ruby-msg 1.2.17

data/data/named_map.yaml

@@ -0,0 +1,114 @@
+# this file provides for the mapping of the keys of named properties
+# to symbolic names (as opposed to mapitags.yaml, which is currently
+# in a different format, has a different source, and is only fixed
+# code properties)
+#
+# essentially the symbols are slightly munged versions of the names
+# given to these properties by CDO, or Outlook's object model.
+# it was parsed out of cdo10.htm, and neatened up a bit.
+#
+# interestingly, despite having separate guids, the codes are picked not to
+# clash. further the names themselves have only 3 clashes in all the below.
+{
+[0x8005, PSETID_Address]: file_under,
+[0x8017, PSETID_Address]: last_name_and_first_name,
+[0x8018, PSETID_Address]: company_and_full_name,
+[0x8019, PSETID_Address]: full_name_and_company,
+[0x801a, PSETID_Address]: home_address,
+[0x801b, PSETID_Address]: business_address,
+[0x801c, PSETID_Address]: other_address,
+[0x8022, PSETID_Address]: selected_address,
+[0x802b, PSETID_Address]: web_page,
+[0x802c, PSETID_Address]: yomi_first_name,
+[0x802d, PSETID_Address]: yomi_last_name,
+[0x802e, PSETID_Address]: yomi_company_name,
+[0x8030, PSETID_Address]: last_first_no_space,
+[0x8031, PSETID_Address]: last_first_space_only,
+[0x8032, PSETID_Address]: company_last_first_no_space,
+[0x8033, PSETID_Address]: company_last_first_space_only,
+[0x8034, PSETID_Address]: last_first_no_space_company,
+[0x8035, PSETID_Address]: last_first_space_only_company,
+[0x8036, PSETID_Address]: last_first_and_suffix,
+[0x8045, PSETID_Address]: business_address_street,
+[0x8046, PSETID_Address]: business_address_city,
+[0x8047, PSETID_Address]: business_address_state,
+[0x8048, PSETID_Address]: business_address_postal_code,
+[0x8049, PSETID_Address]: business_address_country,
+[0x804a, PSETID_Address]: business_address_post_office_box,
+[0x804f, PSETID_Address]: user_field1,
+[0x8050, PSETID_Address]: user_field2,
+[0x8051, PSETID_Address]: user_field3,
+[0x8052, PSETID_Address]: user_field4,
+[0x8062, PSETID_Address]: imaddress,
+[0x8082, PSETID_Address]: email_addr_type,
+[0x8083, PSETID_Address]: email_email_address,
+[0x8084, PSETID_Address]: email_original_display_name,
+[0x8085, PSETID_Address]: email_original_entry_id,
+[0x8092, PSETID_Address]: email2_addr_type,
+[0x8093, PSETID_Address]: email2_email_address,
+[0x8094, PSETID_Address]: email2_original_display_name,
+[0x8095, PSETID_Address]: email2_original_entry_id,
+[0x80a2, PSETID_Address]: email3_addr_type,
+[0x80a3, PSETID_Address]: email3_email_address,
+[0x80a4, PSETID_Address]: email3_original_display_name,
+[0x80a5, PSETID_Address]: email3_original_entry_id,
+[0x80d8, PSETID_Address]: internet_free_busy_address,
+[0x8101, PSETID_Task]: status,
+[0x8102, PSETID_Task]: percent_complete,
+[0x8103, PSETID_Task]: team_task,
+[0x8104, PSETID_Task]: start_date,
+[0x8105, PSETID_Task]: due_date,
+[0x8106, PSETID_Task]: duration,
+[0x810f, PSETID_Task]: date_completed,
+[0x8110, PSETID_Task]: actual_work,
+[0x8111, PSETID_Task]: total_work,
+[0x811c, PSETID_Task]: complete,
+[0x811f, PSETID_Task]: owner,
+[0x8126, PSETID_Task]: is_recurring,
+[0x8205, PSETID_Appointment]: busy_status,
+[0x8208, PSETID_Appointment]: location,
+[0x820d, PSETID_Appointment]: start_date,
+[0x820e, PSETID_Appointment]: end_date,
+[0x8213, PSETID_Appointment]: duration,
+[0x8214, PSETID_Appointment]: colors,
+[0x8216, PSETID_Appointment]: recurrence_state,
+[0x8218, PSETID_Appointment]: response_status,
+[0x8222, PSETID_Appointment]: reply_time,
+[0x8223, PSETID_Appointment]: is_recurring,
+[0x822e, PSETID_Appointment]: organizer,
+[0x8231, PSETID_Appointment]: recurrence_type,
+[0x8232, PSETID_Appointment]: recurrence_pattern,
+# also had CdoPR_FLAG_DUE_BY, when applied to messages. i don't currently
+# use message class specific names
+[0x8502, PSETID_Common]: reminder_time,
+[0x8503, PSETID_Common]: reminder_set,
+[0x8516, PSETID_Common]: common_start,
+[0x8517, PSETID_Common]: common_end,
+[0x851c, PSETID_Common]: reminder_override,
+[0x851e, PSETID_Common]: reminder_sound,
+[0x851f, PSETID_Common]: reminder_file,
+# this one only listed as CdoPR_FLAG_TEXT. maybe should be
+# reminder_text
+[0x8530, PSETID_Common]: flag_text,
+[0x8534, PSETID_Common]: mileage,
+[0x8535, PSETID_Common]: billing_information,
+[0x8539, PSETID_Common]: companies,
+[0x853a, PSETID_Common]: contact_names,
+# had CdoPR_FLAG_DUE_BY_NEXT for this one also
+[0x8560, PSETID_Common]: reminder_next_time,
+[0x8700, PSETID_Log]: entry,
+[0x8704, PSETID_Log]: start_date,
+[0x8705, PSETID_Log]: start_time,
+[0x8706, PSETID_Log]: start,
+[0x8707, PSETID_Log]: duration,
+[0x8708, PSETID_Log]: end,
+[0x870e, PSETID_Log]: doc_printed,
+[0x870f, PSETID_Log]: doc_saved,
+[0x8710, PSETID_Log]: doc_routed,
+[0x8711, PSETID_Log]: doc_posted,
+[0x8712, PSETID_Log]: entry_type,
+[0x8b00, PSETID_Note]: color,
+[0x8b02, PSETID_Note]: width,
+[0x8b03, PSETID_Note]: height,
+["Keywords", PS_PUBLIC_STRINGS]: categories
+}

data/data/types.yaml

@@ -0,0 +1,15 @@
+---
+# grep ' PT_' mapitags.yaml | sort -u > types.yaml
+- PT_BINARY
+- PT_BOOLEAN
+- PT_CLSID
+- PT_I8
+- PT_LONG
+- PT_MV_BINARY
+- PT_MV_LONG
+- PT_MV_TSTRING
+- PT_OBJECT
+- PT_SHORT
+- PT_STRING8
+- PT_SYSTIME
+- PT_TSTRING

data/lib/blah.rb

@@ -0,0 +1,106 @@
+#! /usr/bin/ruby
+
+class SimpleGrammar
+    def initialize(grammar)
+        @grammar = grammar
+    end
+    def scan(cursor,buffer)
+        @grammar.scan(cursor,buffer)
+    end
+    def |(other); Code.new { |cursor,buffer|
+        scan(cursor,buffer) || other.scan(cursor,buffer)
+    } end
+    def +(other); Code.new { |cursor,buffer|
+        scan(cursor,buffer) && other.scan(cursor,buffer)
+    } end
+    def filter(buf0,&block); Code.new { |cursor,buffer|
+        buf = buf0.clone
+        scan(cursor,buf) && buffer.concat(block[buf])
+    } end
+    def discard; Code.new { |cursor,buffer|
+        scan(cursor,[])
+    } end
+    class Code < SimpleGrammar
+        def initialize(&block)
+            @block = block
+        end
+        def scan(cursor,buffer)
+            @block[cursor,buffer]
+        end
+    end
+    class Recurse < SimpleGrammar
+        def initialize(&block)
+            @grammar = block[self]
+        end
+        def scan(cursor,buffer)
+            @grammar.scan(cursor,buffer)
+        end
+    end
+    class Element < SimpleGrammar
+        def initialize(pattern)
+            @pattern = pattern
+        end
+        def scan(cursor,buffer)
+            c = cursor.read1after
+            if @pattern===c
+                buffer << c
+                cursor.skip1next
+                true
+            end
+        end
+    end
+    # Make methods for our classes that call new for us
+    constants.each { |klass|
+        eval("
+            def #{klass}(*args,&block)
+                #{klass}.new(*args,&block)
+            end
+            def self.#{klass}(*args,&block)
+                #{klass}.new(*args,&block)
+            end
+        ")
+    }
+    NULL = Code.new { true }
+end
+
+class IO
+    # implement just the methods we need to look like a cursor
+    def read1after;c=getc;ungetc(c);c;end
+    def skip1next;getc&&true;end
+end
+
+class Expression < SimpleGrammar::Recurse
+    def initialize; super() { |expr|
+        digit = Element(?0..?9)
+        int = Recurse { |int| digit+(int|NULL) }
+        number =
+            (int + (
+                Element(?.)+int |
+                NULL
+            )).filter("") { |n| [n.to_f] }
+        primary = Recurse { |primary|
+            number |
+            Element(?-).discard + primary + Code { |_,b| b[-1]=-b[-1] } |
+            Element(?().discard + expr + Element(?)).discard
+        }
+        product = Recurse { |product|
+            primary + (
+                Element(?*).discard + product + Code { |_,b|
+b[-2]*=b[-1];b.pop } |
+                Element(?/).discard + product + Code { |_,b|
+b[-2]/=b[-1];b.pop } |
+                NULL
+            )
+        }
+        sum = Recurse { |sum|
+            product + (
+                Element(?+).discard + sum + Code { |_,b| b[-2]+=b[-1];b.pop } |
+                Element(?-).discard + sum + Code { |_,b| b[-2]-=b[-1];b.pop } |
+                NULL
+            )
+        }
+    } end
+end
+
+Expression.new.scan(STDIN,buf=[]) && p(buf[0])
+

data/lib/mime-new.rb

@@ -0,0 +1,210 @@
+require 'stringio'
+require 'ole/storage'
+
+# the IO equivalent to StringScanner.
+# in fact its mostly scan and scan_until that are useful to me. maybe this can be
+# mixed in to the underlying io class, as providing a lot of stuff, like simple reads
+# and peeks seems redundant extra layer cruft.
+class IOScanner
+	# needs seekable +io+, as we may read more than we should, and don't
+	# have a way (that i know of?) to push back those bytes onto temporary
+	# buffer. (ungetc?)
+	attr_reader :io
+	def initialize io
+		@io = io
+	end
+
+	def pos
+		@io.pos
+	end
+
+	def pos= pos
+		@io.pos = pos
+	end
+
+	def scan rx
+		rx = Regexp.new rx
+		# now we have to see whether we match
+	end
+end
+
+#
+# = Introduction
+#
+# A *basic* mime class for _really_ _basic_ and probably non-standard parsing
+# and construction of MIME messages.
+#
+# Intended for two main purposes in this project:
+# 1. As the container that is used to build up the message for eventual
+#    serialization as an eml.
+# 2. For assistance in parsing the +transport_message_headers+ provided in .msg files,
+#    which are then kept through to the final eml.
+#
+# = TODO
+#
+# * Better streaming support, rather than an all-in-string approach.
+# * Add +OrderedHash+ optionally, to not lose ordering in headers.
+# * A fair bit remains to be done for this class, its fairly immature. But generally I'd like
+#   to see it be more generally useful.
+# * All sorts of correctness issues, encoding particular.
+# * Duplication of work in net/http.rb's +HTTPHeader+? Don't know if the overlap is sufficient.
+#   I don't want to lower case things, just for starters.
+# * Mime was the original place I wrote #to_tree, intended as a quick debug hack.
+#
+class Mime
+	Hash = begin
+		require 'orderedhash'
+		OrderedHash
+	rescue LoadError
+		Hash
+	end
+
+	attr_reader :headers, :body, :parts, :content_type, :preamble, :epilogue
+
+	# Create a Mime object from +io+ as an initial serialization, which must contain headers
+	# and a body (even if empty). Needs work.
+	attr_reader :io
+	def initialize io=nil
+		# use a string backed io if not specified
+		@io = io || StringIO.new('')
+		# read headers. flexible about new line endings here
+		headers = ''
+		while s = @io.gets
+			s.chomp!
+			break if s.empty?
+			headers << s + "\r\n"
+		end
+		# what remains is the body. maybe rangesio should support infinite ranges
+		@body = RangesIO.new @io, [@io.pos...(@io.stat.size rescue @io.size)]
+		#headers, @body = $~[1..-1] if str[/(.*?\r?\n)(?:\r?\n(.*))?\Z/m]
+
+		# don't like the way i'm creatinh on access.
+		@headers = Hash.new { |hash, key| hash[key] = [] }
+		#@body ||= ''
+		headers.to_s.scan(/^\S+:\s*.*(?:\n\t.*)*/).each do |header|
+			@headers[header[/(\S+):/, 1]] << header[/\S+:\s*(.*)/m, 1].gsub(/\s+/m, ' ').strip # this is kind of wrong
+		end
+
+		# don't have to have content type i suppose
+		@content_type, attrs = nil, {}
+		if @headers.include? 'Content-Type'
+			@content_type, attrs = Mime.split_header @headers['Content-Type'][0]
+		end
+
+		if multipart?
+			# special case
+			if body.size == 0
+				@preamble = ''
+				@epilogue = ''
+				@parts = []
+			else
+				# we need to split the message at the boundary
+				# instead of raising, it should just probably become a single part message.
+				boundary = attrs['boundary'] or raise "no boundary for multipart message"
+
+				# splitting the body:
+				# talk about crap. all that work to use io, and then i do this :)
+				# each multipart should become a sub-io, which gets passed to Mime.new
+				parts = body.read.split(/--#{Regexp.quote boundary}/m)
+				unless parts[-1] =~ /^--/; warn "bad multipart boundary (missing trailing --)"
+				else parts[-1][0..1] = ''
+				end
+				parts.each_with_index do |part, i|
+					part =~ /^(\r?\n)?(.*?)(\r?\n)?\Z/m
+					part.replace $2
+					warn "bad multipart boundary" if (1...parts.length-1) === i and !($1 && $3)
+				end
+				@preamble = parts.shift
+				@epilogue = parts.pop
+				@parts = parts.map { |part| Mime.new part }
+			end
+		end
+	end
+
+	def multipart?
+		@content_type && @content_type =~ /^multipart/ ? true : false
+	end
+
+	def inspect
+		# add some extra here.
+		"#<Mime content_type=#{@content_type.inspect}>"
+	end
+
+	def to_tree
+		if multipart?
+			str = "- #{inspect}\n"
+			parts.each_with_index do |part, i|
+				last = i == parts.length - 1
+				part.to_tree.split(/\n/).each_with_index do |line, j|
+					str << "  #{last ? (j == 0 ? "\\" : ' ') : '|'}" + line + "\n"
+				end
+			end
+			str
+		else
+			"- #{inspect}\n"
+		end
+	end
+
+	def to_s opts={}
+		opts = {:boundary_counter => 0}.merge opts
+		if multipart?
+			boundary = Mime.make_boundary opts[:boundary_counter] += 1, self
+			@body = [preamble, parts.map { |part| "\r\n" + part.to_s(opts) + "\r\n" }, "--\r\n" + epilogue].
+				flatten.join("\r\n--" + boundary)
+			content_type, attrs = Mime.split_header @headers['Content-Type'][0]
+			attrs['boundary'] = boundary
+			@headers['Content-Type'] = [([content_type] + attrs.map { |key, val| %{#{key}="#{val}"} }).join('; ')]
+		end
+
+		str = ''
+		@headers.each do |key, vals|
+			vals.each { |val| str << "#{key}: #{val}\r\n" }
+		end
+		str << "\r\n" + @body
+	end
+
+	def self.split_header header
+		# FIXME: haven't read standard. not sure what its supposed to do with " in the name, or if other
+		# escapes are allowed. can't test on windows as " isn't allowed anyway. can be fixed with more
+		# accurate parser later.
+		# maybe move to some sort of Header class. but not all headers should be of it i suppose.
+		# at least add a join_header then, taking name and {}. for use in Mime#to_s (for boundary
+		# rewrite), and Attachment#to_mime, among others...
+		attrs = {}
+		header.scan(/;\s*([^\s=]+)\s*=\s*("[^"]*"|[^\s;]*)\s*/m).each do |key, value|
+			if attrs[key]; warn "ignoring duplicate header attribute #{key.inspect}"
+			else attrs[key] = value[/^"/] ? value[1..-2] : value
+			end
+		end
+
+		[header[/^[^;]+/].strip, attrs]
+	end
+
+	# +i+ is some value that should be unique for all multipart boundaries for a given message
+	def self.make_boundary i, extra_obj = Mime
+		"----_=_NextPart_#{'%03d' % i}_#{'%08x' % extra_obj.object_id}.#{'%08x' % Time.now}"
+	end
+end
+
+=begin
+things to consider for header work.
+encoded words:
+Subject: =?iso-8859-1?q?p=F6stal?=
+
+and other mime funkyness:
+Content-Disposition: attachment;
+	filename*0*=UTF-8''09%20%D7%90%D7%A5;
+	filename*1*=%20%D7%A1%D7%91-;
+	filename*2*=%D7%A7%95%A5.wma
+Content-Transfer-Encoding: base64
+
+and another, doing a test with an embedded newline in an attachment name, I
+get this output from evolution. I get the feeling that this is probably a bug
+with their implementation though, they weren't expecting new lines in filenames.
+Content-Disposition: attachment; filename="asdf'b\"c
+d   efgh=i: ;\\j"
+d   efgh=i: ;\\j"; charset=us-ascii
+Content-Type: text/plain; name="asdf'b\"c"; charset=us-ascii
+
+=end
+

data/lib/mime.rb

@@ -0,0 +1,165 @@
+#
+# = Introduction
+#
+# A *basic* mime class for _really_ _basic_ and probably non-standard parsing
+# and construction of MIME messages.
+#
+# Intended for two main purposes in this project:
+# 1. As the container that is used to build up the message for eventual
+#    serialization as an eml.
+# 2. For assistance in parsing the +transport_message_headers+ provided in .msg files,
+#    which are then kept through to the final eml.
+#
+# = TODO
+#
+# * Better streaming support, rather than an all-in-string approach.
+# * Add +OrderedHash+ optionally, to not lose ordering in headers.
+# * A fair bit remains to be done for this class, its fairly immature. But generally I'd like
+#   to see it be more generally useful.
+# * All sorts of correctness issues, encoding particular.
+# * Duplication of work in net/http.rb's +HTTPHeader+? Don't know if the overlap is sufficient.
+#   I don't want to lower case things, just for starters.
+# * Mime was the original place I wrote #to_tree, intended as a quick debug hack.
+#
+class Mime
+	Hash = begin
+		require 'orderedhash'
+		OrderedHash
+	rescue LoadError
+		Hash
+	end
+
+	attr_reader :headers, :body, :parts, :content_type, :preamble, :epilogue
+
+	# Create a Mime object using +str+ as an initial serialization, which must contain headers
+	# and a body (even if empty). Needs work.
+	def initialize str, ignore_body=false
+		headers, @body = $~[1..-1] if str[/(.*?\r?\n)(?:\r?\n(.*))?\Z/m]
+
+		@headers = Hash.new { |hash, key| hash[key] = [] }
+		@body ||= ''
+		headers.to_s.scan(/^\S+:\s*.*(?:\n\t.*)*/).each do |header|
+			@headers[header[/(\S+):/, 1]] << header[/\S+:\s*(.*)/m, 1].gsub(/\s+/m, ' ').strip # this is kind of wrong
+		end
+
+		# don't have to have content type i suppose
+		@content_type, attrs = nil, {}
+		if content_type = @headers['Content-Type'][0]
+			@content_type, attrs = Mime.split_header content_type
+		end
+
+		return if ignore_body
+
+		if multipart?
+			if body.empty?
+				@preamble = ''
+				@epilogue = ''
+				@parts = []
+			else
+				# we need to split the message at the boundary
+				boundary = attrs['boundary'] or raise "no boundary for multipart message"
+
+				# splitting the body:
+				parts = body.split(/--#{Regexp.quote boundary}/m)
+				unless parts[-1] =~ /^--/; warn "bad multipart boundary (missing trailing --)"
+				else parts[-1][0..1] = ''
+				end
+				parts.each_with_index do |part, i|
+					part =~ /^(\r?\n)?(.*?)(\r?\n)?\Z/m
+					part.replace $2
+					warn "bad multipart boundary" if (1...parts.length-1) === i and !($1 && $3)
+				end
+				@preamble = parts.shift
+				@epilogue = parts.pop
+				@parts = parts.map { |part| Mime.new part }
+			end
+		end
+	end
+
+	def multipart?
+		@content_type && @content_type =~ /^multipart/ ? true : false
+	end
+
+	def inspect
+		# add some extra here.
+		"#<Mime content_type=#{@content_type.inspect}>"
+	end
+
+	def to_tree
+		if multipart?
+			str = "- #{inspect}\n"
+			parts.each_with_index do |part, i|
+				last = i == parts.length - 1
+				part.to_tree.split(/\n/).each_with_index do |line, j|
+					str << "  #{last ? (j == 0 ? "\\" : ' ') : '|'}" + line + "\n"
+				end
+			end
+			str
+		else
+			"- #{inspect}\n"
+		end
+	end
+
+	def to_s opts={}
+		opts = {:boundary_counter => 0}.merge opts
+		if multipart?
+			boundary = Mime.make_boundary opts[:boundary_counter] += 1, self
+			@body = [preamble, parts.map { |part| "\r\n" + part.to_s(opts) + "\r\n" }, "--\r\n" + epilogue].
+				flatten.join("\r\n--" + boundary)
+			content_type, attrs = Mime.split_header @headers['Content-Type'][0]
+			attrs['boundary'] = boundary
+			@headers['Content-Type'] = [([content_type] + attrs.map { |key, val| %{#{key}="#{val}"} }).join('; ')]
+		end
+
+		str = ''
+		@headers.each do |key, vals|
+			vals.each { |val| str << "#{key}: #{val}\r\n" }
+		end
+		str << "\r\n" + @body
+	end
+
+	def self.split_header header
+		# FIXME: haven't read standard. not sure what its supposed to do with " in the name, or if other
+		# escapes are allowed. can't test on windows as " isn't allowed anyway. can be fixed with more
+		# accurate parser later.
+		# maybe move to some sort of Header class. but not all headers should be of it i suppose.
+		# at least add a join_header then, taking name and {}. for use in Mime#to_s (for boundary
+		# rewrite), and Attachment#to_mime, among others...
+		attrs = {}
+		header.scan(/;\s*([^\s=]+)\s*=\s*("[^"]*"|[^\s;]*)\s*/m).each do |key, value|
+			if attrs[key]; warn "ignoring duplicate header attribute #{key.inspect}"
+			else attrs[key] = value[/^"/] ? value[1..-2] : value
+			end
+		end
+
+		[header[/^[^;]+/].strip, attrs]
+	end
+
+	# +i+ is some value that should be unique for all multipart boundaries for a given message
+	def self.make_boundary i, extra_obj = Mime
+		"----_=_NextPart_#{'%03d' % i}_#{'%08x' % extra_obj.object_id}.#{'%08x' % Time.now}"
+	end
+end
+
+=begin
+things to consider for header work.
+encoded words:
+Subject: =?iso-8859-1?q?p=F6stal?=
+
+and other mime funkyness:
+Content-Disposition: attachment;
+	filename*0*=UTF-8''09%20%D7%90%D7%A5;
+	filename*1*=%20%D7%A1%D7%91-;
+	filename*2*=%D7%A7%95%A5.wma
+Content-Transfer-Encoding: base64
+
+and another, doing a test with an embedded newline in an attachment name, I
+get this output from evolution. I get the feeling that this is probably a bug
+with their implementation though, they weren't expecting new lines in filenames.
+Content-Disposition: attachment; filename="asdf'b\"c
+d   efgh=i: ;\\j"
+d   efgh=i: ;\\j"; charset=us-ascii
+Content-Type: text/plain; name="asdf'b\"c"; charset=us-ascii
+
+=end
+