
+        }
+      end
+
+      it 'raises an error' do
+        expect { lookup('a') }.to raise_error(Puppet::Error, /hiera configuration version 3 cannot be used in an environment/)
+      end
     end
 
-    context 'with provided default' do
-      it 'will return default when lookup fails' do
-        resources = assemble_and_compile('${r}', "'abc::x'", 'String', 'undef', "'dflt_x'")
-        expect(resources).to include('dflt_x')
+    context 'and a global empty Hiera configuration' do
+      let(:hiera_yaml_path) { File.join(code_dir, 'hiera.yaml') }
+      let(:code_dir_files) do
+        {
+          'hiera.yaml' => '',
+        }
       end
 
-      it 'can precede default parameter with undef as the value_type and undef as the merge type' do
-        resources = assemble_and_compile('${r}', "'abc::x'", 'undef', 'undef', "'dflt_x'")
-        expect(resources).to include('dflt_x')
+      let(:environment_files) do
+        {
+          env_name => {
+            'hieradata' => {
+              'common.yaml' =>  <<-YAML.unindent,
+                x: value x (from environment)
+                YAML
+            }
+          }
+        }
       end
 
-      it 'can use array' do
-        resources = assemble_and_compile('${r[0]}_${r[1]}', "'abc::x'", 'Array[String]', 'undef', "['dflt_x', 'dflt_y']")
-        expect(resources).to include('dflt_x_dflt_y')
+      before(:each) do
+        # Need to set here since spec_helper defines these settings in its "before each"
+        Puppet.settings[:hiera_config] = hiera_yaml_path
       end
 
-      it 'can use hash' do
-        resources = assemble_and_compile('${r[a]}_${r[b]}', "'abc::x'", 'Hash[String,String]', 'undef', "{'a' => 'dflt_x', 'b' => 'dflt_y'}")
-        expect(resources).to include('dflt_x_dflt_y')
+      it 'uses a Hiera version 3 defaults' do
+        expect(lookup('x')).to eql('value x (from environment)')
       end
 
-      it 'fails unless default is an instance of value_type' do
-        expect do
-          assemble_and_compile('${r[a]}_${r[b]}', "'abc::x'", 'Hash[String,String]', 'undef', "{'a' => 'dflt_x', 'b' => 32}")
-        end.to raise_error(Puppet::ParseError, /default_value value has wrong type/)
+      context 'obtained using /dev/null', :unless => Puppet.features.microsoft_windows? do
+        let(:code_dir_files) { {} }
+
+        it 'uses a Hiera version 3 defaults' do
+          Puppet[:hiera_config] = '/dev/null'
+          expect(lookup('x')).to eql('value x (from environment)')
+        end
       end
     end
 
-    context 'with a default block' do
-      it 'will be called when lookup fails' do
-        resources = assemble_and_compile_with_block('${r}', "'dflt_x'", "'abc::x'")
-        expect(resources).to include('dflt_x')
+    context 'and a global configuration' do
+      let(:hiera_yaml) do
+        <<-YAML.unindent
+        ---
+        :backends:
+          - yaml
+          - json
+          - custom
+          - hocon
+        :yaml:
+          :datadir: #{code_dir}/hieradata
+        :json:
+          :datadir: #{code_dir}/hieradata
+        :hocon:
+          :datadir: #{code_dir}/hieradata
+        :hierarchy:
+          - common
+          - "%{domain}"
+        :merge_behavior: deeper
+        YAML
       end
 
-      it 'will not called when lookup succeeds but the found value is nil' do
-        resources = assemble_and_compile_with_block('${r}', "'dflt_x'", "'abc::n'")
-        expect(resources).to include('no_value')
+      let(:code_dir_files) do
+        {
+          'hiera.yaml' => hiera_yaml,
+          'ruby_stuff' => {
+            'hiera' => {
+              'backend' => {
+                'custom_backend.rb' => <<-RUBY.unindent,
+                  class Hiera::Backend::Custom_backend
+                    def lookup(key, scope, order_override, resolution_type, context)
+                      case key
+                      when 'hash_c'
+                        { 'hash_ca' => { 'cad' => 'value hash_c.hash_ca.cad (from global custom)' }}
+                      when 'datasources'
+                        Hiera::Backend.datasources(scope, order_override) { |source| source }
+                      else
+                        throw :no_such_key
+                      end
+                    end
+                  end
+                  RUBY
+                'other_backend.rb' => <<-RUBY.unindent,
+                  class Hiera::Backend::Other_backend
+                    def lookup(key, scope, order_override, resolution_type, context)
+                      value = Hiera::Config[:other][key.to_sym]
+                      throw :no_such_key if value.nil?
+                      value
+                    end
+                  end
+                  RUBY
+              }
+            }
+          },
+          'hieradata' => {
+            'common.yaml' =>  <<-YAML.unindent,
+              a: value a (from global)
+              hash_b:
+                hash_ba:
+                  bab: value hash_b.hash_ba.bab (from global)
+              hash_c:
+                hash_ca:
+                  cab: value hash_c.hash_ca.cab (from global)
+              YAML
+            'example.com.yaml' =>  <<-YAML.unindent,
+              x: value x (from global example.com.yaml)
+              YAML
+            'common.json' =>  <<-JSON.unindent,
+              {
+                "hash_b": {
+                  "hash_ba": {
+                    "bac": "value hash_b.hash_ba.bac (from global json)"
+                  }
+                },
+                "hash_c": {
+                  "hash_ca": {
+                    "cac": "value hash_c.hash_ca.cac (from global json)"
+                  }
+                }
+              }
+              JSON
+            'common.conf' =>  <<-HOCON.unindent,
+              // The 'xs' is a value used for testing
+              xs = { subkey = value xs.subkey (from global hocon) }
+              HOCON
+          }
+        }
       end
 
-      it 'can use array' do
-        resources = assemble_and_compile_with_block('${r[0]}_${r[1]}', "['dflt_x', 'dflt_y']", "'abc::x'")
-        expect(resources).to include('dflt_x_dflt_y')
+      before(:each) do
+        # Need to set here since spec_helper defines these settings in its "before each"
+        Puppet.settings[:codedir] = populated_code_dir
+        Puppet.settings[:hiera_config] = File.join(code_dir, 'hiera.yaml')
       end
 
-      it 'can use hash' do
-        resources = assemble_and_compile_with_block('${r[a]}_${r[b]}', "{'a' => 'dflt_x', 'b' => 'dflt_y'}", "'abc::x'")
-        expect(resources).to include('dflt_x_dflt_y')
+      around(:each) do |example|
+        # Faking the load path to enable 'require' to load from 'ruby_stuff'. It removes the need for a static fixture
+        # for the custom backend
+        $LOAD_PATH.unshift(File.join(code_dir, 'ruby_stuff'))
+        begin
+          Puppet.override(:environments => environments, :current_environment => env) do
+            example.run
+          end
+        ensure
+          Hiera::Backend.send(:remove_const, :Custom_backend) if Hiera::Backend.const_defined?(:Custom_backend)
+          Hiera::Backend.send(:remove_const, :Other_backend) if Hiera::Backend.const_defined?(:Other_backend)
+          $LOAD_PATH.shift
+        end
       end
 
-      it 'can return undef from block' do
-        resources = assemble_and_compile_with_block('${r}', 'undef', "'abc::x'")
-        expect(resources).to include('no_value')
-      end
+      context 'version 3' do
+        it 'finds data in the environment and reports deprecation warnings for both environment.conf and hiera.yaml' do
+          expect(lookup('a')).to eql('value a (from global)')
+          expect(warnings).to include(/Use of 'hiera.yaml' version 3 is deprecated. It should be converted to version 5/)
+        end
+
+        it 'explain contains output from global layer' do
+          explanation = explain('a')
+          expect(explanation).to include('Global Data Provider (hiera configuration version 3)')
+          expect(explanation).to include('Hierarchy entry "yaml"')
+          expect(explanation).to include('Hierarchy entry "json"')
+          expect(explanation).to include('Found key: "a" value: "value a (from global)"')
+        end
 
-      it 'fails unless block returns an instance of value_type' do
-        expect do
-          assemble_and_compile_with_block('${r[a]}_${r[b]}', "{'a' => 'dflt_x', 'b' => 32}", "'abc::x'", 'Hash[String,String]')
-        end.to raise_error(Puppet::ParseError, /default_block value has wrong type/)
+        it 'uses the merge behavior specified in global hiera.yaml to merge only global backends' do
+          expect(lookup('hash_b')).to eql(
+            { 'hash_ba' => { 'bab' => 'value hash_b.hash_ba.bab (from global)', 'bac' => 'value hash_b.hash_ba.bac (from global json)' } })
+        end
+
+        it 'uses the merge from lookup options to merge all layers and override merge_behavior specified in global hiera.yaml' do
+          expect(lookup('hash_c')).to eql(
+            { 'hash_ca' => { 'cab' => 'value hash_c.hash_ca.cab (from global)' } })
+        end
+
+        it 'uses the explicitly given merge to override lookup options and to merge all layers' do
+          expect(lookup('hash_c', 'merge' => 'deep')).to eql(
+            {
+              'hash_ca' =>
+              {
+                'caa' => 'value hash_c.hash_ca.caa (from environment)',
+                'cab' => 'value hash_c.hash_ca.cab (from global)',
+                'cac' => 'value hash_c.hash_ca.cac (from global json)',
+                'cad' => 'value hash_c.hash_ca.cad (from global custom)'
+              }
+            })
+        end
+
+        it 'paths are interpolated' do
+          expect(lookup('x')).to eql('value x (from global example.com.yaml)')
+        end
+
+        it 'backend data sources are propagated to custom backend' do
+          expect(lookup('datasources')).to eql(['common', 'example.com'])
+        end
+
+        it 'delegates configured hocon backend to hocon_data function' do
+          expect(explain('xs')).to match(/Hierarchy entry "hocon"\n.*\n.*\n.*"common"\n\s*Found key: "xs"/m)
+        end
+
+        it 'can dig down into subkeys provided by hocon_data function' do
+          expect(lookup('xs.subkey')).to eql('value xs.subkey (from global hocon)')
+        end
+
+        context 'using deep_merge_options supported by deep_merge gem but not supported by Puppet' do
+
+          let(:hiera_yaml) do
+            <<-YAML.unindent
+              ---
+              :backends:
+                - yaml
+              :yaml:
+                :datadir: #{code_dir}/hieradata
+              :hierarchy:
+                - other
+                - common
+              :merge_behavior: deeper
+              :deep_merge_options:
+                :unpack_arrays: ','
+              YAML
+          end
+
+          let(:code_dir_files) do
+            {
+              'hiera.yaml' => hiera_yaml,
+              'hieradata' => {
+                'common.yaml' => <<-YAML.unindent,
+                  a:
+                    - x1,x2
+                  YAML
+                'other.yaml' => <<-YAML.unindent,
+                  a:
+                    - x3
+                    - x4
+                  YAML
+              }
+            }
+          end
+
+          it 'honors option :unpack_arrays: (unsupported by puppet)' do
+            expect(lookup('a')).to eql(%w(x1 x2 x3 x4))
+          end
+        end
+
+        context 'using relative datadir paths' do
+          let(:hiera_yaml) do
+            <<-YAML.unindent
+          ---
+          :backends:
+            - yaml
+          :yaml:
+            :datadir: relative_data
+          :hierarchy:
+            - common
+            YAML
+          end
+
+          let(:populated_code_dir) do
+            dir_contained_in(code_dir, code_dir_files.merge({
+              'fake_cwd' => {
+                'relative_data' => {
+                  'common.yaml' => <<-YAML.unindent
+                    a: value a (from fake_cwd/relative_data/common.yaml)
+                  YAML
+                }
+              }
+            }))
+            code_dir
+          end
+
+          around(:each) do |example|
+            cwd = Dir.pwd
+            Dir.chdir(File.join(code_dir, 'fake_cwd'))
+            begin
+              example.run
+            ensure
+              Dir.chdir(cwd)
+            end
+          end
+
+          it 'finds data from data file beneath relative datadir' do
+            expect(lookup('a')).to eql('value a (from fake_cwd/relative_data/common.yaml)')
+          end
+        end
       end
 
-      it 'receives a single name parameter' do
-        resources = assemble_and_compile_with_block('${r}', 'true', "'name_x'")
-        expect(resources).to include('name_x')
+      context 'version 5' do
+        let(:hiera_yaml) do
+          <<-YAML.unindent
+          ---
+          version: 5
+          defaults:
+            datadir: hieradata
+
+          hierarchy:
+            - name: Yaml
+              data_hash: yaml_data
+              paths:
+                - common.yaml
+                - "%{domain}.yaml"
+            - name: Json
+              data_hash: json_data
+              paths:
+                - common.json
+                - "%{domain}.json"
+            - name: Hocon
+              data_hash: hocon_data
+              paths:
+                - common.conf
+                - "%{domain}.conf"
+            - name: Custom
+              hiera3_backend: custom
+              paths:
+                - common.custom
+                - "%{domain}.custom"
+            - name: Other
+              hiera3_backend: other
+              options:
+                other_option: value of other_option
+              paths:
+                - common.other
+                - "%{domain}.other"
+              YAML
+        end
+
+        it 'finds data in the environment and reports no deprecation warnings' do
+          expect(lookup('a')).to eql('value a (from global)')
+          expect(warnings).to be_empty
+        end
+
+        it 'explain contains output from global layer' do
+          explanation = explain('a')
+          expect(explanation).to include('Global Data Provider (hiera configuration version 5)')
+          expect(explanation).to include('Hierarchy entry "Yaml"')
+          expect(explanation).to include('Hierarchy entry "Json"')
+          expect(explanation).to include('Hierarchy entry "Hocon"')
+          expect(explanation).to include('Hierarchy entry "Custom"')
+          expect(explanation).to include('Found key: "a" value: "value a (from global)"')
+        end
+
+        it 'uses the explicitly given merge to override lookup options and to merge all layers' do
+          expect(lookup('hash_c', 'merge' => 'deep')).to eql(
+            {
+              'hash_ca' =>
+                {
+                  'caa' => 'value hash_c.hash_ca.caa (from environment)',
+                  'cab' => 'value hash_c.hash_ca.cab (from global)',
+                  'cac' => 'value hash_c.hash_ca.cac (from global json)',
+                  'cad' => 'value hash_c.hash_ca.cad (from global custom)'
+                }
+            })
+        end
+
+        it 'backend data sources are propagated to custom backend' do
+          expect(lookup('datasources')).to eql(['common', 'example.com'])
+        end
+
+        it 'backend specific options are propagated to custom backend' do
+          expect(lookup('other_option')).to eql('value of other_option')
+        end
+
+        it 'multiple hiera3_backend declarations can be used and are merged into the generated config' do
+          expect(lookup(['datasources', 'other_option'])).to eql([['common', 'example.com'], 'value of other_option'])
+          expect(Hiera::Config.instance_variable_get(:@config)).to eql(
+            {
+              :backends => ['custom', 'other'],
+              :hierarchy => ['common', '%{domain}'],
+              :custom => { :datadir => "#{code_dir}/hieradata" },
+              :other => { :other_option => 'value of other_option', :datadir=>"#{code_dir}/hieradata" },
+              :logger => 'puppet'
+            })
+        end
+
+        it 'provides a sensible error message when the hocon library is not loaded' do
+          Puppet.features.stubs(:hocon?).returns(false)
+
+          expect { lookup('a') }.to raise_error do |e|
+            expect(e.message).to match(/Lookup using Hocon data_hash function is not supported without hocon library/)
+          end
+        end
       end
 
-      it 'receives an array name parameter' do
-        resources = assemble_and_compile_with_block('${r[0]}_${r[1]}', 'true', "['name_x', 'name_y']")
-        expect(resources).to include('name_x_name_y')
+      context 'with a hiera3_backend that has no paths' do
+        let(:hiera_yaml) do
+          <<-YAML.unindent
+          ---
+          version: 5
+          hierarchy:
+            - name: Custom
+              hiera3_backend: custom
+          YAML
+        end
+
+        it 'calls the backend' do
+          expect(lookup('hash_c')).to eql(
+            { 'hash_ca' => { 'cad' => 'value hash_c.hash_ca.cad (from global custom)' }})
+        end
       end
     end
-  end
 
-  context 'when passing a hash as the only parameter' do
-    it 'can pass a single name correctly' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::a'}")
-      expect(resources).to include('env_a')
-    end
+    context 'and a module' do
+      let(:mod_a_files) { {} }
 
-    it 'can pass a an array of names correctly' do
-      resources = assemble_and_compile('${r}', "{name => ['abc::b', 'abc::a']}")
-      expect(resources).to include('module_b')
-    end
+      let(:populated_env_dir) do
+        dir_contained_in(env_dir, DeepMerge.deep_merge!(environment_files, env_name => { 'modules' => mod_a_files }))
+        env_dir
+      end
 
-    it 'can pass an override map and find values there even though they would be found' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::a', override => { abc::a => 'override_a'}}")
-      expect(resources).to include('override_a')
-    end
+      context 'that has no lookup configured' do
+        let(:mod_a_files) do
+          {
+            'mod_a' => {
+              'data' => {
+                'common.yaml' => <<-YAML.unindent
+                ---
+                mod_a::b: value mod_a::b (from mod_a)
+              YAML
+              }
+            }
+          }
+        end
 
-    it 'can pass an default_values_hash and find values there correctly' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::x', default_values_hash => { abc::x => 'extra_x'}}")
-      expect(resources).to include('extra_x')
-    end
+        it 'does not find data in the module' do
+          expect { lookup('mod_a::b') }.to raise_error(Puppet::DataBinding::LookupError, /did not find a value for the name 'mod_a::b'/)
+        end
 
-    it 'can pass an default_values_hash but not use it when value is found elsewhere' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::a', default_values_hash => { abc::a => 'extra_a'}}")
-      expect(resources).to include('env_a')
-    end
+        context 'with a Hiera v3 configuration' do
+          let(:mod_a_files) do
+            {
+              'mod_a' => {
+                'hiera.yaml' => <<-YAML.unindent
+                  ---
+                  :backends: yaml
+                  YAML
+              }
+            }
+          end
 
-    it 'can pass an default_values_hash but not use it when value is found elsewhere even when found value is undef' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::n', default_values_hash => { abc::n => 'extra_n'}}")
-      expect(resources).to include('no_value')
-    end
+          it 'raises an error' do
+            expect { lookup('mod_a::a') }.to raise_error(Puppet::Error, /hiera configuration version 3 cannot be used in a module/)
+          end
+        end
 
-    it 'can pass an override and an default_values_hash and find the override value' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::x', override => { abc::x => 'override_x'}, default_values_hash => { abc::x => 'extra_x'}}")
-      expect(resources).to include('override_x')
-    end
+        context "but a metadata.json with 'module_data_provider=hiera'" do
+          let(:mod_a_files_1) { DeepMerge.deep_merge!(mod_a_files, 'mod_a' => { 'metadata.json' => <<-JSON.unindent }) }
+              {
+                  "name": "example/mod_a",
+                  "version": "0.0.2",
+                  "source": "git@github.com/example/mod_a.git",
+                  "dependencies": [],
+                  "author": "Bob the Builder",
+                  "license": "Apache-2.0",
+                  "data_provider": "hiera"
+              }
+              JSON
 
-    it 'will raise an exception when value is not found for single key and no default is provided' do
-      expect do
-        assemble_and_compile('${r}', "{name => 'abc::x'}")
-      end.to raise_error(Puppet::ParseError, /did not find a value for the name 'abc::x'/)
-    end
+          let(:populated_env_dir) do
+            dir_contained_in(env_dir, DeepMerge.deep_merge!(environment_files, env_name => { 'modules' => mod_a_files_1 }))
+            env_dir
+          end
 
-    it 'will not raise an exception when value is not found default value is nil' do
-      resources = assemble_and_compile('${r}', "{name => 'abc::x', default_value => undef}")
-      expect(resources).to include('no_value')
-    end
-  end
+          it 'finds data in the module and reports deprecation warning for metadata.json' do
+            expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+            expect(warnings).to include(/Defining "data_provider": "hiera" in metadata.json is deprecated. A 'hiera.yaml' file should be used instead/)
+          end
 
-  context 'when accessing from outside a module' do
-    it 'will raise an exception when key in the function provided module data is not prefixed' do
-      Puppet[:code] = "include bad_data\nlookup(bad_data::b)"
-      expect do
-        compiler.compile()
-      end.to raise_error(Puppet::ParseError, /data for module 'bad_data' must use keys qualified with the name of the module/)
-    end
+          context 'and a hiera.yaml file' do
+            let(:mod_a_files_2) { DeepMerge.deep_merge!(mod_a_files_1, 'mod_a' => { 'hiera.yaml' => <<-YAML.unindent }) }
+            ---
+            version: 4
+            hierarchy:
+              - name: common
+                backend: yaml
+            YAML
 
-    it 'will resolve global, environment, and module correctly' do
-      Hiera.any_instance.expects(:lookup).with('bca::e', any_parameters).returns({ 'k1' => 'global_e1' })
-      Puppet[:code] = <<-END.gsub(/^ {8}/, '')
-        include bca
-        $r = lookup(bca::e, Hash[String,String], hash)
-        notify { "${r[k1]}_${r[k2]}_${r[k3]}": }
-      END
-      resources = compiler.compile().resources.map(&:ref).select { |r| r.start_with?('Notify[') }.map { |r| r[7..-2] }
-      expect(resources).to include('global_e1_module_bca_e2_env_bca_e3')
-    end
+            let(:populated_env_dir) do
+              dir_contained_in(env_dir, DeepMerge.deep_merge!(environment_files, env_name => { 'modules' => mod_a_files_2 }))
+              env_dir
+            end
 
-    it 'will resolve global and environment correctly when module has no provider' do
-      Hiera.any_instance.expects(:lookup).with('no_provider::e', any_parameters).returns({ 'k1' => 'global_e1' })
-      Puppet[:code] = <<-END.gsub(/^ {8}/, '')
-        include no_provider
-        $r = lookup(no_provider::e, Hash[String,String], hash)
-        notify { "${r[k1]}_${r[k2]}_${r[k3]}": }
-      END
-      resources = compiler.compile().resources.map(&:ref).select { |r| r.start_with?('Notify[') }.map { |r| r[7..-2] }
-      expect(resources).to include('global_e1__env_no_provider_e3') # k2 is missing
-    end
-  end
+            it 'finds data in the module and reports deprecation warnings for both metadata.json and hiera.yaml' do
+              expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+              expect(warnings).to include(/Defining "data_provider": "hiera" in metadata.json is deprecated/)
+              expect(warnings).to include(/Use of 'hiera.yaml' version 4 is deprecated. It should be converted to version 5/)
+            end
+          end
+        end
+      end
+
+      context 'using a data_hash that reads a yaml file' do
+        let(:mod_a_files) do
+          {
+            'mod_a' => {
+              'data' => {
+                'common.yaml' => <<-YAML.unindent
+                ---
+                mod_a::a: value mod_a::a (from mod_a)
+                mod_a::b: value mod_a::b (from mod_a)
+                mod_a::xo: value mod_a::xo (from mod_a)
+                mod_a::xd_found: value mod_a::xd_found (from mod_a)
+                mod_a::interpolate_xo: "-- %{lookup('mod_a::xo')} --"
+                mod_a::interpolate_xd: "-- %{lookup('mod_a::xd')} --"
+                mod_a::interpolate_scope_xo: "-- %{scope_xo} --"
+                mod_a::interpolate_scope_xd: "-- %{scope_xd} --"
+                mod_a::hash_a:
+                  a: value mod_a::hash_a.a (from mod_a)
+                  b: value mod_a::hash_a.b (from mod_a)
+                mod_a::hash_b:
+                  a: value mod_a::hash_b.a (from mod_a)
+                  b: value mod_a::hash_b.b (from mod_a)
+                mod_a::interpolated: "-- %{lookup('mod_a::a')} --"
+                mod_a::a_a: "-- %{lookup('mod_a::hash_a.a')} --"
+                mod_a::a_b: "-- %{lookup('mod_a::hash_a.b')} --"
+                mod_a::b_a: "-- %{lookup('mod_a::hash_b.a')} --"
+                mod_a::b_b: "-- %{lookup('mod_a::hash_b.b')} --"
+                mod_a::interpolate_array:
+                  - "-- %{lookup('mod_a::a')} --"
+                  - "-- %{lookup('mod_a::b')} --"
+                mod_a::interpolate_literal: "-- %{literal('hello')} --"
+                mod_a::interpolate_scope: "-- %{scope_scalar} --"
+                mod_a::interpolate_scope_not_found: "-- %{scope_nope} --"
+                mod_a::interpolate_scope_dig: "-- %{scope_hash.a} --"
+                mod_a::interpolate_scope_dig_not_found: "-- %{scope_hash.nope} --"
+                mod_a::quoted_interpolation: '-- %{lookup(''"mod_a::a.quoted.key"'')} --'
+                "mod_a::a.quoted.key": "value mod_a::a.quoted.key (from mod_a)"
+              YAML
+              },
+              'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: "Common"
+                  data_hash: yaml_data
+                  path: "common.yaml"
+            YAML
+            }
+          }
+        end
+
+        it 'finds data in the module' do
+          expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+        end
+
+        it 'environment data has higher priority than module data' do
+          expect(lookup('mod_a::a')).to eql('value mod_a::a (from environment)')
+        end
+
+        it 'environment data has higher priority than module data in interpolated module data' do
+          expect(lookup('mod_a::interpolated')).to eql('-- value mod_a::a (from environment) --')
+        end
+
+        it 'overrides have higher priority than found data' do
+          expect(lookup('mod_a::xo', { 'override' => overrides })).to eql('value mod_a::xo (from override)')
+        end
+
+        it 'overrides have higher priority than found data in lookup interpolations' do
+          expect(lookup('mod_a::interpolate_xo', { 'override' => overrides })).to eql('-- value mod_a::xo (from override) --')
+        end
+
+        it 'overrides have higher priority than found data in scope interpolations' do
+          expect(lookup('mod_a::interpolate_scope_xo', { 'override' => overrides })).to eql('-- value scope_xo (from override) --')
+        end
+
+        it 'defaults have lower priority than found data' do
+          expect(lookup('mod_a::xd_found', { 'default_values_hash' => defaults })).to eql('value mod_a::xd_found (from mod_a)')
+        end
+
+        it 'defaults are used when data is not found' do
+          expect(lookup('mod_a::xd', { 'default_values_hash' => defaults })).to eql('value mod_a::xd (from default)')
+        end
+
+        it 'defaults are used when data is not found in lookup interpolations' do
+          expect(lookup('mod_a::interpolate_xd', { 'default_values_hash' => defaults })).to eql('-- value mod_a::xd (from default) --')
+        end
+
+        it 'defaults are used when data is not found in scope interpolations' do
+          expect(lookup('mod_a::interpolate_scope_xd', { 'default_values_hash' => defaults })).to eql('-- value scope_xd (from default) --')
+        end
+
+        it 'merges hashes from environment and module unless strategy hash is used' do
+          expect(lookup('mod_a::hash_a')).to eql({'a' => 'value mod_a::hash_a.a (from environment)'})
+        end
+
+        it 'merges hashes from environment and module when merge strategy hash is used' do
+          expect(lookup('mod_a::hash_a', :merge => 'hash')).to eql(
+            {'a' => 'value mod_a::hash_a.a (from environment)', 'b' => 'value mod_a::hash_a.b (from mod_a)'})
+        end
+
+        it 'will not merge hashes from environment and module in interpolated expressions' do
+          expect(lookup(['mod_a::a_a', 'mod_a::a_b'])).to eql(
+            ['-- value mod_a::hash_a.a (from environment) --', '--  --']) # root key found in environment, no hash merge is performed
+        end
+
+        it 'interpolates arrays' do
+          expect(lookup('mod_a::interpolate_array')).to eql(['-- value mod_a::a (from environment) --', '-- value mod_a::b (from mod_a) --'])
+        end
+
+        it 'can dig into arrays using subkeys' do
+          expect(lookup('mod_a::interpolate_array.1')).to eql('-- value mod_a::b (from mod_a) --')
+        end
+
+        it 'treats an out of range subkey as not found' do
+          expect(explain('mod_a::interpolate_array.2')).to match(/No such key: "2"/)
+        end
 
-  context 'when accessing bad data' do
-    it 'will raise an exception when key in the function provided module data is not prefixed' do
-      Puppet[:code] = 'include bad_data'
-      expect do
-        compiler.compile()
-      end.to raise_error(Puppet::ParseError, /data for module 'bad_data' must use keys qualified with the name of the module/)
+        it 'interpolates a literal' do
+          expect(lookup('mod_a::interpolate_literal')).to eql('-- hello --')
+        end
+
+        it 'interpolates scalar from scope' do
+          expect(lookup('mod_a::interpolate_scope')).to eql('-- scope scalar value --')
+        end
+
+        it 'interpolates not found in scope as empty string' do
+          expect(lookup('mod_a::interpolate_scope_not_found')).to eql('--  --')
+        end
+
+        it 'interpolates dotted key from scope' do
+          expect(lookup('mod_a::interpolate_scope_dig')).to eql('-- scope hash a --')
+        end
+
+        it 'treates interpolated dotted key but not found in scope as empty string' do
+          expect(lookup('mod_a::interpolate_scope_dig_not_found')).to eql('--  --')
+        end
+
+        it 'can use quoted keys in interpolation' do
+          expect(lookup('mod_a::quoted_interpolation')).to eql('-- value mod_a::a.quoted.key (from mod_a) --') # root key found in environment, no hash merge is performed
+        end
+
+        it 'merges hashes from environment and module in interpolated expressions if hash merge is specified in lookup options' do
+          expect(lookup(['mod_a::b_a', 'mod_a::b_b'])).to eql(
+            ['-- value mod_a::hash_b.a (from environment) --', '-- value mod_a::hash_b.b (from mod_a) --'])
+        end
+      end
+
+      context 'using a lookup_key that uses a path' do
+        let(:mod_a_files) do
+          {
+            'mod_a' => {
+              'functions' => {
+                'pp_lookup_key.pp' => <<-PUPPET.unindent
+                  function mod_a::pp_lookup_key($key, $options, $context) {
+                    if !$context.cache_has_key(undef) {
+                      $context.cache_all(yaml_data($options, $context))
+                      $context.cache(undef, true)
+                    }
+                    if $context.cache_has_key($key) { $context.cached_value($key) } else { $context.not_found }
+                  }
+                  PUPPET
+              },
+              'hiera.yaml' => <<-YAML.unindent,
+                ---
+                version: 5
+                hierarchy:
+                  - name: "Common"
+                    lookup_key: mod_a::pp_lookup_key
+                    path: common.yaml
+                YAML
+              'data' => {
+                'common.yaml' => <<-YAML.unindent
+                  mod_a::b: value mod_a::b (from mod_a)
+                  YAML
+              }
+            }
+          }
+        end
+
+        it 'finds data in the module' do
+          expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+        end
+      end
+
+      context 'using a lookup_key that is a puppet function' do
+        let(:mod_a_files) do
+          {
+            'mod_a' => {
+              'functions' => {
+                'pp_lookup_key.pp' => <<-PUPPET.unindent
+                function mod_a::pp_lookup_key($key, $options, $context) {
+                  case $key {
+                    'mod_a::really_interpolated': { $context.interpolate("-- %{lookup('mod_a::a')} --") }
+                    'mod_a::recursive': { lookup($key) }
+                    default: {
+                      if $context.cache_has_key(mod_a::a) {
+                        $context.explain || { 'reusing cache' }
+                      } else {
+                        $context.explain || { 'initializing cache' }
+                        $context.cache_all({
+                          mod_a::a => 'value mod_a::a (from mod_a)',
+                          mod_a::b => 'value mod_a::b (from mod_a)',
+                          mod_a::c => 'value mod_a::c (from mod_a)',
+                          mod_a::hash_a => {
+                            a => 'value mod_a::hash_a.a (from mod_a)',
+                            b => 'value mod_a::hash_a.b (from mod_a)'
+                          },
+                          mod_a::hash_b => {
+                            a => 'value mod_a::hash_b.a (from mod_a)',
+                            b => 'value mod_a::hash_b.b (from mod_a)'
+                          },
+                          mod_a::interpolated => "-- %{lookup('mod_a::a')} --",
+                          mod_a::a_a => "-- %{lookup('mod_a::hash_a.a')} --",
+                          mod_a::a_b => "-- %{lookup('mod_a::hash_a.b')} --",
+                          mod_a::b_a => "-- %{lookup('mod_a::hash_b.a')} --",
+                          mod_a::b_b => "-- %{lookup('mod_a::hash_b.b')} --",
+                          'mod_a::a.quoted.key' => 'value mod_a::a.quoted.key (from mod_a)',
+                          mod_a::sensitive => Sensitive('reduct me please'),
+                          mod_a::type => Object[{name => 'FindMe', 'attributes' => {'x' => String}}],
+                          mod_a::version => SemVer('3.4.1'),
+                          mod_a::version_range => SemVerRange('>=3.4.1'),
+                          mod_a::timestamp => Timestamp("1994-03-25T19:30:00"),
+                          mod_a::timespan => Timespan("3-10:00:00")
+                        })
+                      }
+                      if !$context.cache_has_key($key) {
+                        $context.not_found
+                      }
+                      $context.explain || { "returning value for $key" }
+                      $context.cached_value($key)
+                    }
+                  }
+                }
+              PUPPET
+              },
+              'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: "Common"
+                  lookup_key: mod_a::pp_lookup_key
+            YAML
+            }
+          }
+        end
+
+        it 'finds data in the module' do
+          expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+        end
+
+        it 'environment data has higher priority than module data' do
+          expect(lookup('mod_a::a')).to eql('value mod_a::a (from environment)')
+        end
+
+        it 'finds quoted keys in the module' do
+          expect(lookup('"mod_a::a.quoted.key"')).to eql('value mod_a::a.quoted.key (from mod_a)')
+        end
+
+        it 'will not resolve interpolated expressions' do
+          expect(lookup('mod_a::interpolated')).to eql("-- %{lookup('mod_a::a')} --")
+        end
+
+        it 'resolves interpolated expressions using Context#interpolate' do
+          expect(lookup('mod_a::really_interpolated')).to eql("-- value mod_a::a (from environment) --")
+        end
+
+        it 'will not merge hashes from environment and module unless strategy hash is used' do
+          expect(lookup('mod_a::hash_a')).to eql({ 'a' => 'value mod_a::hash_a.a (from environment)' })
+        end
+
+        it 'merges hashes from environment and module when merge strategy hash is used' do
+          expect(lookup('mod_a::hash_a', :merge => 'hash')).to eql({ 'a' => 'value mod_a::hash_a.a (from environment)', 'b' => 'value mod_a::hash_a.b (from mod_a)' })
+        end
+
+        it 'traps recursive lookup trapped' do
+          expect(explain('mod_a::recursive')).to include('Recursive lookup detected')
+        end
+
+        it 'private cache is persisted over multiple calls' do
+          collect_notices("notice(lookup('mod_a::b')) notice(lookup('mod_a::c'))", true)
+          expect(notices).to eql(['value mod_a::b (from mod_a)', 'value mod_a::c (from mod_a)'])
+          expect(explanation).to match(/initializing cache.*reusing cache/m)
+          expect(explanation).not_to match(/initializing cache.*initializing cache/m)
+        end
+
+        it 'the same key is requested only once' do
+          collect_notices("notice(lookup('mod_a::b')) notice(lookup('mod_a::b'))", true)
+          expect(notices).to eql(['value mod_a::b (from mod_a)', 'value mod_a::b (from mod_a)'])
+          expect(explanation).to match(/Found key: "mod_a::b".*Found key: "mod_a::b"/m)
+          expect(explanation).to match(/returning value for mod_a::b/m)
+          expect(explanation).not_to match(/returning value for mod_a::b.*returning value for mod_a::b/m)
+        end
+
+        context 'and calling function via API' do
+          it 'finds and delivers rich data' do
+            collect_notices("notice('success')") do |scope|
+              expect(lookup_func.call(scope, 'mod_a::sensitive')).to be_a(Puppet::Pops::Types::PSensitiveType::Sensitive)
+              expect(lookup_func.call(scope, 'mod_a::type')).to be_a(Puppet::Pops::Types::PObjectType)
+              expect(lookup_func.call(scope, 'mod_a::version')).to eql(SemanticPuppet::Version.parse('3.4.1'))
+              expect(lookup_func.call(scope, 'mod_a::version_range')).to eql(SemanticPuppet::VersionRange.parse('>=3.4.1'))
+              expect(lookup_func.call(scope, 'mod_a::timestamp')).to eql(Puppet::Pops::Time::Timestamp.parse('1994-03-25T19:30:00'))
+              expect(lookup_func.call(scope, 'mod_a::timespan')).to eql(Puppet::Pops::Time::Timespan.parse('3-10:00:00'))
+            end
+            expect(notices).to eql(['success'])
+          end
+        end
+      end
+
+      context 'using a data_dig that is a ruby function' do
+        let(:mod_a_files) do
+          {
+            'mod_a' => {
+              'lib' => {
+                'puppet' => {
+                  'functions' => {
+                    'mod_a' => {
+                      'ruby_dig.rb' => <<-RUBY.unindent
+                      Puppet::Functions.create_function(:'mod_a::ruby_dig') do
+                        dispatch :ruby_dig do
+                          param 'Array[String[1]]', :segments
+                          param 'Hash[String,Any]', :options
+                          param 'Puppet::LookupContext', :context
+                        end
+
+                        def ruby_dig(segments, options, context)
+                          sub_segments = segments.dup
+                          root_key = sub_segments.shift
+                          case root_key
+                          when 'mod_a::options'
+                            hash = { 'mod_a::options' => options }
+                          when 'mod_a::lookup'
+                            return call_function('lookup', segments.join('.'))
+                          else
+                            hash = {
+                              'mod_a::a' => 'value mod_a::a (from mod_a)',
+                              'mod_a::b' => 'value mod_a::b (from mod_a)',
+                              'mod_a::hash_a' => {
+                                'a' => 'value mod_a::hash_a.a (from mod_a)',
+                                'b' => 'value mod_a::hash_a.b (from mod_a)'
+                              },
+                              'mod_a::hash_b' => {
+                                'a' => 'value mod_a::hash_b.a (from mod_a)',
+                                'b' => 'value mod_a::hash_b.b (from mod_a)'
+                              },
+                              'mod_a::interpolated' => "-- %{lookup('mod_a::a')} --",
+                              'mod_a::really_interpolated' => "-- %{lookup('mod_a::a')} --",
+                              'mod_a::a_a' => "-- %{lookup('mod_a::hash_a.a')} --",
+                              'mod_a::a_b' => "-- %{lookup('mod_a::hash_a.b')} --",
+                              'mod_a::b_a' => "-- %{lookup('mod_a::hash_b.a')} --",
+                              'mod_a::b_b' => "-- %{lookup('mod_a::hash_b.b')} --",
+                              'mod_a::bad_type' => :oops,
+                              'mod_a::bad_type_in_hash' => { 'a' => :oops },
+                            }
+                            end
+                          context.not_found unless hash.include?(root_key)
+                          value = sub_segments.reduce(hash[root_key]) do |memo, segment|
+                            context.not_found unless memo.is_a?(Hash) && memo.include?(segment)
+                            memo[segment]
+                          end
+                          root_key == 'mod_a::really_interpolated' ? context.interpolate(value) : value
+                        end
+                      end
+                    RUBY
+                    }
+                  }
+                }
+              },
+              'hiera.yaml' => <<-YAML.unindent,
+              ---
+              version: 5
+              hierarchy:
+                - name: "Common"
+                  data_dig: mod_a::ruby_dig
+                  uri: "http://www.example.com/passed/as/option"
+                  options:
+                    option_a: Option value a
+                    option_b:
+                      x: Option value b.x
+                      y: Option value b.y
+            YAML
+            }
+          }
+        end
+
+        it 'finds data in the module' do
+          expect(lookup('mod_a::b')).to eql('value mod_a::b (from mod_a)')
+        end
+
+        it 'environment data has higher priority than module data' do
+          expect(lookup('mod_a::a')).to eql('value mod_a::a (from environment)')
+        end
+
+        it 'will not resolve interpolated expressions' do
+          expect(lookup('mod_a::interpolated')).to eql("-- %{lookup('mod_a::a')} --")
+        end
+
+        it 'resolves interpolated expressions using Context#interpolate' do
+          expect(lookup('mod_a::really_interpolated')).to eql("-- value mod_a::a (from environment) --")
+        end
+
+        it 'does not accept return of runtime type from function' do
+          expect(explain('mod_a::bad_type')).to include('Value returned from Hierarchy entry "Common" has wrong type')
+        end
+
+        it 'does not accept return of runtime type embedded in hash from function' do
+          expect(explain('mod_a::bad_type_in_hash')).to include('Value returned from Hierarchy entry "Common" has wrong type')
+        end
+
+        it 'will not merge hashes from environment and module unless strategy hash is used' do
+          expect(lookup('mod_a::hash_a')).to eql({'a' => 'value mod_a::hash_a.a (from environment)'})
+        end
+
+        it 'hierarchy entry options are passed to the function' do
+          expect(lookup('mod_a::options.option_b.x')).to eql('Option value b.x')
+        end
+
+        it 'hierarchy entry "uri" is passed as location option to the function' do
+          expect(lookup('mod_a::options.uri')).to eql('http://www.example.com/passed/as/option')
+        end
+
+        it 'recursive lookup is trapped' do
+          expect(explain('mod_a::lookup.mod_a::lookup')).to include('Recursive lookup detected')
+        end
+
+        context 'with merge strategy hash' do
+          it 'merges hashes from environment and module' do
+            expect(lookup('mod_a::hash_a', :merge => 'hash')).to eql({'a' => 'value mod_a::hash_a.a (from environment)', 'b' => 'value mod_a::hash_a.b (from mod_a)'})
+          end
+
+          it 'will "undig" value from data_dig function, merge root hashes, and then dig to get values by subkey' do
+            expect(lookup(['mod_a::hash_a.a', 'mod_a::hash_a.b'], :merge => 'hash')).to eql(
+              ['value mod_a::hash_a.a (from environment)', 'value mod_a::hash_a.b (from mod_a)'])
+          end
+        end
+      end
     end
   end
 end