puppet-retrospec 1.5.0 → 1.6.0

data/vendor/pup410/lib/puppet/pops/parser/locator.rb

@@ -0,0 +1,321 @@
+module Puppet::Pops
+module Parser
+# Helper class that keeps track of where line breaks are located and can answer questions about positions.
+#
+class Locator
+  # Creates, or recreates a Locator. A Locator is created if index is not given (a scan is then
+  # performed of the given source string.
+  #
+  def self.locator(string, file, index = nil, char_offsets = false)
+    if char_offsets
+      LocatorForChars.new(string, file, index)
+    else
+      Locator19.new(string, file, index)
+    end
+  end
+
+  # Returns the file name associated with the string content
+  def file
+  end
+
+  # Returns the string content
+  def string
+  end
+
+  # Returns the position on line (first position on a line is 1)
+  def pos_on_line(offset)
+  end
+
+  # Returns the line number (first line is 1) for the given offset
+  def line_for_offset(offset)
+  end
+
+  # Returns the offset on line (first offset on a line is 0).
+  #
+  def offset_on_line(offset)
+  end
+
+  # Returns the character offset for a given reported offset
+  def char_offset(byte_offset)
+  end
+
+  # Returns the length measured in number of characters from the given start and end reported offset
+  def char_length(offset, end_offset)
+  end
+
+  # Returns the length measured in number of characters from the given start and end byte offseta
+  def char_length(offset, end_offset)
+  end
+
+  # Extracts the text from offset with given length (measured in what the locator uses for offset)
+  # @returns String - the extracted text
+  def extract_text(offset, length)
+  end
+
+  # Returns the line index - an array of line offsets for the start position of each line, starting at 0 for
+  # the first line.
+  #
+  def line_index()
+  end
+
+  # A Sublocator locates a concrete locator (subspace) in a virtual space.
+  # The `leading_line_count` is the (virtual) number of lines preceding the first line in the concrete locator.
+  # The `leading_offset` is the (virtual) byte offset of the first byte in the concrete locator.
+  # The `leading_line_offset` is the (virtual) offset / margin in characters for each line.
+  #
+  # This illustrates characters in the sublocator (`.`) inside the subspace (`X`):
+  #
+  #      1:XXXXXXXX
+  #      2:XXXX.... .. ... ..
+  #      3:XXXX. . .... ..
+  #      4:XXXX............
+  #
+  # This sublocator would be configured with leading_line_count = 1,
+  # leading_offset=8, and leading_line_offset=4
+  #
+  # Note that leading_offset must be the same for all lines and measured in characters.
+  #
+  class SubLocator < Locator
+    attr_reader :locator
+    attr_reader :leading_line_count
+    attr_reader :leading_offset
+    attr_reader :leading_line_offset
+
+    def self.sub_locator(string, file, leading_line_count, leading_offset, leading_line_offset)
+      self.new(Locator.locator(string, file),
+        leading_line_count,
+        leading_offset,
+        leading_line_offset)
+    end
+
+    def initialize(locator, leading_line_count, leading_offset, leading_line_offset)
+      @locator = locator
+      @leading_line_count = leading_line_count
+      @leading_offset = leading_offset
+      @leading_line_offset = leading_line_offset
+    end
+
+    def file
+      @locator.file
+    end
+
+    def string
+      @locator.string
+    end
+
+    # Given offset is offset in the subspace
+    def line_for_offset(offset)
+      @locator.line_for_offset(offset) + @leading_line_count
+    end
+
+    # Given offset is offset in the subspace
+    def offset_on_line(offset)
+      @locator.offset_on_line(offset) + @leading_line_offset
+    end
+
+    # Given offset is offset in the subspace
+    def char_offset(offset)
+      effective_line = @locator.line_for_offset(offset)
+      locator.char_offset(offset) + (effective_line * @leading_line_offset) + @leading_offset
+    end
+
+    # Given offsets are offsets in the subspace
+    def char_length(offset, end_offset)
+      effective_line = @locator.line_for_offset(end_offset) - @locator.line_for_offset(offset)
+      locator.char_length(offset, end_offset) + (effective_line * @leading_line_offset)
+    end
+
+    def pos_on_line(offset)
+      offset_on_line(offset) +1
+    end
+  end
+
+  private
+
+  class AbstractLocator < Locator
+    attr_accessor :line_index
+    attr_accessor :string
+    attr_accessor :prev_offset
+    attr_accessor :prev_line
+    attr_reader   :string
+    attr_reader   :file
+
+    # Create a locator based on a content string, and a boolean indicating if ruby version support multi-byte strings
+    # or not.
+    #
+    def initialize(string, file, line_index = nil)
+      @string = string.freeze
+      @file = file.freeze
+      @prev_offset = nil
+      @prev_line = nil
+      @line_index = line_index
+      compute_line_index if line_index.nil?
+    end
+
+    # Returns the position on line (first position on a line is 1)
+    def pos_on_line(offset)
+      offset_on_line(offset) +1
+    end
+
+    def to_location_hash(reported_offset, end_offset)
+      pos        = pos_on_line(reported_offset)
+      offset     = char_offset(reported_offset)
+      length     = char_length(reported_offset, end_offset)
+      start_line = line_for_offset(reported_offset)
+      { :line => start_line, :pos => pos, :offset => offset, :length => length}
+    end
+
+    # Returns the index of the smallest item for which the item > the given value
+    # This is a min binary search. Although written in Ruby it is only slightly slower than
+    # the corresponding method in C in Ruby 2.0.0 - the main benefit to use this method over
+    # the Ruby C version is that it returns the index (not the value) which means there is not need
+    # to have an additional structure to get the index (or record the index in the structure). This
+    # saves both memory and CPU. It also does not require passing a block that is called since this
+    # method is specialized to search the line index.
+    #
+    def ary_bsearch_i(ary, value)
+      low = 0
+      high = ary.length
+      mid = nil
+      smaller = false
+      satisfied = false
+      v = nil
+
+      while low < high do
+          mid = low + ((high - low) / 2)
+          v = (ary[mid] > value)
+          if v == true
+            satisfied = true
+            smaller = true
+          elsif !v
+            smaller = false
+          else
+            raise TypeError, "wrong argument, must be boolean or nil, got '#{v.class}'"
+          end
+
+          if smaller
+            high = mid
+          else
+            low = mid + 1;
+          end
+      end
+
+      return nil if low == ary.length
+      return nil if !satisfied
+      return low
+    end
+
+    def hash
+      [string, file, line_index].hash
+    end
+
+    # Equal method needed by serializer to perform tabulation
+    def eql?(o)
+      self.class == o.class && string == o.string && file == o.file && line_index == o.line_index
+    end
+
+    # Common impl for 18 and 19 since scanner is byte based
+    def compute_line_index
+      scanner = StringScanner.new(string)
+      result = [0] # first line starts at 0
+      while scanner.scan_until(/\n/)
+        result << scanner.pos
+      end
+      self.line_index = result.freeze
+    end
+
+    # Returns the line number (first line is 1) for the given offset
+    def line_for_offset(offset)
+      if prev_offset == offset
+        # use cache
+        return prev_line
+      end
+      if line_nbr = ary_bsearch_i(line_index, offset)
+        # cache
+        prev_offset = offset
+        prev_line = line_nbr
+        return line_nbr
+      end
+      # If not found it is after last
+      # clear cache
+      prev_offset = prev_line = nil
+      return line_index.size
+    end
+  end
+
+  class LocatorForChars < AbstractLocator
+
+    def offset_on_line(offset)
+      line_offset = line_index[ line_for_offset(offset)-1 ]
+      offset - line_offset
+    end
+
+    def char_offset(char_offset)
+      char_offset
+    end
+
+    def char_length(offset, end_offset)
+      end_offset - offset
+    end
+
+    # Extracts the text from char offset with given byte length
+    # @returns String - the extracted text
+    def extract_text(offset, length)
+      string.slice(offset, length)
+    end
+
+  end
+
+  # This implementation is for Ruby19 and Ruby20. It uses byteslice to get strings from byte based offsets.
+  # For Ruby20 this is faster than using the Stringscanner.charpos method (byteslice outperforms it, when
+  # strings are frozen).
+  #
+  class Locator19 < AbstractLocator
+    include Types::PuppetObject
+
+    def self._ptype
+      @type
+    end
+
+    # The Locator is not part of the Ecore model so no ObjectType is automatically inferred. Instead the
+    # type is explicitly added here.
+    # TODO: LocatorForChars is never added. It looks like it could be removed (remnant from Ruby 1.8 compatibility?)
+    # @api private
+    def self.register_ptype(loader, ir)
+      @type = Pcore::create_object_type(loader, ir, self, 'Puppet::AST::Locator', 'Any',
+        'string' => Types::PStringType::DEFAULT,
+        'file' => Types::PStringType::DEFAULT,
+        'line_index' => {
+          Types::KEY_TYPE => Types::POptionalType.new(Types::PArrayType.new(Types::PIntegerType::DEFAULT)),
+          Types::KEY_VALUE => nil
+        }
+      ).resolve(Types::TypeParser.singleton, loader)
+    end
+
+    # Returns the offset on line (first offset on a line is 0).
+    # Ruby 19 is multibyte but has no character position methods, must use byteslice
+    def offset_on_line(offset)
+      line_offset = line_index[ line_for_offset(offset)-1 ]
+      string.byteslice(line_offset, offset-line_offset).length
+    end
+
+    # Returns the character offset for a given byte offset
+    # Ruby 19 is multibyte but has no character position methods, must use byteslice
+    def char_offset(byte_offset)
+      string.byteslice(0, byte_offset).length
+    end
+
+    # Returns the length measured in number of characters from the given start and end byte offseta
+    def char_length(offset, end_offset)
+      string.byteslice(offset, end_offset - offset).length
+    end
+
+    # Extracts the text from byte offset with given byte length
+    # @returns String - the extracted text
+    def extract_text(offset, length)
+      string.byteslice(offset, length)
+    end
+  end
+end
+end
+end

data/vendor/pup410/lib/puppet/pops/parser/parser_support.rb

@@ -0,0 +1,251 @@
+require 'puppet/parser/functions'
+require 'puppet/parser/files'
+require 'puppet/resource/type_collection'
+require 'puppet/resource/type'
+require 'monitor'
+
+module Puppet::Pops
+module Parser
+# Supporting logic for the parser.
+# This supporting logic has slightly different responsibilities compared to the original Puppet::Parser::Parser.
+# It is only concerned with parsing.
+#
+class Parser
+  # Note that the name of the contained class and the file name (currently parser_support.rb)
+  # needs to be different as the class is generated by Racc, and this file (parser_support.rb) is included as a mix in
+  #
+
+  # Simplify access to the Model factory
+  # Note that the parser/parser support does not have direct knowledge about the Model.
+  # All model construction/manipulation is made by the Factory.
+  #
+  Factory = Model::Factory
+
+  attr_accessor :lexer
+  attr_reader :definitions
+
+  # Returns the token text of the given lexer token, or nil, if token is nil
+  def token_text t
+    return t if t.nil?
+    t = t.current if t.respond_to?(:current)
+    return t.value if t.is_a? Model::QualifiedName
+
+    # else it is a lexer token
+    t[:value]
+  end
+
+  # Produces the fully qualified name, with the full (current) namespace for a given name.
+  #
+  # This is needed because class bodies are lazily evaluated and an inner class' container(s) may not
+  # have been evaluated before some external reference is made to the inner class; its must therefore know its complete name
+  # before evaluation-time.
+  #
+  def classname(name)
+    [namespace, name].join('::').sub(/^::/, '')
+  end
+
+  # Raises a Parse error with location information. Information about file is always obtained from the
+  # lexer. Line and position is produced if the given semantic is a Positioned object and have been given an offset.
+  #
+  def error(semantic, message)
+    except = Puppet::ParseError.new(message)
+    if semantic.is_a?(LexerSupport::TokenValue)
+      except.file = semantic[:file];
+      except.line = semantic[:line];
+      except.pos = semantic[:pos];
+    else
+      semantic = semantic.current() if semantic.is_a?(Factory)
+
+      # Adapt the model so it is possible to get location information.
+      # The model may not have been added to the source tree, so give it the lexer's locator
+      # directly instead of searching for the root Program where the locator is normally stored.
+      #
+      if semantic.is_a?(Model::Positioned)
+        adapter = Adapters::SourcePosAdapter.adapt(semantic)
+        adapter.locator = @lexer.locator
+      else
+        adapter = nil
+      end
+      except.file = @lexer.locator.file
+      except.line = adapter.line if adapter
+      except.pos = adapter.pos if adapter
+    end
+    raise except
+  end
+
+  # Parses a file expected to contain pp DSL logic.
+  def parse_file(file)
+    unless Puppet::FileSystem.exist?(file)
+      unless file =~ /\.pp$/
+        file = file + ".pp"
+      end
+    end
+    @lexer.file = file
+    _parse()
+  end
+
+  def initialize()
+    @lexer = Lexer2.new
+    @namestack = []
+    @definitions = []
+  end
+
+  # This is a callback from the generated parser (when an error occurs while parsing)
+  #
+  def on_error(token,value,stack)
+    if token == 0 # denotes end of file
+      value_at = 'end of file'
+    else
+      value_at = "'#{value[:value]}'"
+    end
+    error = Issues::SYNTAX_ERROR.format(:where => value_at)
+    error = "#{error}, token: #{token}" if @yydebug
+
+    # Note, old parser had processing of "expected token here" - do not try to reinstate:
+    # The 'expected' is only of value at end of input, otherwise any parse error involving a
+    # start of a pair will be reported as expecting the close of the pair - e.g. "$x.each |$x {|", would
+    # report that "seeing the '{', the '}' is expected. That would be wrong.
+    # Real "expected" tokens are very difficult to compute (would require parsing of racc output data). Output of the stack
+    # could help, but can require extensive backtracking and produce many options.
+    #
+    # The lexer should handle the "expected instead of end of file for strings, and interpolation", other expectancies
+    # must be handled by the grammar. The lexer may have enqueued tokens far ahead - the lexer's opinion about this
+    # is not trustworthy.
+    #
+    file = nil
+    line = nil
+    pos  = nil
+    if token != 0
+      file = value[:file]
+      line = value[:line]
+      pos  = value[:pos]
+    else
+      # At end of input, use what the lexer thinks is the source file
+      file = lexer.file
+    end
+    file = nil unless file.is_a?(String) && !file.empty?
+    raise Puppet::ParseErrorWithIssue.new(error, file, line, pos, nil, issue_code = Issues::SYNTAX_ERROR.issue_code)
+  end
+
+  # Parses a String of pp DSL code.
+  #
+  def parse_string(code, path = '')
+    @lexer.lex_string(code, path)
+    _parse()
+  end
+
+  # Mark the factory wrapped model object with location information
+  # @return [Factory] the given factory
+  # @api private
+  #
+  def loc(factory, start_locateable, end_locateable = nil)
+    factory.record_position(start_locateable, end_locateable)
+  end
+
+  # Mark the factory wrapped heredoc model object with location information
+  # @return [Factory] the given factory
+  # @api private
+  #
+  def heredoc_loc(factory, start_locateabke, end_locateable = nil)
+    factory.record_heredoc_position(start_locatable, end_locatable)
+  end
+
+  def aryfy(o)
+    o = [o] unless o.is_a?(Array)
+    o
+  end
+
+  def namespace
+    @namestack.join('::')
+  end
+
+  def namestack(name)
+    @namestack << name
+  end
+
+  def namepop()
+    @namestack.pop
+  end
+
+  def add_definition(definition)
+    @definitions << definition.current
+    definition
+  end
+
+  def add_mapping(produces)
+    # The actual handling of mappings happens in PopsBridge
+    add_definition(produces)
+  end
+
+  # Transforms an array of expressions containing literal name expressions to calls if followed by an
+  # expression, or expression list
+  #
+  def transform_calls(expressions)
+    # Factory transform raises an error if a non qualified name is followed by an argument list
+    # since there is no way that that can be transformed back to sanity. This occurs in situations like this:
+    #
+    #  $a = 10, notice hello
+    #
+    # where the "10, notice" forms an argument list. The parser builds an Array with the expressions and includes
+    # the comma tokens to enable the error to be reported against the first comma.
+    #
+    begin
+      Factory.transform_calls(expressions)
+    rescue Factory::ArgsToNonCallError => e
+      # e.args[1] is the first comma token in the list
+      # e.name_expr is the function name expression
+      if e.name_expr.is_a?(Model::QualifiedName)
+        error(e.args[1], "attempt to pass argument list to the function '#{e.name_expr.value}' which cannot be called without parentheses")
+      else
+        error(e.args[1], "illegal comma separated argument list")
+      end
+    end
+  end
+
+  # Transforms a LEFT followed by the result of attribute_operations, this may be a call or an invalid sequence
+  def transform_resource_wo_title(left, resource, lbrace_token, rbrace_token)
+    Factory.transform_resource_wo_title(left, resource, lbrace_token, rbrace_token)
+  end
+
+  # Creates a program with the given body.
+  #
+  def create_program(body)
+    locator = @lexer.locator
+    Factory.PROGRAM(body, definitions, locator)
+  end
+
+  # Creates an empty program with a single No-op at the input's EOF offset with 0 length.
+  #
+  def create_empty_program()
+    locator = @lexer.locator
+    no_op = Factory.literal(nil)
+    # Create a synthetic NOOP token at EOF offset with 0 size. The lexer does not produce an EOF token that is
+    # visible to the grammar rules. Creating this token is mainly to reuse the positioning logic as it
+    # expects a token decorated with location information.
+    token_sym, token = @lexer.emit_completed([:NOOP,'',0], locator.string.bytesize)
+    loc(no_op, token)
+    # Program with a Noop
+    program = Factory.PROGRAM(no_op, [], locator)
+    program
+  end
+
+  # Performs the parsing and returns the resulting model.
+  # The lexer holds state, and this is setup with {#parse_string}, or {#parse_file}.
+  #
+  # @api private
+  #
+  def _parse()
+    begin
+      @yydebug = false
+      main = yyparse(@lexer,:scan)
+    end
+    return main
+  ensure
+    @lexer.clear
+    @namestack = []
+    @definitions = []
+  end
+
+end
+end
+end