diff --git a/README_DEVELOPER.md b/README_DEVELOPER.md
new file mode 100644
index 000000000..faadde828
--- /dev/null
+++ b/README_DEVELOPER.md
@@ -0,0 +1,63 @@
+# Developer README #
+
+This file is intended to provide a place for developers and contributors to
+document what other developers need to know about changes made to Puppet.
+
+# UTF-8 Handling #
+
+As Ruby 1.9 becomes more commonly used with Puppet, developers should be aware
+of major changes to the way Strings and Regexp objects are handled.
+Specifically, every instance of these two classes will have an encoding
+attribute determined in a number of ways.
+
+ * If the source file has an encoding specified in the magic comment at the
+   top, the instance will take on that encoding.
+ * Otherwise, the encoding will be determined by the LC\_LANG or LANG
+   environment variables.
+ * Otherwise, the encoding will default to ASCII-8BIT
+
+## References ##
+
+Excellent information about the differences between encodings in Ruby 1.8 and
+Ruby 1.9 is published in this blog series:
+[Understanding M17n](http://links.puppetlabs.com/understanding_m17n)
+
+## Encodings of Regexp and String instances ##
+
+In general, please be aware that Ruby 1.9 regular expressions need to be
+compatible with the encoding of a string being used to match them.  If they are
+not compatible you can expect to receive and error such as:
+
+    Encoding::CompatibilityError: incompatible encoding regexp match (ASCII-8BIT
+    regexp with UTF-8 string)
+
+In addition, some escape sequences where were valid in Ruby 1.8 are no longer
+valid in 1.9 if the regular expression is not marked as an ASCII-8BIT object.
+You may expect errors like this in this situation:
+
+    SyntaxError: (irb):7: invalid multibyte escape: /\xFF/
+
+This error is particularly comment when serializing a string to other
+representations like JSON or YAML.  To resolve the problem you can explicitly
+mark the regular expression as ASCII-8BIT using the /n flag:
+
+    "a" =~ /\342\230\203/n
+
+Finally, any time you're thinking of a string as an array of bytes rather than
+an array of characters, common when escaping a string, you should work with
+everything in ASCII-8BIT.  Changing the encoding will not change the data
+itself and allow the Regexp and the String to deal with bytes rather than
+characters.
+
+Puppet provides a monkey patch to String which returns an encoding suitable for
+byte manipulations:
+
+    # Example of how to escape non ASCII printable characters for YAML.
+    >> snowman = "☃"
+    >> snowman.to_ascii8bit.gsub(/([\x80-\xFF])/n) { |x| "\\x#{x.unpack("C")[0].to_s(16)} }
+    => "\\xe2\\x98\\x83"
+
+If the Regexp is not marked as ASCII-8BIT using /n, then you can expect the
+SyntaxError, invalid multibyte escape as mentioned above.
+
+EOF
diff --git a/lib/puppet/util/zaml.rb b/lib/puppet/util/zaml.rb
index bbb2af2d2..e07a2d7b9 100644
--- a/lib/puppet/util/zaml.rb
+++ b/lib/puppet/util/zaml.rb
@@ -1,321 +1,355 @@
+# encoding: UTF-8
 #
+# The above encoding line is a magic comment to set the default source encoding
+# of this file for the Ruby interpreter.  It must be on the first or second
+# line of the file if an interpreter is in use.  In Ruby 1.9 and later, the
+# source encoding determines the encoding of String and Regexp objects created
+# from this source file.  This explicit encoding is important becuase otherwise
+# Ruby will pick an encoding based on LANG or LC_CTYPE environment variables.
+# These may be different from site to site so it's important for us to
+# establish a consistent behavior.  For more information on M17n please see:
+# http://links.puppetlabs.com/understanding_m17n
+
 # ZAML -- A partial replacement for YAML, writen with speed and code clarity
 #         in mind.  ZAML fixes one YAML bug (loading Exceptions) and provides
 #         a replacement for YAML.dump unimaginatively called ZAML.dump,
 #         which is faster on all known cases and an order of magnitude faster
 #         with complex structures.
 #
 # http://github.com/hallettj/zaml
 #
 # Authors: Markus Roberts, Jesse Hallett, Ian McIntosh, Igal Koshevoy, Simon Chiang
 #
 
 require 'yaml'
 
 class ZAML
   VERSION = "0.1.1"
   #
   # Class Methods
   #
   def self.dump(stuff, where='')
     z = new
     stuff.to_zaml(z)
     where << z.to_s
   end
   #
   # Instance Methods
   #
   def initialize
     @result = []
     @indent = nil
     @structured_key_prefix = nil
     @previously_emitted_object = {}
     @next_free_label_number = 0
     emit('--- ')
   end
   def nested(tail='  ')
     old_indent = @indent
     @indent = "#{@indent || "\n"}#{tail}"
     yield
     @indent = old_indent
   end
   class Label
     #
     # YAML only wants objects in the datastream once; if the same object
     #    occurs more than once, we need to emit a label ("&idxxx") on the
     #    first occurrence and then emit a back reference (*idxxx") on any
     #    subsequent occurrence(s).
     #
     # To accomplish this we keeps a hash (by object id) of the labels of
     #    the things we serialize as we begin to serialize them.  The labels
     #    initially serialize as an empty string (since most objects are only
     #    going to be be encountered once), but can be changed to a valid
     #    (by assigning it a number) the first time it is subsequently used,
     #    if it ever is.  Note that we need to do the label setup BEFORE we
     #    start to serialize the object so that circular structures (in
     #    which we will encounter a reference to the object as we serialize
     #    it can be handled).
     #
     attr_accessor :this_label_number
     def initialize(obj,indent)
       @indent = indent
       @this_label_number = nil
       @obj = obj # prevent garbage collection so that object id isn't reused
     end
     def to_s
       @this_label_number ? ('&id%03d%s' % [@this_label_number, @indent]) : ''
     end
     def reference
       @reference         ||= '*id%03d' % @this_label_number
     end
   end
   def label_for(obj)
     @previously_emitted_object[obj.object_id]
   end
   def new_label_for(obj)
     label = Label.new(obj,(Hash === obj || Array === obj) ? "#{@indent || "\n"}  " : ' ')
     @previously_emitted_object[obj.object_id] = label
     label
   end
   def first_time_only(obj)
     if label = label_for(obj)
       label.this_label_number ||= (@next_free_label_number += 1)
       emit(label.reference)
     else
       if @structured_key_prefix and not obj.is_a? String
         emit(@structured_key_prefix)
         @structured_key_prefix = nil
       end
       emit(new_label_for(obj))
       yield
     end
   end
   def emit(s)
     @result << s
     @recent_nl = false unless s.kind_of?(Label)
   end
   def nl(s='')
     emit(@indent || "\n") unless @recent_nl
     emit(s)
     @recent_nl = true
   end
   def to_s
     @result.join
   end
   def prefix_structured_keys(x)
     @structured_key_prefix = x
     yield
     nl unless @structured_key_prefix
     @structured_key_prefix = nil
   end
 end
 
 ################################################################
 #
 #   Behavior for custom classes
 #
 ################################################################
 
 class Object
   def to_yaml_properties
     instance_variables.sort        # Default YAML behavior
   end
   def yaml_property_munge(x)
     x
   end
   def zamlized_class_name(root)
     cls = self.class
     "!ruby/#{root.name.downcase}#{cls == root ? '' : ":#{cls.respond_to?(:name) ? cls.name : cls}"}"
   end
   def to_zaml(z)
     z.first_time_only(self) {
       z.emit(zamlized_class_name(Object))
       z.nested {
         instance_variables = to_yaml_properties
         if instance_variables.empty?
           z.emit(" {}")
         else
           instance_variables.each { |v|
             z.nl
             v[1..-1].to_zaml(z)       # Remove leading '@'
             z.emit(': ')
             yaml_property_munge(instance_variable_get(v)).to_zaml(z)
           }
         end
       }
     }
   end
 end
 
 ################################################################
 #
 #   Behavior for built-in classes
 #
 ################################################################
 
 class NilClass
   def to_zaml(z)
     z.emit('')        # NOTE: blank turns into nil in YAML.load
   end
 end
 
 class Symbol
   def to_zaml(z)
     z.emit(self.inspect)
   end
 end
 
 class TrueClass
   def to_zaml(z)
     z.emit('true')
   end
 end
 
 class FalseClass
   def to_zaml(z)
     z.emit('false')
   end
 end
 
 class Numeric
   def to_zaml(z)
     z.emit(self)
   end
 end
 
 class Regexp
   def to_zaml(z)
     z.first_time_only(self) { z.emit("#{zamlized_class_name(Regexp)} #{inspect}") }
   end
 end
 
 class Exception
   def to_zaml(z)
     z.emit(zamlized_class_name(Exception))
     z.nested {
       z.nl("message: ")
       message.to_zaml(z)
     }
   end
   #
   # Monkey patch for buggy Exception restore in YAML
   #
   #     This makes it work for now but is not very future-proof; if things
   #     change we'll most likely want to remove this.  To mitigate the risks
   #     as much as possible, we test for the bug before appling the patch.
   #
   if respond_to? :yaml_new and yaml_new(self, :tag, "message" => "blurp").message != "blurp"
     def self.yaml_new( klass, tag, val )
       o = YAML.object_maker( klass, {} ).exception(val.delete( 'message'))
       val.each_pair do |k,v|
         o.instance_variable_set("@#{k}", v)
       end
       o
     end
   end
 end
 
 class String
   ZAML_ESCAPES = %w{\x00 \x01 \x02 \x03 \x04 \x05 \x06 \a \x08 \t \n \v \f \r \x0e \x0f \x10 \x11 \x12 \x13 \x14 \x15 \x16 \x17 \x18 \x19 \x1a \e \x1c \x1d \x1e \x1f }
   def escaped_for_zaml
-    gsub( /\x5C/, "\\\\\\" ).  # Demi-kludge for Maglev/rubinius; the regexp should be /\\/ but parsetree chokes on that.
-    gsub( /"/, "\\\"" ).
-    gsub( /([\x00-\x1F])/ ) { |x| ZAML_ESCAPES[ x.unpack("C")[0] ] }.
-    gsub( /([\x80-\xFF])/ ) { |x| "\\x#{x.unpack("C")[0].to_s(16)}" }
+    # JJM (Note the trailing dots to construct a multi-line method chain.) This
+    # code is meant to escape all bytes which are not ASCII-8BIT printable
+    # characters.  Multi-byte unicode characters are handled just fine because
+    # each byte of the character results in an escaped string emitted to the
+    # YAML stream.  When the YAML is de-serialized back into a String the bytes
+    # will be reconstructed properly into the unicode character.
+    self.to_ascii8bit.gsub( /\x5C/n, "\\\\\\" ).  # Demi-kludge for Maglev/rubinius; the regexp should be /\\/ but parsetree chokes on that.
+    gsub( /"/n, "\\\"" ).
+    gsub( /([\x00-\x1F])/n ) { |x| ZAML_ESCAPES[ x.unpack("C")[0] ] }.
+    gsub( /([\x80-\xFF])/n ) { |x| "\\x#{x.unpack("C")[0].to_s(16)}" }
   end
   def to_zaml(z)
     z.first_time_only(self) {
       num = '[-+]?(0x)?\d+\.?\d*'
       case
         when self == ''
           z.emit('""')
         # when self =~ /[\x00-\x08\x0B\x0C\x0E-\x1F\x80-\xFF]/
         #   z.emit("!binary |\n")
         #   z.emit([self].pack("m*"))
         when (
           (self =~ /\A(true|false|yes|no|on|null|off|#{num}(:#{num})*|!|=|~)$/i) or
           (self =~ /\A\n* /) or
           (self =~ /[\s:]$/) or
           (self =~ /^[>|][-+\d]*\s/i) or
           (self[-1..-1] =~ /\s/) or
-          (self =~ /[\x00-\x08\x0B\x0C\x0E-\x1F\x80-\xFF]/) or
+          # This regular expression assumes the string is a byte sequence.
+          # It does not concern itself with characters so we convert the string
+          # to ASCII-8BIT for Ruby 1.9 to match up encodings.
+          (self.to_ascii8bit=~ /[\x00-\x08\x0B\x0C\x0E-\x1F\x80-\xFF]/n) or
           (self =~ /[,\[\]\{\}\r\t]|:\s|\s#/) or
           (self =~ /\A([-:?!#&*'"]|<<|%.+:.)/)
           )
           z.emit("\"#{escaped_for_zaml}\"")
         when self =~ /\n/
           if self[-1..-1] == "\n" then z.emit('|+') else z.emit('|-') end
           z.nested { split("\n",-1).each { |line| z.nl; z.emit(line.chomp("\n")) } }
         else
           z.emit(self)
       end
     }
   end
+
+  # Return a guranteed ASCII-8BIT encoding for Ruby 1.9 This is a helper
+  # method for other methods that perform regular expressions against byte
+  # sequences deliberately rather than dealing with characters.
+  # The method may or may not return a new instance.
+  def to_ascii8bit
+    if self.respond_to?(:encoding) and self.encoding.name != "ASCII-8BIT" then
+      str = self.dup
+      str.force_encoding("ASCII-8BIT")
+      return str
+    else
+      return self
+    end
+  end
 end
 
 class Hash
   def to_zaml(z)
     z.first_time_only(self) {
       z.nested {
         if empty?
           z.emit('{}')
         else
           each_pair { |k, v|
             z.nl
             z.prefix_structured_keys('? ') { k.to_zaml(z) }
             z.emit(': ')
             v.to_zaml(z)
           }
         end
       }
     }
   end
 end
 
 class Array
   def to_zaml(z)
     z.first_time_only(self) {
       z.nested {
         if empty?
           z.emit('[]')
         else
           each { |v| z.nl('- '); v.to_zaml(z) }
         end
       }
     }
   end
 end
 
 class Time
   def to_zaml(z)
     # 2008-12-06 10:06:51.373758 -07:00
     ms = ("%0.6f" % (usec * 1e-6)).sub(/^\d+\./,'')
     offset = "%+0.2i:%0.2i" % [utc_offset / 3600, (utc_offset / 60) % 60]
     z.emit(self.strftime("%Y-%m-%d %H:%M:%S.#{ms} #{offset}"))
   end
 end
 
 class Date
   def to_zaml(z)
     z.emit(strftime('%Y-%m-%d'))
   end
 end
 
 class Range
   def to_zaml(z)
     z.first_time_only(self) {
       z.emit(zamlized_class_name(Range))
       z.nested {
         z.nl
         z.emit('begin: ')
         z.emit(first)
         z.nl
         z.emit('end: ')
         z.emit(last)
         z.nl
         z.emit('excl: ')
         z.emit(exclude_end?)
       }
     }
   end
 end
diff --git a/spec/unit/util/zaml_spec.rb b/spec/unit/util/zaml_spec.rb
index d77cf99d2..858ae6044 100755
--- a/spec/unit/util/zaml_spec.rb
+++ b/spec/unit/util/zaml_spec.rb
@@ -1,62 +1,99 @@
 #!/usr/bin/env rspec
+# encoding: UTF-8
+#
+# The above encoding line is a magic comment to set the default source encoding
+# of this file for the Ruby interpreter.  It must be on the first or second
+# line of the file if an interpreter is in use.  In Ruby 1.9 and later, the
+# source encoding determines the encoding of String and Regexp objects created
+# from this source file.  This explicit encoding is important becuase otherwise
+# Ruby will pick an encoding based on LANG or LC_CTYPE environment variables.
+# These may be different from site to site so it's important for us to
+# establish a consistent behavior.  For more information on M17n please see:
+# http://links.puppetlabs.com/understanding_m17n
+
 require 'spec_helper'
 
 require 'puppet/util/monkey_patches'
 
 describe "Pure ruby yaml implementation" do
   {
   7            => "--- 7",
   3.14159      => "--- 3.14159",
   'test'       => "--- test",
   []           => "--- []",
   :symbol      => "--- !ruby/sym symbol",
   {:a => "A"}  => "--- \n  !ruby/sym a: A",
   {:a => "x\ny"} => "--- \n  !ruby/sym a: |-\n    x\n    y" 
   }.each { |o,y|
     it "should convert the #{o.class} #{o.inspect} to yaml" do
       o.to_yaml.should == y
     end
     it "should produce yaml for the #{o.class} #{o.inspect} that can be reconstituted" do
       YAML.load(o.to_yaml).should == o
     end
   }
   #
   # Can't test for equality on raw objects
   {
   Object.new                   => "--- !ruby/object {}",
   [Object.new]                 => "--- \n  - !ruby/object {}",
   {Object.new => Object.new}   => "--- \n  ? !ruby/object {}\n  : !ruby/object {}"
   }.each { |o,y|
     it "should convert the #{o.class} #{o.inspect} to yaml" do
       o.to_yaml.should == y
     end
     it "should produce yaml for the #{o.class} #{o.inspect} that can be reconstituted" do
       lambda { YAML.load(o.to_yaml) }.should_not raise_error
     end
   }
 
   it "should emit proper labels and backreferences for common objects" do
     # Note: this test makes assumptions about the names ZAML chooses
     # for labels.
     x = [1, 2]
     y = [3, 4]
     z = [x, y, x, y]
     z.to_yaml.should == "--- \n  - &id001\n    - 1\n    - 2\n  - &id002\n    - 3\n    - 4\n  - *id001\n  - *id002"
     z2 = YAML.load(z.to_yaml)
     z2.should == z
     z2[0].should equal(z2[2])
     z2[1].should equal(z2[3])
   end
 
   it "should emit proper labels and backreferences for recursive objects" do
     x = [1, 2]
     x << x
     x.to_yaml.should == "--- &id001\n  \n  - 1\n  - 2\n  - *id001"
     x2 = YAML.load(x.to_yaml)
     x2.should be_a(Array)
     x2.length.should == 3
     x2[0].should == 1
     x2[1].should == 2
     x2[2].should equal(x2)
   end
 end
+
+# Note, many of these tests will pass on Ruby 1.8 but fail on 1.9 if the patch
+# fix is not applied to Puppet or there's a regression.  These version
+# dependant failures are intentional since the string encoding behavior changed
+# significantly in 1.9.
+describe "UTF-8 encoded String#to_yaml (Bug #11246)" do
+  # JJM All of these snowmen are different representations of the same
+  # UTF-8 encoded string.
+  let(:snowman)         { 'Snowman: [☃]' }
+  let(:snowman_escaped) { "Snowman: [\xE2\x98\x83]" }
+
+  describe "UTF-8 String Literal" do
+    subject { snowman }
+
+    it "should serialize to YAML" do
+      subject.to_yaml
+    end
+    it "should serialize and deserialize to the same thing" do
+      YAML.load(subject.to_yaml).should == subject
+    end
+    it "should serialize and deserialize to a String compatible with a UTF-8 encoded Regexp" do
+      YAML.load(subject.to_yaml).should =~ /☃/u
+    end
+  end
+end