mortar 0.1.0

data/lib/mortar/project.rb

@@ -0,0 +1,156 @@
+#
+# Copyright 2012 Mortar Data Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+require 'fileutils'
+
+module Mortar
+  module Project
+    class ProjectError < RuntimeError; end
+    
+    class Project
+      def self.required_directories
+        ["macros", "pigscripts", "udfs"]
+      end
+      
+      attr_accessor :name
+      attr_accessor :remote
+      attr_accessor :root_path
+      
+      def initialize(name, root_path, remote)
+        @name = name
+        @root_path = root_path
+        @remote = remote
+      end
+      
+      def python_udfs_path
+        File.join(@root_path, "udfs/python")
+      end
+
+      def python_udfs
+        @python_udfs ||= PythonUDFs.new(
+          python_udfs_path,
+          "python",
+          ".py")
+      end
+      
+      def pigscripts_path
+        File.join(@root_path, "pigscripts")
+      end
+
+      def pigscripts
+        @pigscripts ||= PigScripts.new(
+          pigscripts_path,
+          "pigscripts",
+          ".pig")
+        @pigscripts
+      end
+      
+      def tmp_path
+        path = File.join(@root_path, "tmp")
+        unless File.directory? path
+          FileUtils.mkdir_p path
+        end
+        path
+      end
+    end
+    
+    class ProjectEntity
+      
+      include Enumerable
+      
+      def initialize(path, name, filename_extension)
+        @path = path
+        @name = name
+        @filename_extension = filename_extension
+        @elements = elements
+      end
+      
+      def method_missing(method, *args)
+        method_name = method.to_s
+        return @elements[method_name] if @elements[method_name]
+        super
+      end
+      
+      def each
+        @elements.each {|key, value| yield [key, value]}
+      end
+      
+      def [](key)
+        @elements[key]
+      end
+      
+      def keys
+        @elements.keys
+      end
+      
+      protected
+      
+      def element_name(element_path)
+        File.basename(element_path, @filename_extension)
+      end
+
+      def elements
+        unless File.directory? @path
+          raise ProjectError, "Unable to find #{@name} directory in project"
+        end
+
+        # get {script_name => full_path}
+        file_paths = Dir[File.join(@path, "**", "*#{@filename_extension}")]
+        file_paths_hsh = file_paths.collect{|element_path| [element_name(element_path), element(element_name(element_path), element_path)]}.flatten
+        Hash[*file_paths_hsh]
+      end
+      
+      def element(path)
+        raise NotImplementedError, "Implement in subclass"
+      end
+    end
+    
+    class PigScripts < ProjectEntity
+      def element(name, path)
+        Script.new(name, path)
+      end
+    end
+
+    class PythonUDFs < ProjectEntity
+      def element(name, path)
+        Script.new(name, path)
+      end
+    end
+    
+    class Script
+      
+      attr_reader :name
+      attr_reader :path
+      
+      def initialize(name, path)
+        @name = name
+        @path = path
+      end
+
+      def code
+        script_file = File.open(@path, "r")
+        script_contents = script_file.read
+        script_file.close
+        script_contents
+      end
+      
+      def to_s
+        code
+      end
+    end
+    
+  end
+end

data/lib/mortar/snapshot.rb

@@ -0,0 +1,39 @@
+#
+# Copyright 2012 Mortar Data Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+module Mortar
+  module Snapshot
+
+    extend self
+    
+    def create_and_push_snapshot_branch(git, project)
+      # create / push a snapshot branch
+      snapshot_branch = action("Taking code snapshot") do
+        git.create_snapshot_branch()
+      end
+
+      git_ref = action("Sending code snapshot to Mortar") do
+        # push the code
+        git.push(project.remote, snapshot_branch)
+
+        # grab the commit hash and clean out the branch from the local branches
+        ref = git.git_ref(snapshot_branch)
+        git.branch_delete(snapshot_branch)
+        ref
+      end
+    end
+  end
+end

data/lib/mortar/templates/macro/macro.pig

@@ -0,0 +1,14 @@
+/**
+ * <%= macro_name %>: Pig macros for use in pigscripts.
+ *
+ */
+
+/**
+ * A simple example macro function that returns the entity passed in.
+ */
+DEFINE <%= macro_name.capitalize %>_EXAMPLE(input_relation)
+returns output_relation {
+    -- just an example
+    $output_relation = FOREACH $input_relation
+                      GENERATE *;
+};

data/lib/mortar/templates/pigscript/pigscript.pig

@@ -0,0 +1,38 @@
+/**
+ * <%= script_name %>
+ *
+ * Required parameters:
+ *
+ * -param INPUT_PATH Input path for script data (e.g. s3n://hawk-example-data/tutorial/excite.log.bz2)
+ * -param OUTPUT_PATH Output path for script data (e.g. s3n://my-output-bucket/<%= script_name %>)
+ */
+
+<% if not options[:skip_udf] %>
+/**
+ * User-Defined Functions (UDFs)
+ */
+
+REGISTER '../udfs/python/<%= script_name %>.py' using streaming_python as <%= script_name %>;
+<% end %>
+
+-- This is an example of loading up input data
+my_input_data = LOAD '$INPUT_PATH' 
+               USING PigStorage('\t') 
+                  AS (field0:chararray, field1:chararray, field2:chararray);
+
+-- This is an example pig operation
+filtered = FILTER my_input_data
+               BY field0 IS NOT NULL;
+
+-- This is an example call to a python user-defined function
+with_udf_output = FOREACH filtered 
+                 GENERATE field0..field2, 
+                          <%= script_name %>.example_udf(field0) AS example_udf_field;
+
+-- remove any existing data
+rmf $OUTPUT_PATH;
+
+-- store the results
+STORE with_udf_output 
+ INTO '$OUTPUT_PATH' 
+USING PigStorage('\t');

data/lib/mortar/templates/pigscript/python_udf.py

@@ -0,0 +1,13 @@
+from pig_util import outputSchema
+
+# 
+# This is where we write python UDFs (User-Defined Functions) that we can call from pig.
+# Pig needs to know the schema of the data coming out of the function, 
+# which we specify using the @outputSchema decorator.
+#
+@outputSchema('example_udf:int')
+def example_udf(input_str):
+    """
+    A simple example function that just returns the length of the string passed in.
+    """
+    return len(input_str) if input_str else None

data/lib/mortar/templates/project/Gemfile

@@ -0,0 +1,3 @@
+source "http://rubygems.org"
+
+gem "mortar"

data/lib/mortar/templates/project/README.md

@@ -0,0 +1,8 @@
+# Weclome to Mortar!
+
+Copy TK
+
+
+# Getting Started
+
+Copy TK

data/lib/mortar/templates/project/gitignore

@@ -0,0 +1,4 @@
+*.pyc
+*.class
+*.log
+tmp

data/lib/mortar/templates/project/pigscripts/pigscript.pig

@@ -0,0 +1,35 @@
+/**
+ * <%= project_name %>
+ *
+ * Required parameters:
+ *
+ * -param INPUT_PATH Input path for script data (e.g. s3n://hawk-example-data/tutorial/excite.log.bz2)
+ * -param OUTPUT_PATH Output path for script data (e.g. s3n://my-output-bucket/<%= project_name %>)
+ */
+
+/**
+ * User-Defined Functions (UDFs)
+ */
+REGISTER '../udfs/python/<%= project_name %>.py' using streaming_python as <%= project_name %>;
+
+-- This is an example of loading up input data
+my_input_data = LOAD '$INPUT_PATH' 
+               USING PigStorage('\t') 
+                  AS (field0:chararray, field1:chararray, field2:chararray);
+
+-- This is an example pig operation
+filtered = FILTER my_input_data
+               BY field0 IS NOT NULL;
+
+-- This is an example call to a python user-defined function
+with_udf_output = FOREACH filtered 
+                 GENERATE field0..field2, 
+                          <%= project_name %>.example_udf(field0) AS example_udf_field;
+
+-- remove any existing data
+rmf $OUTPUT_PATH;
+
+-- store the results
+STORE with_udf_output 
+ INTO '$OUTPUT_PATH' 
+USING PigStorage('\t');

data/lib/mortar/templates/project/udfs/python/python_udf.py

@@ -0,0 +1,13 @@
+from pig_util import outputSchema
+
+# 
+# This is where we write python UDFs (User-Defined Functions) that we can call from pig.
+# Pig needs to know the schema of the data coming out of the function, 
+# which we specify using the @outputSchema decorator.
+#
+@outputSchema('example_udf:int')
+def example_udf(input_str):
+    """
+    A simple example function that just returns the length of the string passed in.
+    """
+    return len(input_str) if input_str else None

data/lib/mortar/templates/udf/python_udf.py

@@ -0,0 +1,13 @@
+from pig_util import outputSchema
+
+# 
+# This is where we write python UDFs (User-Defined Functions) that we can call from pig.
+# Pig needs to know the schema of the data coming out of the function, 
+# which we specify using the @outputSchema decorator.
+#
+@outputSchema('example_udf:int')
+def example_udf(input_str):
+    """
+    A simple example function that just returns the length of the string passed in.
+    """
+    return len(input_str) if input_str else None

data/lib/mortar/version.rb

@@ -0,0 +1,20 @@
+#
+# Copyright 2012 Mortar Data Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+module Mortar
+  # see http://semver.org/
+  VERSION = "0.1.0"
+end

data/lib/vendor/mortar/okjson.rb

@@ -0,0 +1,598 @@
+# encoding: UTF-8
+#
+# Copyright 2011, 2012 Keith Rarick
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+# THE SOFTWARE.
+
+# See https://github.com/kr/okjson for updates.
+
+require 'stringio'
+
+# Some parts adapted from
+# http://golang.org/src/pkg/json/decode.go and
+# http://golang.org/src/pkg/utf8/utf8.go
+module Mortar
+  module OkJson
+    extend self
+
+
+    # Decodes a json document in string s and
+    # returns the corresponding ruby value.
+    # String s must be valid UTF-8. If you have
+    # a string in some other encoding, convert
+    # it first.
+    #
+    # String values in the resulting structure
+    # will be UTF-8.
+    def decode(s)
+      ts = lex(s)
+      v, ts = textparse(ts)
+      if ts.length > 0
+        raise Error, 'trailing garbage'
+      end
+      v
+    end
+
+
+    # Parses a "json text" in the sense of RFC 4627.
+    # Returns the parsed value and any trailing tokens.
+    # Note: this is almost the same as valparse,
+    # except that it does not accept atomic values.
+    def textparse(ts)
+      if ts.length < 0
+        raise Error, 'empty'
+      end
+
+      typ, _, val = ts[0]
+      case typ
+      when '{' then objparse(ts)
+      when '[' then arrparse(ts)
+      else
+        raise Error, "unexpected #{val.inspect}"
+      end
+    end
+
+
+    # Parses a "value" in the sense of RFC 4627.
+    # Returns the parsed value and any trailing tokens.
+    def valparse(ts)
+      if ts.length < 0
+        raise Error, 'empty'
+      end
+
+      typ, _, val = ts[0]
+      case typ
+      when '{' then objparse(ts)
+      when '[' then arrparse(ts)
+      when :val,:str then [val, ts[1..-1]]
+      else
+        raise Error, "unexpected #{val.inspect}"
+      end
+    end
+
+
+    # Parses an "object" in the sense of RFC 4627.
+    # Returns the parsed value and any trailing tokens.
+    def objparse(ts)
+      ts = eat('{', ts)
+      obj = {}
+
+      if ts[0][0] == '}'
+        return obj, ts[1..-1]
+      end
+
+      k, v, ts = pairparse(ts)
+      obj[k] = v
+
+      if ts[0][0] == '}'
+        return obj, ts[1..-1]
+      end
+
+      loop do
+        ts = eat(',', ts)
+
+        k, v, ts = pairparse(ts)
+        obj[k] = v
+
+        if ts[0][0] == '}'
+          return obj, ts[1..-1]
+        end
+      end
+    end
+
+
+    # Parses a "member" in the sense of RFC 4627.
+    # Returns the parsed values and any trailing tokens.
+    def pairparse(ts)
+      (typ, _, k), ts = ts[0], ts[1..-1]
+      if typ != :str
+        raise Error, "unexpected #{k.inspect}"
+      end
+      ts = eat(':', ts)
+      v, ts = valparse(ts)
+      [k, v, ts]
+    end
+
+
+    # Parses an "array" in the sense of RFC 4627.
+    # Returns the parsed value and any trailing tokens.
+    def arrparse(ts)
+      ts = eat('[', ts)
+      arr = []
+
+      if ts[0][0] == ']'
+        return arr, ts[1..-1]
+      end
+
+      v, ts = valparse(ts)
+      arr << v
+
+      if ts[0][0] == ']'
+        return arr, ts[1..-1]
+      end
+
+      loop do
+        ts = eat(',', ts)
+
+        v, ts = valparse(ts)
+        arr << v
+
+        if ts[0][0] == ']'
+          return arr, ts[1..-1]
+        end
+      end
+    end
+
+
+    def eat(typ, ts)
+      if ts[0][0] != typ
+        raise Error, "expected #{typ} (got #{ts[0].inspect})"
+      end
+      ts[1..-1]
+    end
+
+
+    # Scans s and returns a list of json tokens,
+    # excluding white space (as defined in RFC 4627).
+    def lex(s)
+      ts = []
+      while s.length > 0
+        typ, lexeme, val = tok(s)
+        if typ == nil
+          raise Error, "invalid character at #{s[0,10].inspect}"
+        end
+        if typ != :space
+          ts << [typ, lexeme, val]
+        end
+        s = s[lexeme.length..-1]
+      end
+      ts
+    end
+
+
+    # Scans the first token in s and
+    # returns a 3-element list, or nil
+    # if s does not begin with a valid token.
+    #
+    # The first list element is one of
+    # '{', '}', ':', ',', '[', ']',
+    # :val, :str, and :space.
+    #
+    # The second element is the lexeme.
+    #
+    # The third element is the value of the
+    # token for :val and :str, otherwise
+    # it is the lexeme.
+    def tok(s)
+      case s[0]
+      when ?{  then ['{', s[0,1], s[0,1]]
+      when ?}  then ['}', s[0,1], s[0,1]]
+      when ?:  then [':', s[0,1], s[0,1]]
+      when ?,  then [',', s[0,1], s[0,1]]
+      when ?[  then ['[', s[0,1], s[0,1]]
+      when ?]  then [']', s[0,1], s[0,1]]
+      when ?n  then nulltok(s)
+      when ?t  then truetok(s)
+      when ?f  then falsetok(s)
+      when ?"  then strtok(s)
+      when Spc then [:space, s[0,1], s[0,1]]
+      when ?\t then [:space, s[0,1], s[0,1]]
+      when ?\n then [:space, s[0,1], s[0,1]]
+      when ?\r then [:space, s[0,1], s[0,1]]
+      else          numtok(s)
+      end
+    end
+
+
+    def nulltok(s);  s[0,4] == 'null'  ? [:val, 'null',  nil]   : [] end
+    def truetok(s);  s[0,4] == 'true'  ? [:val, 'true',  true]  : [] end
+    def falsetok(s); s[0,5] == 'false' ? [:val, 'false', false] : [] end
+
+
+    def numtok(s)
+      m = /-?([1-9][0-9]+|[0-9])([.][0-9]+)?([eE][+-]?[0-9]+)?/.match(s)
+      if m && m.begin(0) == 0
+        if m[3] && !m[2]
+          [:val, m[0], Integer(m[1])*(10**Integer(m[3][1..-1]))]
+        elsif m[2]
+          [:val, m[0], Float(m[0])]
+        else
+          [:val, m[0], Integer(m[0])]
+        end
+      else
+        []
+      end
+    end
+
+
+    def strtok(s)
+      m = /"([^"\\]|\\["\/\\bfnrt]|\\u[0-9a-fA-F]{4})*"/.match(s)
+      if ! m
+        raise Error, "invalid string literal at #{abbrev(s)}"
+      end
+      [:str, m[0], unquote(m[0])]
+    end
+
+
+    def abbrev(s)
+      t = s[0,10]
+      p = t['`']
+      t = t[0,p] if p
+      t = t + '...' if t.length < s.length
+      '`' + t + '`'
+    end
+
+
+    # Converts a quoted json string literal q into a UTF-8-encoded string.
+    # The rules are different than for Ruby, so we cannot use eval.
+    # Unquote will raise an error if q contains control characters.
+    def unquote(q)
+      q = q[1...-1]
+      a = q.dup # allocate a big enough string
+      rubydoesenc = false
+      # In ruby >= 1.9, a[w] is a codepoint, not a byte.
+      if a.class.method_defined?(:force_encoding)
+        a.force_encoding('UTF-8')
+        rubydoesenc = true
+      end
+      r, w = 0, 0
+      while r < q.length
+        c = q[r]
+        case true
+        when c == ?\\
+          r += 1
+          if r >= q.length
+            raise Error, "string literal ends with a \"\\\": \"#{q}\""
+          end
+
+          case q[r]
+          when ?",?\\,?/,?'
+            a[w] = q[r]
+            r += 1
+            w += 1
+          when ?b,?f,?n,?r,?t
+            a[w] = Unesc[q[r]]
+            r += 1
+            w += 1
+          when ?u
+            r += 1
+            uchar = begin
+              hexdec4(q[r,4])
+            rescue RuntimeError => e
+              raise Error, "invalid escape sequence \\u#{q[r,4]}: #{e}"
+            end
+            r += 4
+            if surrogate? uchar
+              if q.length >= r+6
+                uchar1 = hexdec4(q[r+2,4])
+                uchar = subst(uchar, uchar1)
+                if uchar != Ucharerr
+                  # A valid pair; consume.
+                  r += 6
+                end
+              end
+            end
+            if rubydoesenc
+              a[w] = '' << uchar
+              w += 1
+            else
+              w += ucharenc(a, w, uchar)
+            end
+          else
+            raise Error, "invalid escape char #{q[r]} in \"#{q}\""
+          end
+        when c == ?", c < Spc
+          raise Error, "invalid character in string literal \"#{q}\""
+        else
+          # Copy anything else byte-for-byte.
+          # Valid UTF-8 will remain valid UTF-8.
+          # Invalid UTF-8 will remain invalid UTF-8.
+          # In ruby >= 1.9, c is a codepoint, not a byte,
+          # in which case this is still what we want.
+          a[w] = c
+          r += 1
+          w += 1
+        end
+      end
+      a[0,w]
+    end
+
+
+    # Encodes unicode character u as UTF-8
+    # bytes in string a at position i.
+    # Returns the number of bytes written.
+    def ucharenc(a, i, u)
+      case true
+      when u <= Uchar1max
+        a[i] = (u & 0xff).chr
+        1
+      when u <= Uchar2max
+        a[i+0] = (Utag2 | ((u>>6)&0xff)).chr
+        a[i+1] = (Utagx | (u&Umaskx)).chr
+        2
+      when u <= Uchar3max
+        a[i+0] = (Utag3 | ((u>>12)&0xff)).chr
+        a[i+1] = (Utagx | ((u>>6)&Umaskx)).chr
+        a[i+2] = (Utagx | (u&Umaskx)).chr
+        3
+      else
+        a[i+0] = (Utag4 | ((u>>18)&0xff)).chr
+        a[i+1] = (Utagx | ((u>>12)&Umaskx)).chr
+        a[i+2] = (Utagx | ((u>>6)&Umaskx)).chr
+        a[i+3] = (Utagx | (u&Umaskx)).chr
+        4
+      end
+    end
+
+
+    def hexdec4(s)
+      if s.length != 4
+        raise Error, 'short'
+      end
+      (nibble(s[0])<<12) | (nibble(s[1])<<8) | (nibble(s[2])<<4) | nibble(s[3])
+    end
+
+
+    def subst(u1, u2)
+      if Usurr1 <= u1 && u1 < Usurr2 && Usurr2 <= u2 && u2 < Usurr3
+        return ((u1-Usurr1)<<10) | (u2-Usurr2) + Usurrself
+      end
+      return Ucharerr
+    end
+
+
+    def surrogate?(u)
+      Usurr1 <= u && u < Usurr3
+    end
+
+
+    def nibble(c)
+      case true
+      when ?0 <= c && c <= ?9 then c.ord - ?0.ord
+      when ?a <= c && c <= ?z then c.ord - ?a.ord + 10
+      when ?A <= c && c <= ?Z then c.ord - ?A.ord + 10
+      else
+        raise Error, "invalid hex code #{c}"
+      end
+    end
+
+
+    # Encodes x into a json text. It may contain only
+    # Array, Hash, String, Numeric, true, false, nil.
+    # (Note, this list excludes Symbol.)
+    # X itself must be an Array or a Hash.
+    # No other value can be encoded, and an error will
+    # be raised if x contains any other value, such as
+    # Nan, Infinity, Symbol, and Proc, or if a Hash key
+    # is not a String.
+    # Strings contained in x must be valid UTF-8.
+    def encode(x)
+      case x
+      when Hash    then objenc(x)
+      when Array   then arrenc(x)
+      else
+        raise Error, 'root value must be an Array or a Hash'
+      end
+    end
+
+
+    def valenc(x)
+      case x
+      when Hash    then objenc(x)
+      when Array   then arrenc(x)
+      when String  then strenc(x)
+      when Numeric then numenc(x)
+      when true    then "true"
+      when false   then "false"
+      when nil     then "null"
+      else
+        raise Error, "cannot encode #{x.class}: #{x.inspect}"
+      end
+    end
+
+
+    def objenc(x)
+      '{' + x.map{|k,v| keyenc(k) + ':' + valenc(v)}.join(',') + '}'
+    end
+
+
+    def arrenc(a)
+      '[' + a.map{|x| valenc(x)}.join(',') + ']'
+    end
+
+
+    def keyenc(k)
+      case k
+      when String then strenc(k)
+      else
+        raise Error, "Hash key is not a string: #{k.inspect}"
+      end
+    end
+
+
+    def strenc(s)
+      t = StringIO.new
+      t.putc(?")
+      r = 0
+
+      # In ruby >= 1.9, s[r] is a codepoint, not a byte.
+      rubydoesenc = s.class.method_defined?(:encoding)
+
+      while r < s.length
+        case s[r]
+        when ?"  then t.print('\\"')
+        when ?\\ then t.print('\\\\')
+        when ?\b then t.print('\\b')
+        when ?\f then t.print('\\f')
+        when ?\n then t.print('\\n')
+        when ?\r then t.print('\\r')
+        when ?\t then t.print('\\t')
+        else
+          c = s[r]
+          case true
+          when rubydoesenc
+            begin
+              c.ord # will raise an error if c is invalid UTF-8
+              t.write(c)
+            rescue
+              t.write(Ustrerr)
+            end
+          when Spc <= c && c <= ?~
+            t.putc(c)
+          else
+            n = ucharcopy(t, s, r) # ensure valid UTF-8 output
+            r += n - 1 # r is incremented below
+          end
+        end
+        r += 1
+      end
+      t.putc(?")
+      t.string
+    end
+
+
+    def numenc(x)
+      if ((x.nan? || x.infinite?) rescue false)
+        raise Error, "Numeric cannot be represented: #{x}"
+      end
+      "#{x}"
+    end
+
+
+    # Copies the valid UTF-8 bytes of a single character
+    # from string s at position i to I/O object t, and
+    # returns the number of bytes copied.
+    # If no valid UTF-8 char exists at position i,
+    # ucharcopy writes Ustrerr and returns 1.
+    def ucharcopy(t, s, i)
+      n = s.length - i
+      raise Utf8Error if n < 1
+
+      c0 = s[i].ord
+
+      # 1-byte, 7-bit sequence?
+      if c0 < Utagx
+        t.putc(c0)
+        return 1
+      end
+
+      raise Utf8Error if c0 < Utag2 # unexpected continuation byte?
+
+      raise Utf8Error if n < 2 # need continuation byte
+      c1 = s[i+1].ord
+      raise Utf8Error if c1 < Utagx || Utag2 <= c1
+
+      # 2-byte, 11-bit sequence?
+      if c0 < Utag3
+        raise Utf8Error if ((c0&Umask2)<<6 | (c1&Umaskx)) <= Uchar1max
+        t.putc(c0)
+        t.putc(c1)
+        return 2
+      end
+
+      # need second continuation byte
+      raise Utf8Error if n < 3
+
+      c2 = s[i+2].ord
+      raise Utf8Error if c2 < Utagx || Utag2 <= c2
+
+      # 3-byte, 16-bit sequence?
+      if c0 < Utag4
+        u = (c0&Umask3)<<12 | (c1&Umaskx)<<6 | (c2&Umaskx)
+        raise Utf8Error if u <= Uchar2max
+        t.putc(c0)
+        t.putc(c1)
+        t.putc(c2)
+        return 3
+      end
+
+      # need third continuation byte
+      raise Utf8Error if n < 4
+      c3 = s[i+3].ord
+      raise Utf8Error if c3 < Utagx || Utag2 <= c3
+
+      # 4-byte, 21-bit sequence?
+      if c0 < Utag5
+        u = (c0&Umask4)<<18 | (c1&Umaskx)<<12 | (c2&Umaskx)<<6 | (c3&Umaskx)
+        raise Utf8Error if u <= Uchar3max
+        t.putc(c0)
+        t.putc(c1)
+        t.putc(c2)
+        t.putc(c3)
+        return 4
+      end
+
+      raise Utf8Error
+    rescue Utf8Error
+      t.write(Ustrerr)
+      return 1
+    end
+
+
+    class Utf8Error < ::StandardError
+    end
+
+
+    class Error < ::StandardError
+    end
+
+
+    Utagx = 0x80 # 1000 0000
+    Utag2 = 0xc0 # 1100 0000
+    Utag3 = 0xe0 # 1110 0000
+    Utag4 = 0xf0 # 1111 0000
+    Utag5 = 0xF8 # 1111 1000
+    Umaskx = 0x3f # 0011 1111
+    Umask2 = 0x1f # 0001 1111
+    Umask3 = 0x0f # 0000 1111
+    Umask4 = 0x07 # 0000 0111
+    Uchar1max = (1<<7) - 1
+    Uchar2max = (1<<11) - 1
+    Uchar3max = (1<<16) - 1
+    Ucharerr = 0xFFFD # unicode "replacement char"
+    Ustrerr = "\xef\xbf\xbd" # unicode "replacement char"
+    Usurrself = 0x10000
+    Usurr1 = 0xd800
+    Usurr2 = 0xdc00
+    Usurr3 = 0xe000
+
+    Spc = ' '[0]
+    Unesc = {?b=>?\b, ?f=>?\f, ?n=>?\n, ?r=>?\r, ?t=>?\t}
+  end
+end