pdf-reader 2.2.0 → 2.11.0

data/lib/pdf/reader/buffer.rb

@@ -1,4 +1,5 @@
 # coding: ASCII-8BIT
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -48,6 +49,18 @@ class PDF::Reader
     ID = "ID"
     FWD_SLASH = "/"
     NULL_BYTE = "\x00"
+    CR = "\r"
+    LF = "\n"
+    CRLF = "\r\n"
+    WHITE_SPACE = [LF, CR, ' ']
+
+    # Quite a few PDFs have trailing junk.
+    # This can be several k of nuls in some cases
+    # Allow for this here
+    TRAILING_BYTECOUNT = 5000
+
+    # must match whole tokens
+    DIGITS_ONLY = %r{\A\d+\z}
 
     attr_reader :pos
 
@@ -55,7 +68,7 @@ class PDF::Reader
     #
     # Params:
     #
-    #   io - an IO stream or string with the raw data to tokenise
+    #   io - an IO stream (usually a StringIO) with the raw data to tokenise
     #
     # options:
     #
@@ -86,9 +99,12 @@ class PDF::Reader
     #
     # options:
     #
-    #   :skip_eol - if true, the IO stream is advanced past a CRLF or LF that
-    #               is sitting under the io cursor.
-    #
+    #   :skip_eol - if true, the IO stream is advanced past a CRLF, CR or LF
+    #               that is sitting under the io cursor.
+    #   Note:
+    #   Skipping a bare CR is not spec-compliant.
+    #   This is because the data may start with LF.
+    #   However we check for CRLF first, so the ambiguity is avoided.
     def read(bytes, opts = {})
       reset_pos
 
@@ -97,9 +113,9 @@ class PDF::Reader
         str = @io.read(2)
         if str.nil?
           return nil
-        elsif str == "\r\n"
+        elsif str == CRLF # This MUST be done before checking for CR alone
           # do nothing
-        elsif str[0,1] == "\n"
+        elsif str[0, 1] == LF || str[0, 1] == CR # LF or CR alone
           @io.seek(-1, IO::SEEK_CUR)
         else
           @io.seek(-2, IO::SEEK_CUR)
@@ -127,8 +143,10 @@ class PDF::Reader
     #
     def find_first_xref_offset
       check_size_is_non_zero
-      @io.seek(-1024, IO::SEEK_END) rescue @io.seek(0)
-      data = @io.read(1024)
+      @io.seek(-TRAILING_BYTECOUNT, IO::SEEK_END) rescue @io.seek(0)
+      data = @io.read(TRAILING_BYTECOUNT)
+
+      raise MalformedPDFError, "PDF does not contain EOF marker" if data.nil?
 
       # the PDF 1.7 spec (section #3.4) says that EOL markers can be either \r, \n, or both.
       lines = data.split(/[\n\r]+/).reverse
@@ -136,7 +154,12 @@ class PDF::Reader
 
       raise MalformedPDFError, "PDF does not contain EOF marker" if eof_index.nil?
       raise MalformedPDFError, "PDF EOF marker does not follow offset" if eof_index >= lines.size-1
-      lines[eof_index+1].to_i
+      offset = lines[eof_index+1].to_i
+
+      # a byte offset < 0 doesn't make much sense. This is unlikely to happen, but in theory some
+      # corrupted PDFs might have a line that looks like a negative int preceding the `%%EOF`
+      raise MalformedPDFError, "invalid xref offset" if offset < 0
+      offset
     end
 
     private
@@ -217,45 +240,73 @@ class PDF::Reader
       return if @tokens.size < 3
       return if @tokens[2] != "R"
 
-      if @tokens[0].match(/\d+/) && @tokens[1].match(/\d+/)
-        @tokens[0] = PDF::Reader::Reference.new(@tokens[0].to_i, @tokens[1].to_i)
-        @tokens[1] = nil
-        @tokens[2] = nil
-        @tokens.compact!
+      token_one = @tokens[0]
+      token_two = @tokens[1]
+      if token_one.is_a?(String) && token_two.is_a?(String) && token_one.match(DIGITS_ONLY) && token_two.match(DIGITS_ONLY)
+        @tokens[0] = PDF::Reader::Reference.new(token_one.to_i, token_two.to_i)
+        @tokens.delete_at(2)
+        @tokens.delete_at(1)
       end
     end
 
+    # Extract data between ID and EI
+    # If the EI follows white-space the space is dropped from the data
+    # The EI must followed by white-space or end of buffer
+    # This is to reduce the chance of accidentally matching an embedded EI
     def prepare_inline_token
-      str = "".dup
-
-      buffer = []
-
-      until buffer[0] =~ /\s|\0/ && buffer[1, 2] == ["E", "I"]
+      idstart = @io.pos
+      prevchr = ''
+      eisize = 0 # how many chars in the end marker
+      seeking = 'E' # what are we looking for now?
+      loop do
         chr = @io.read(1)
-        buffer << chr
-
-        if buffer.length > 3
-          str << buffer.shift
+        break if chr.nil?
+        case seeking
+        when 'E'
+          if chr == 'E'
+            seeking = 'I'
+            if WHITE_SPACE.include? prevchr
+              eisize = 3 # include whitespace in delimiter, i.e. drop from data
+            else # assume the EI immediately follows the data
+              eisize = 2 # leave prevchr in data
+            end
+          end
+        when 'I'
+          if chr == 'I'
+            seeking = ''
+          else
+            seeking = 'E'
+          end
+        when ''
+          if WHITE_SPACE.include? chr
+            eisize += 1 # Drop trailer
+            break
+          else
+            seeking = 'E'
+          end
         end
+        prevchr = chr.is_a?(String) ? chr : ''
       end
-
-      str << NULL_BYTE if buffer.first == NULL_BYTE
-
-      @tokens << string_token(str)
-      @io.seek(-3, IO::SEEK_CUR) unless chr.nil?
+      unless seeking == ''
+        raise MalformedPDFError, "EI terminator not found"
+      end
+      eiend = @io.pos
+      @io.seek(idstart, IO::SEEK_SET)
+      str = @io.read(eiend - eisize - idstart) # get the ID content
+      @tokens << str.freeze if str
     end
 
     # if we're currently inside a hex string, read hex nibbles until
     # we find a closing >
     #
     def prepare_hex_token
+      finished = :false
       str = "".dup
-      finished = false
 
-      while !finished
+      until finished == :true
         byte = @io.getbyte
         if byte.nil?
-          finished = true # unbalanced params
+          finished = :true # unbalanced params
         elsif (48..57).include?(byte) || (65..90).include?(byte) || (97..122).include?(byte)
           str << byte
         elsif byte <= 32
@@ -264,7 +315,7 @@ class PDF::Reader
           @tokens << str if str.size > 0
           @tokens << ">" if byte != 0x3E # '>'
           @tokens << byte.chr
-          finished = true
+          finished = :true
         end
       end
     end
@@ -311,14 +362,17 @@ class PDF::Reader
     def prepare_regular_token
       tok = "".dup
 
-      while byte = @io.getbyte
+      loop do
+        byte = @io.getbyte
+
         case byte
+        when nil
+          break
         when 0x25
           # comment, ignore everything until the next EOL char
-          done = false
-          while !done
-            byte = @io.getbyte
-            done = true if byte.nil? || byte == 0x0A || byte == 0x0D
+          loop do
+            commentbyte = @io.getbyte
+            break if commentbyte.nil? || commentbyte == 0x0A || commentbyte == 0x0D
           end
         when *TOKEN_WHITESPACE
           # white space, token finished
@@ -388,15 +442,5 @@ class PDF::Reader
       byte
     end
 
-    # for a handful of tokens we want to tell the parser how to convert them
-    # into higher level tokens. This methods adds a to_token() method
-    # to tokens that should remain as strings.
-    #
-    def string_token(token)
-      def token.to_token
-        to_s
-      end
-      token
-    end
   end
 end

data/lib/pdf/reader/cid_widths.rb

@@ -1,8 +1,7 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
-#
-
 require 'forwardable'
 
 class PDF::Reader
@@ -32,10 +31,10 @@ class PDF::Reader
         params << array.shift
 
         if params.size == 2 && params.last.is_a?(Array)
-          widths.merge! parse_first_form(params.first, params.last)
+          widths.merge! parse_first_form(params.first.to_i, Array(params.last))
           params = []
         elsif params.size == 3
-          widths.merge! parse_second_form(params[0], params[1], params[2])
+          widths.merge! parse_second_form(params[0].to_i, params[1].to_i, params[2].to_i)
           params = []
         end
       end
@@ -53,6 +52,10 @@ class PDF::Reader
 
     # this is the form 10 20 123 where all index between 10 and 20 have width 123
     def parse_second_form(first, final, width)
+      if first > final
+        raise MalformedPDFError, "CidWidths: #{first} must be less than #{final}"
+      end
+
       (first..final).inject({}) { |accum, index|
         accum[index] = width
         accum

data/lib/pdf/reader/cmap.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -32,16 +33,17 @@ class PDF::Reader
   # extracting various useful information.
   #
   class CMap # :nodoc:
+
     CMAP_KEYWORDS = {
-      "begincodespacerange" => 1,
-      "endcodespacerange" => 1,
-      "beginbfchar" => 1,
-      "endbfchar" => 1,
-      "beginbfrange" => 1,
-      "endbfrange" => 1,
-      "begin" => 1,
-      "begincmap" => 1,
-      "def" => 1
+      "begincodespacerange" => :noop,
+      "endcodespacerange" => :noop,
+      "beginbfchar" => :noop,
+      "endbfchar" => :noop,
+      "beginbfrange" => :noop,
+      "endbfrange" => :noop,
+      "begin" => :noop,
+      "begincmap" => :noop,
+      "def" => :noop
     }
 
     attr_reader :map
@@ -51,30 +53,6 @@ class PDF::Reader
       process_data(data)
     end
 
-    def process_data(data)
-      parser = build_parser(data)
-      mode = nil
-      instructions = []
-
-      while token = parser.parse_token(CMAP_KEYWORDS)
-        if token == "beginbfchar"
-          mode = :char
-        elsif token == "endbfchar"
-          process_bfchar_instructions(instructions)
-          instructions = []
-          mode = nil
-        elsif token == "beginbfrange"
-          mode = :range
-        elsif token == "endbfrange"
-          process_bfrange_instructions(instructions)
-          instructions = []
-          mode = nil
-        elsif mode == :char || mode == :range
-          instructions << token
-        end
-      end
-    end
-
     def size
       @map.size
     end
@@ -84,44 +62,84 @@ class PDF::Reader
     # Returns an array of Integers.
     #
     def decode(c)
-      # TODO: implement the conversion
-      return c unless Integer === c
-      @map[c]
+      @map.fetch(c, [])
     end
 
     private
 
+    def process_data(data, initial_mode = :none)
+      parser = build_parser(data)
+      mode = initial_mode
+      instructions = []
+
+      while token = parser.parse_token(CMAP_KEYWORDS)
+        if token.is_a?(String) || token.is_a?(Array)
+          if token == "beginbfchar"
+            mode = :char
+          elsif token == "endbfchar"
+            process_bfchar_instructions(instructions)
+            instructions = []
+            mode = :none
+          elsif token == "beginbfrange"
+            mode = :range
+          elsif token == "endbfrange"
+            process_bfrange_instructions(instructions)
+            instructions = []
+            mode = :none
+          elsif mode == :char
+            instructions << token.to_s
+          elsif mode == :range
+            instructions << token
+          end
+        end
+      end
+    end
+
+
     def build_parser(instructions)
       buffer = Buffer.new(StringIO.new(instructions))
       Parser.new(buffer)
     end
 
+    # The following includes some manual decoding of UTF-16BE strings into unicode codepoints. In
+    # theory we could replace all the UTF-16 code with something based on Ruby's encoding support:
+    #
+    #    str.dup.force_encoding("utf-16be").encode!("utf-8").unpack("U*")
+    #
+    # However, some cmaps contain broken surrogate pairs and the ruby encoding support raises an
+    # exception when we try converting broken UTF-16 to UTF-8
+    #
     def str_to_int(str)
-      return nil if str.nil? || str.size == 0
-      unpacked_string = if str.size == 1 # UTF-8
+      unpacked_string = if str.bytesize == 1 # UTF-8
         str.unpack("C*")
       else # UTF-16
          str.unpack("n*")
       end
-      if unpacked_string.size == 1
-        unpacked_string
-      elsif unpacked_string.size == 2 && unpacked_string[0] > 0xD800 && unpacked_string[0] < 0xDBFF
-        # this is a Unicode UTF-16 "Surrogate Pair" see Unicode Spec. Chapter 3.7
-        # lets convert to a UTF-32. (the high bit is between 0xD800-0xDBFF, the
-        # low bit is between 0xDC00-0xDFFF) for example: U+1D44E (U+D835 U+DC4E)
-        [(unpacked_string[0] - 0xD800) * 0x400 + (unpacked_string[1] - 0xDC00) + 0x10000]
-      else
-        # it is a bad idea to just return the first 16 bits, as this doesn't allow
-        # for ligatures for example fi (U+0066 U+0069)
-        unpacked_string
+      result = []
+      while unpacked_string.any? do
+        if unpacked_string.size >= 2 &&
+            unpacked_string.first.to_i > 0xD800 &&
+            unpacked_string.first.to_i < 0xDBFF
+          # this is a Unicode UTF-16 "Surrogate Pair" see Unicode Spec. Chapter 3.7
+          # lets convert to a UTF-32. (the high bit is between 0xD800-0xDBFF, the
+          # low bit is between 0xDC00-0xDFFF) for example: U+1D44E (U+D835 U+DC4E)
+          point_one = unpacked_string.shift.to_i
+          point_two = unpacked_string.shift.to_i
+          result << (point_one - 0xD800) * 0x400 + (point_two - 0xDC00) + 0x10000
+        else
+          result << unpacked_string.shift
+        end
       end
+      result
     end
 
     def process_bfchar_instructions(instructions)
       instructions.each_slice(2) do |one, two|
-        find    = str_to_int(one)
-        replace = str_to_int(two)
-        @map[find.first] = replace
+        find    = str_to_int(one.to_s)
+        replace = str_to_int(two.to_s)
+        if find.any? && replace.any?
+          @map[find.first.to_i] = replace
+        end
       end
     end
 
@@ -132,30 +150,36 @@ class PDF::Reader
         elsif start.kind_of?(String) && finish.kind_of?(String) && to.kind_of?(Array)
           bfrange_type_two(start, finish, to)
         else
-          raise "invalid bfrange section"
+          raise MalformedPDFError, "invalid bfrange section"
         end
       end
     end
 
     def bfrange_type_one(start_code, end_code, dst)
-      start_code = str_to_int(start_code)[0]
-      end_code   = str_to_int(end_code)[0]
+      start_code = str_to_int(start_code).first
+      end_code   = str_to_int(end_code).first
       dst        = str_to_int(dst)
 
+      return if start_code.nil? || end_code.nil?
+
       # add all values in the range to our mapping
       (start_code..end_code).each_with_index do |val, idx|
-        @map[val] = dst.length == 1 ? [dst[0] + idx] : [dst[0], dst[1] + 1]
+        @map[val] = dst.length == 1 ? [dst[0].to_i + idx] : [dst[0].to_i, dst[1].to_i + 1]
       end
     end
 
     def bfrange_type_two(start_code, end_code, dst)
-      start_code = str_to_int(start_code)[0]
-      end_code   = str_to_int(end_code)[0]
+      start_code = str_to_int(start_code).first
+      end_code   = str_to_int(end_code).first
+
+      return if start_code.nil? || end_code.nil?
+
       from_range = (start_code..end_code)
 
       # add all values in the range to our mapping
       from_range.each_with_index do |val, idx|
-        @map[val] = str_to_int(dst[idx])
+        dst_char = dst[idx]
+        @map[val.to_i] = str_to_int(dst_char) if dst_char
       end
     end
   end

data/lib/pdf/reader/encoding.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -40,20 +41,22 @@ class PDF::Reader
       @mapping  = default_mapping # maps from character codes to Unicode codepoints
       @string_cache  = {} # maps from character codes to UTF-8 strings.
 
-      if enc.kind_of?(Hash)
-        self.differences = enc[:Differences] if enc[:Differences]
-        enc = enc[:Encoding] || enc[:BaseEncoding]
-      elsif enc != nil
-        enc = enc.to_sym
+      @enc_name = if enc.kind_of?(Hash)
+        enc[:Encoding] || enc[:BaseEncoding]
+      elsif enc && enc.respond_to?(:to_sym)
+        enc.to_sym
       else
-        enc = nil
+        :StandardEncoding
       end
 
-      @enc_name = enc
-      @unpack   = get_unpack(enc)
-      @map_file = get_mapping_file(enc)
+      @unpack   = get_unpack(@enc_name)
+      @map_file = get_mapping_file(@enc_name)
 
       load_mapping(@map_file) if @map_file
+
+      if enc.is_a?(Hash) && enc[:Differences]
+        self.differences = enc[:Differences]
+      end
     end
 
     # set the differences table for this encoding. should be an array in the following format:
@@ -66,16 +69,16 @@ class PDF::Reader
     #
     #   [25, :A, :B]
     def differences=(diff)
-      raise ArgumentError, "diff must be an array" unless diff.kind_of?(Array)
+      PDF::Reader::Error.validate_type(diff, "diff", Array)
 
       @differences = {}
       byte = 0
       diff.each do |val|
         if val.kind_of?(Numeric)
           byte = val.to_i
-        else
+        elsif codepoint = glyphlist.name_to_unicode(val)
           @differences[byte] = val
-          @mapping[byte] = glyphlist.name_to_unicode(val)
+          @mapping[byte] = codepoint
           byte += 1
         end
       end
@@ -164,7 +167,7 @@ class PDF::Reader
     end
 
     def convert_to_utf8(str)
-      ret = str.unpack(unpack).map! { |c| @mapping[c] || c }.pack("U*")
+      ret = str.unpack(unpack).map! { |c| @mapping[c.to_i] || c }.pack("U*")
       ret.force_encoding("UTF-8")
       ret
     end
@@ -206,7 +209,7 @@ class PDF::Reader
     def load_mapping(file)
       File.open(file, "r:BINARY") do |f|
         f.each do |l|
-          _m, single_byte, unicode = *l.match(/([0-9A-Za-z]+);([0-9A-F]{4})/)
+          _m, single_byte, unicode = *l.match(/\A([0-9A-Za-z]+);([0-9A-F]{4})/)
           @mapping["0x#{single_byte}".hex] = "0x#{unicode}".hex if single_byte
         end
       end

data/lib/pdf/reader/error.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -33,19 +34,30 @@ class PDF::Reader
     def self.str_assert(lvalue, rvalue, chars=nil)
       raise MalformedPDFError, "PDF malformed, expected string but found #{lvalue.class} instead" if chars and !lvalue.kind_of?(String)
       lvalue = lvalue[0,chars] if chars
-      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found #{lvalue} instead"  if lvalue != rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead"  if lvalue != rvalue
     end
     ################################################################################
     def self.str_assert_not(lvalue, rvalue, chars=nil)
       raise MalformedPDFError, "PDF malformed, expected string but found #{lvalue.class} instead" if chars and !lvalue.kind_of?(String)
       lvalue = lvalue[0,chars] if chars
-      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found #{lvalue} instead"  if lvalue == rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead"  if lvalue == rvalue
     end
     ################################################################################
     def self.assert_equal(lvalue, rvalue)
-      raise MalformedPDFError, "PDF malformed, expected #{rvalue} but found #{lvalue} instead" if lvalue != rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead" if lvalue != rvalue
     end
     ################################################################################
+    def self.validate_type(object, name, klass)
+      raise ArgumentError, "#{name} (#{object}) must be a #{klass}" unless object.is_a?(klass)
+    end
+    ################################################################################
+    def self.validate_type_as_malformed(object, name, klass)
+      raise MalformedPDFError, "#{name} (#{object}) must be a #{klass}" unless object.is_a?(klass)
+    end
+    ################################################################################
+    def self.validate_not_nil(object, name)
+      raise ArgumentError, "#{object} must not be nil" if object.nil?
+    end
   end
 
   ################################################################################

data/lib/pdf/reader/filter/ascii85.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 require 'ascii85'
@@ -7,6 +8,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the Ascii85 filter
     class Ascii85
+
       def initialize(options = {})
         @options = options
       end
@@ -17,7 +19,11 @@ class PDF::Reader
       #
       def filter(data)
         data = "<~#{data}" unless data.to_s[0,2] == "<~"
-        ::Ascii85::decode(data)
+        if defined?(::Ascii85Native)
+          ::Ascii85Native::decode(data)
+        else
+          ::Ascii85::decode(data)
+        end
       rescue Exception => e
         # Oops, there was a problem decoding the stream
         raise MalformedPDFError,

data/lib/pdf/reader/filter/ascii_hex.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 #
@@ -6,6 +7,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the AsciiHex stream filter
     class AsciiHex
+
       def initialize(options = {})
         @options = options
       end
@@ -16,9 +18,12 @@ class PDF::Reader
       def filter(data)
         data.chop! if data[-1,1] == ">"
         data = data[1,data.size] if data[0,1] == "<"
+
+        return "" if data.nil?
+
         data.gsub!(/[^A-Fa-f0-9]/,"")
         data << "0" if data.size % 2 == 1
-        data.scan(/.{2}/).map { |s| s.hex.chr }.join("")
+        data.scan(/.{2}/).flatten.map { |s| s.hex.chr }.join("")
       rescue Exception => e
         # Oops, there was a problem decoding the stream
         raise MalformedPDFError,