pdf-reader 0.5.1 → 0.6

data/lib/pdf/reader/parser.rb

@@ -47,19 +47,19 @@ class PDF::Reader
       token = @buffer.token
 
       case token
-      when "/"                        : return Name.new(@buffer.token)
-      when "<<"                       : return dictionary()
-      when "["                        : return array()
-      when "("                        : return string()
-      when "<"                        : return hex_string()
-      when "true"                     : return true
-      when "false"                    : return false
-      when "null"                     : return nil
-      when "obj", "endobj"            : return Token.new(token)
-      when "stream", "endstream"      : return Token.new(token)
-      when ">>", "]", ">"             : return Token.new(token)
+      when "/"                        then return Name.new(@buffer.token)
+      when "<<"                       then return dictionary()
+      when "["                        then return array()
+      when "("                        then return string()
+      when "<"                        then return hex_string()
+      when "true"                     then return true
+      when "false"                    then return false
+      when "null"                     then return nil
+      when "obj", "endobj"            then return Token.new(token)
+      when "stream", "endstream"      then return Token.new(token)
+      when ">>", "]", ">"             then return Token.new(token)
       else                          
-        if operators.has_key?(token)  : return Token.new(token)
+        if operators.has_key?(token)  then return Token.new(token)
         else                            return token.to_f
         end
       end
@@ -132,15 +132,15 @@ class PDF::Reader
         when "\\"
           to_remove = 2
           case @buffer.raw[1, 1]
-          when ""   : to_remove = 1
-          when "n"  : str << "\n"
-          when "r"  : str << "\r"
-          when "t"  : str << "\t"
-          when "b"  : str << "\b"
-          when "f"  : str << "\f"
-          when "("  : str << "("
-          when ")"  : str << ")"
-          when "\\" : str << "\\"
+          when ""   then to_remove = 1
+          when "n"  then str << "\n"
+          when "r"  then str << "\r"
+          when "t"  then str << "\t"
+          when "b"  then str << "\b"
+          when "f"  then str << "\f"
+          when "("  then str << "("
+          when ")"  then str << ")"
+          when "\\" then str << "\\"
           else
             if m = @buffer.raw.match(/^\\(\d{1,3})/)
               to_remove = m[0].size
@@ -168,8 +168,8 @@ class PDF::Reader
       post_obj = parse_token
 
       case post_obj
-      when "endobj"   : return obj
-      when "stream"   : return stream(obj)
+      when "endobj"   then return obj
+      when "stream"   then return stream(obj)
       else              raise MalformedPDFError, "PDF malformed, unexpected token #{post_obj}"
       end
     end
@@ -177,8 +177,7 @@ class PDF::Reader
     # Decodes the contents of a PDF Stream and returns it as a Ruby String.
     def stream (dict)
       raise MalformedPDFError, "PDF malformed, missing stream length" unless dict.has_key?('Length')
-      dict['Length'] = @xref.object(dict['Length']) if dict['Length'].kind_of?(Reference)
-      data = @buffer.read(dict['Length'])
+      data = @buffer.read(@xref.object(dict['Length']))
       Error.str_assert(parse_token, "endstream")
       Error.str_assert(parse_token, "endobj")
 
@@ -186,14 +185,17 @@ class PDF::Reader
         options = []
 
         if dict.has_key?('DecodeParms')
-          options = dict['DecodeParms'].to_a
+          options = Array(dict['DecodeParms'])
         end
 
-        dict['Filter'].to_a.each_with_index do |filter, index|
+        Array(dict['Filter']).each_with_index do |filter, index|
           data = Filter.new(filter, options[index]).filter(data)
         end
       end
 
+      # this stream is a cmap
+      data = PDF::Reader::CMap.new(data) if data.include?("begincmap") && data.include?("endcmap")
+
       data
     end
     ################################################################################

data/lib/pdf/reader/register_receiver.rb

@@ -12,7 +12,54 @@ class PDF::Reader
     end
 
     def method_missing(methodname, *args)
-      callbacks << methodname.to_sym
+      callbacks << {:name => methodname.to_sym, :args => args}
+    end
+
+    # count the number of times a callback fired
+    def count(methodname)
+      counter = 0
+      callbacks.each { |cb| counter += 1 if cb[:name] == methodname}
+      return counter
+    end
+
+    # return the details for the first time the specified callback was fired
+    def first_occurance_of(methodname)
+      callbacks.each do |cb|
+        return cb if cb[:name] == methodname
+      end
+      return nil
+    end
+
+    # return the details for the final time the specified callback was fired
+    def final_occurance_of(methodname)
+      returnme = nil
+      callbacks.each do |cb|
+        returnme = cb if cb[:name] == methodname
+      end
+      return returnme
+    end
+
+    # return the first occurance of a particular series of callbacks
+    def series(*methods)
+      return nil if methods.empty?
+
+      indexes = (0..(callbacks.size-1-methods.size))
+      method_indexes = (0..(methods.size-1))
+      match = nil
+
+      indexes.each do |idx|
+        count = methods.size
+        method_indexes.each do |midx|
+          count -= 1 if callbacks[idx+midx][:name] == methods[midx]
+        end
+        match = idx and break if count == 0
+      end
+
+      if match 
+        return callbacks[match, methods.size]
+      else
+        return nil
+      end
     end
   end
 end

data/lib/pdf/reader/xref.rb

@@ -54,6 +54,7 @@ class PDF::Reader
     # by specifying a PDF::Reader::Reference object that contains the objects ID and revision
     # number
     def object (ref, save_pos = true)
+      return ref unless ref.kind_of?(Reference)
       pos = @buffer.pos if save_pos
       parser = Parser.new(@buffer.seek(offset_for(ref)), self).object(ref.id, ref.gen)
       @buffer.seek(pos) if save_pos

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: pdf-reader
 version: !ruby/object:Gem::Version 
-  version: 0.5.1
+  version: "0.6"
 platform: ruby
 authors: 
 - Peter Jones
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2008-01-01 00:00:00 +11:00
+date: 2008-02-26 00:00:00 +11:00
 default_executable: 
 dependencies: []
 
@@ -37,7 +37,11 @@ files:
 - lib/pdf/reader/error.rb
 - lib/pdf/reader/content.rb
 - lib/pdf/reader/parser.rb
+- lib/pdf/reader/cmap.rb
+- lib/pdf/reader/encoding.rb
 - lib/pdf/reader/register_receiver.rb
+- lib/pdf/reader/font.rb
+- lib/pdf/reader/glyphlist.txt
 - lib/pdf/reader.rb
 - Rakefile
 - README