ruby-msg 1.2.17.3 → 1.3.1

data/FIXES

@@ -32,3 +32,25 @@ recent fixes based on importing results into evolution
 
 6. still unsure about how to do my "\r" handling.
 
+7. need to join addresses with , instead of ; i think. evolution only shows the
+   first one otherwise it appears, but all when they are , separated.
+
+8. need to solve ole storage issues with the very large file using extra bat
+   stuff.
+
+9. retest a bit on evolution and thunderbird, and release. tested on a corups
+   of >1000 msg files, so should be starting to get pretty good quality.
+
+10. longer term, things fall into a few basic categories:
+
+- non mail conversions (look further into vcard, ical et al support for other
+  types of msg)
+- further tests and robustness for what i handle now. ie, look into corner
+  cases covered so far, and work on the mime code. fix random charset encoding
+  issues, in the various weird mime ways, do header wrapping etc etc.
+  check fidelity of conversions, and capture some more properties as headers,
+  such as importance which i don't do yet.
+- fix that named property bug. tidy up warnings, exceptions.
+- extend conversion to make better html.
+  this is longer term. as i don't use the rtf, i need to make my html better.
+  emulating some rtf things. harder, not important atm.

data/Rakefile

@@ -6,8 +6,9 @@ require 'rake/gempackagetask'
 require 'rbconfig'
 require 'fileutils'
 
-$: << './lib'
-require 'msg.rb'
+$:.unshift 'lib'
+
+require 'msg'
 
 PKG_NAME = 'ruby-msg'
 PKG_VERSION = Msg::VERSION
@@ -23,18 +24,8 @@ end
 # RDocTask wasn't working for me
 desc 'Build the rdoc HTML Files'
 task :rdoc do
-	system "rdoc -S -N -m Msg -w 2 -t '#{PKG_NAME} documentation' lib"
-end
-
-=begin
-Rake::PackageTask.new(PKG_NAME, PKG_VERSION) do |p|
-	p.need_tar_gz = true
-	p.package_dir = 'build'
-	p.package_files.include("Rakefile", "README")
-	p.package_files.include("contrib/*.c")
-	p.package_files.include("test/test_*.rb", "test/*.doc", "lib/*.rb", "lib/ole/storage.rb")
+	system "rdoc -S -N --main Msg --tab-width 2 --title '#{PKG_NAME} documentation' lib"
 end
-=end
 
 spec = Gem::Specification.new do |s|
 	s.name = PKG_NAME
@@ -46,20 +37,26 @@ spec = Gem::Specification.new do |s|
 	s.homepage = %q{http://code.google.com/p/ruby-msg}
 	#s.rubyforge_project = %q{ruby-msg}
 
-	s.executables = ['msgtool', 'oletool']
+	s.executables = ['msgtool']
 	s.files  = Dir.glob('data/*.yaml') + ['Rakefile', 'README', 'FIXES']
 	s.files += Dir.glob("lib/**/*.rb")
-	s.files += Dir.glob("test/test_*.rb") + Dir.glob("test/*.doc")
+	s.files += Dir.glob("test/test_*.rb")
 	s.files += Dir.glob("bin/*")
 	
 	s.has_rdoc = true
+	s.rdoc_options += ['--main', 'Msg',
+					   '--title', "#{PKG_NAME} documentation",
+					   '--tab-width', '2']
+
 
 	s.autorequire = 'msg'
+
+	s.add_dependency 'ruby-ole', '>=1.2.1'
 end
 
 Rake::GemPackageTask.new(spec) do |p|
 	p.gem_spec = spec
-#	p.need_tar = true
+	p.need_tar = true
 	p.need_zip = false
 	p.package_dir = 'build'
 end

data/bin/msgtool

@@ -32,7 +32,7 @@ def msgtool
 	end
 	# just shut up and convert a message to eml
 	Msg::Log.level = Ole::Log.level = opts[:verbose] ? Logger::WARN : Logger::FATAL
-	# fix for windows
+	# for windows. see issue #2
 	STDOUT.binmode
 	case opts[:action]
 	when :convert

data/lib/msg.rb

@@ -5,7 +5,7 @@ $: << File.dirname(__FILE__)
 require 'yaml'
 require 'base64'
 
-require 'support'
+require 'rubygems'
 require 'ole/storage'
 require 'msg/properties'
 require 'msg/rtf'
@@ -20,7 +20,7 @@ require 'mime'
 #
 
 class Msg
-	VERSION = '1.2.17.3'
+	VERSION = '1.3.1'
 	# we look here for the yaml files in data/, and the exe files for support
 	# decoding at the moment.
 	SUPPORT_DIR = File.dirname(__FILE__) + '/..'
@@ -144,9 +144,9 @@ class Msg
 			# of the ole name, or just leave it if we can't
 			recips = recips_by_type[type]
 			recips = (recips.sort_by { |r| r.obj.name[/\d{8}$/].hex } rescue recips)
-			# are you supposed to use ; or , to separate?
+			# switched to using , for separation, not ;. see issue #4
 			# recips.empty? is strange. i wouldn't have thought it possible, but it was right?
-			headers[type.to_s.sub(/^(.)/) { $1.upcase }] = [recips.join('; ')] unless recips.empty?
+			headers[type.to_s.sub(/^(.)/) { $1.upcase }] = [recips.join(', ')] unless recips.empty?
 		end
 		headers['Subject'] = [props.subject] if props.subject
 
@@ -171,11 +171,28 @@ class Msg
 			headers['Date'] = [Time.iso8601(time.to_s).rfc2822] if time
 		end
 
-		if !headers.has_key?('Message-ID') and props.internet_message_id
-			headers['Message-ID'] = [props.internet_message_id]
-		end
-		if !headers.has_key?('In-Reply-To') and props.in_reply_to_id
-			headers['In-Reply-To'] = [props.in_reply_to_id]
+		# some very simplistic mapping between internet message headers and the
+		# mapi properties
+		# any of these could be causing duplicates due to case issues. the hack in #to_mime
+		# just stops re-duplication at that point. need to move some smarts into the mime
+		# code to handle it.
+		mapi_header_map = [
+			[:internet_message_id, 'Message-ID'],
+			[:in_reply_to_id, 'In-Reply-To'],
+			# don't set these values if they're equal to the defaults anyway
+			[:importance, 'Importance', proc { |val| val.to_s == '1' ? nil : val }],
+			[:priority, 'Priority', proc { |val| val.to_s == '1' ? nil : val }],
+			[:sensitivity, 'Sensitivity', proc { |val| val.to_s == '0' ? nil : val }],
+			# yeah?
+			[:conversation_topic, 'Thread-Topic'],
+			# not sure of the distinction here
+			# :originator_delivery_report_requested ??
+			[:read_receipt_requested, 'Disposition-Notification-To', proc { |val| from }]
+		]
+		mapi_header_map.each do |mapi, mime, *f|
+			next unless q = val = props.send(mapi) or headers.has_key?(mime)
+			next if f[0] and !(val = f[0].call(val))
+			headers[mime] = [val.to_s]
 		end
 	end
 

data/lib/msg/properties.rb

@@ -48,24 +48,28 @@ class Msg
 	# There also needs to be a way to look up properties more specifically:
 	# 
 	#   properties[0x0037] # => gets the subject
-	#   properties[PS_MAPI, 0x0037] # => still gets the subject
-	#   properties[PS_PUBLIC_STRINGS, 'Keywords'] # => gets the above categories
+	#   properties[0x0037, PS_MAPI] # => still gets the subject
+	#   properties['Keywords', PS_PUBLIC_STRINGS] # => gets outlook's categories array
 	# 
-	# The abbreviate versions work by "resolving" the symbols to full keys:
+	# The abbreviated versions work by "resolving" the symbols to full keys:
 	#
-	#   properties.resolve :keywords # => [PS_OUTLOOK, 'Keywords']
-	#   properties.resolve :subject  # => [PS_MAPI, 0x0037]
+	#		# the guid here is just PS_PUBLIC_STRINGS
+	#   properties.resolve :keywords # => #<Key {00020329-0000-0000-c000-000000000046}/"Keywords">
+	#   # the result here is actually also a key
+	#   k = properties.resolve :subject  # => 0x0037
+	#   # it has a guid
+	#   k.guid == Msg::Properties::PS_MAPI # => true
 	#
 	# = Parsing
 	#
 	# There are three objects that need to be parsed to load a +Msg+ property store:
 	# 
-	#	1. The +nameid+ directory (<tt>Properties.parse_nameid</tt>)
+	# 1. The +nameid+ directory (<tt>Properties.parse_nameid</tt>)
 	# 2. The many +substg+ objects, whose names should match <tt>Properties::SUBSTG_RX</tt>
 	#    (<tt>Properties#parse_substg</tt>)
 	# 3. The +properties+ file (<tt>Properties#parse_properties</tt>)
 	#
-	# Understanding of the formats is by no means perfect
+	# Understanding of the formats is by no means perfect.
 	#
 	# = TODO
 	#
@@ -79,7 +83,7 @@ class Msg
 	#   current greedy-loading approach. still want strings to work nicely:
 	#     props.subject
 	#   but don't want to be loading up large binary blobs, typically attachments, eg
-	#     props.attach_data.
+	#     props.attach_data
 	#   probably the easiest solution is that the binary "encoding", be to return an io
 	#   object instead. and you must read it if you want it as a string
 	#   maybe i can avoid the greedy model anyway? rather than parsing the properties completely,
@@ -133,9 +137,13 @@ class Msg
 		attr_reader :unused
 		attr_reader :nameid
 
+		# +nameid+ is to provide a way to inherit from parent (needed for property sets for
+		# attachments and recipients, which inherit from the msg itself. what about nested
+		# msg??)
 		def initialize
 			@raw = {}
 			@unused = []
+			@nameid = nil
 			# FIXME
 			@body_rtf = @body_html = @body = false
 		end
@@ -144,7 +152,7 @@ class Msg
 		# The parsing methods
 		#++
 
-		def self.load obj
+		def self.load obj, ignore=nil
 			prop = Properties.new
 			prop.load obj
 			prop
@@ -154,9 +162,16 @@ class Msg
 		def load obj
 			# we need to do the nameid first, as it provides the map for later user defined properties
 			children = obj.children.dup
-			@nameid = if nameid_obj = children.find { |child| child.name == '__nameid_version1.0' }
+			if nameid_obj = children.find { |child| child.name == '__nameid_version1.0' }
 				children.delete nameid_obj
-				Properties.parse_nameid nameid_obj
+				@nameid = Properties.parse_nameid nameid_obj
+				# hack to make it available to all msg files from the same ole storage object
+				class << obj.ole
+					attr_accessor :msg_nameid
+				end
+				obj.ole.msg_nameid = @nameid
+			elsif obj.ole
+				@nameid = obj.ole.msg_nameid rescue nil
 			end
 			# now parse the actual properties. i think dirs that match the substg should be decoded
 			# as properties to. 0x000d is just another encoding, the dir encoding. it should match
@@ -310,6 +325,8 @@ class Msg
 				elsif real_key = @nameid[key]
 					key = real_key
 				else
+					# i think i hit these when i have a named property, in the PS_MAPI
+					# guid
 					Log.warn "property in named range not in nameid #{key.inspect}"
 					key = Key.new key
 				end

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.9.0
 specification_version: 1
 name: ruby-msg
 version: !ruby/object:Gem::Version 
-  version: 1.2.17.3
-date: 2007-07-12 00:00:00 +10:00
+  version: 1.3.1
+date: 2007-08-21 00:00:00 +10:00
 summary: Ruby Msg library.
 require_paths: 
 - lib
@@ -36,36 +36,37 @@ files:
 - README
 - FIXES
 - bin/msgtool
-- bin/oletool
 - lib/orderedhash.rb
 - lib/rtf.rb
-- lib/support.rb
 - lib/mime.rb
 - lib/msg.rb
-- lib/ole/types.rb
-- lib/ole/file_system.rb
-- lib/ole/storage.rb
-- lib/ole/io_helpers.rb
-- lib/ole/base.rb
 - lib/msg/rtf.rb
 - lib/msg/properties.rb
 - test/test_mime.rb
-- test/test_storage.rb
-- test/test_word_6.doc
-- test/test_word_95.doc
-- test/test_word_97.doc
 test_files: []
 
-rdoc_options: []
-
+rdoc_options: 
+- --main
+- Msg
+- --title
+- ruby-msg documentation
+- --tab-width
+- "2"
 extra_rdoc_files: []
 
 executables: 
 - msgtool
-- oletool
 extensions: []
 
 requirements: []
 
-dependencies: []
-
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: ruby-ole
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Version::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 1.2.1
+    version: 

data/bin/oletool

@@ -1,35 +0,0 @@
-#! /usr/bin/ruby
-
-require 'optparse'
-require 'rubygems'
-require 'ole/storage'
-
-def oletool
-	opts = {:verbose => false, :action => :tree}
-	op = OptionParser.new do |op|
-		op.banner = "Usage: oletool [options] [files]"
-		op.separator ''
-		op.on('-t', '--tree', 'Dump ole trees for files (default)') { opts[:action] = :tree }
-		op.on('-r', '--repack', 'Repack the ole files in canonical form') { opts[:action] = :repack }
-		op.separator ''
-		op.on('-v', '--[no-]verbose', 'Run verbosely') { |v| opts[:verbose] = v }
-		op.on_tail('-h', '--help', 'Show this message') { puts op; exit }
-	end
-	files = op.parse ARGV
-	if files.empty?
-		puts 'Must specify 1 or more msg files.'
-		puts op
-		exit 1
-	end
-	Ole::Log.level = opts[:verbose] ? Logger::WARN : Logger::FATAL
-	files.each do |file|
-		case opts[:action]
-		when :tree
-			Ole::Storage.open(file) { |ole| puts ole.root.to_tree }
-		when :repack
-			Ole::Storage.open(file, &:repack)
-		end
-	end
-end
-
-oletool

data/lib/ole/base.rb

@@ -1,5 +0,0 @@
-
-module Ole # :nodoc:
-	Log = Logger.new_with_callstack
-end
-

data/lib/ole/file_system.rb

@@ -1,181 +0,0 @@
-#
-# = Introduction
-#
-# This file intends to provide file system-like api support, a la <tt>zip/zipfilesystem</tt>.
-#
-# Ideally, this will be the recommended interface, allowing Ole::Storage, Dir, and
-# Zip::ZipFile to be used exchangablyk. It should be possible to write recursive copy using
-# the plain api, such that you can copy dirs/files agnostically between any of ole docs, dirs,
-# and zip files.
-#
-# = Usage
-#
-# Currently you can do something like the following:
-#
-#   Ole::Storage.open 'test.doc' do |ole|
-#     ole.dir.entries '/'         # => [".", "..", "\001Ole", "1Table", "\001CompObj", ...]
-#     ole.file.read "\001CompObj" # => "\001\000\376\377\003\n\000\000\377\377..."
-#   end
-#
-# = Notes
-#
-# *** This file is very incomplete
-# 
-# i think its okay to have an api like this on top, but there are certain things that ole
-# does that aren't captured.
-# <tt>Ole::Storage</tt> can have multiple files with the same name, for example, or with
-# / in the name, and other things that are probably invalid anyway.
-# i think this should remain an addon, built on top of my core api.
-# but still the ideas can be reflected in the core, ie, changing the read/write semantics.
-#
-# once the core changes are complete, this will be a pretty straight forward file to complete.
-#
-
-require 'ole/base'
-
-module Ole # :nodoc:
-	class Storage
-		def file
-			@file ||= FileParent.new self
-		end
-
-		def dir
-			@dir ||= DirParent.new self
-		end
-
-		def dirent_from_path path_str
-			path = path_str.sub(/^\/*/, '').sub(/\/*$/, '')
-			dirent = @root
-			return dirent if path.empty?
-			path = path.split /\/+/
-			until path.empty?
-				raise "invalid path #{path_str.inspect}" if dirent.file?
-				if tmp = dirent[path.shift]
-					dirent = tmp
-				else
-					# allow write etc later.
-					raise "invalid path #{path_str.inspect}"
-				end
-			end
-			dirent
-		end
-
-		class FileParent
-			def initialize ole
-				@ole = ole
-			end
-
-			def open path_str, mode='r', &block
-				dirent = @ole.dirent_from_path path_str
-				# like Errno::EISDIR
-				raise "#{path_str.inspect} is a directory" unless dirent.file?
-				dirent.open(&block)
-			end
-
-			alias new :open
-
-			def read path
-				open(path) { |f| f.read }
-			end
-
-			# crappy copy from Dir.
-			def unlink path
-				dirent = @ole.dirent_from_path path
-				# EPERM
-				raise "operation not permitted #{path.inspect}" unless dirent.file?
-				# i think we should free all of our blocks. i think the best way to do that would be
-				# like:
-				# open(path) { |f| f.truncate 0 }. which should free all our blocks from the
-				# allocation table. then if we remove ourself from our parent, we won't be part of
-				# the bat at save time.
-				# i think if you run repack, all free blocks should get zeroed.
-				open(path) { |f| f.truncate 0 }
-				parent = @ole.dirent_from_path(('/' + path).sub(/\/[^\/]+$/, ''))
-				parent.children.delete dirent
-				1 # hmmm. as per ::File ?
-			end
-		end
-
-		class DirParent
-			def initialize ole
-				@ole = ole
-			end
-
-			def open path_str
-				dirent = @ole.dirent_from_path path_str
-				# like Errno::ENOTDIR
-				raise "#{path_str.inspect} is not a directory" unless dirent.dir?
-				dir = Dir.new dirent, path_str
-				if block_given?
-					yield dir
-				else
-					dir
-				end
-			end
-
-			# certain Dir class methods proxy in this fashion:
-			def entries path
-				open(path) { |dir| dir.entries }
-			end
-
-			# there are some other important ones, like:
-			# chroot (!), mkdir, chdir, rmdir, glob etc etc. for now, i think
-			# mkdir, and rmdir are the main ones we'd need to support
-			def rmdir path
-				dirent = @ole.dirent_from_path path
-				# repeating myself
-				raise "#{path.inspect} is not a directory" unless dirent.dir?
-				# ENOTEMPTY:
-				raise "directory not empty #{path.inspect}" unless dirent.children.empty?
-				# now delete it, how to do that? the canonical representation that is
-				# maintained is the root tree, and the children array. we must remove it
-				# from the children array.
-				# we need the parent then. this sucks but anyway:
-				parent = @ole.dirent_from_path path.sub(/\/[^\/]+$/, '') || '/'
-				# note that the way this currently works, on save and repack time this will get
-				# reflected. to work properly, ie to make a difference now it would have to re-write
-				# the dirent. i think that Ole::Storage#close will handle that. and maybe include a
-				# #repack.
-				parent.children.delete dirent
-				0 # hmmm. as per ::Dir ?
-			end
-
-			class Dir
-				include Enumerable
-				attr_reader :dirent, :path, :entries, :pos
-
-				def initialize dirent, path
-					@dirent, @path = dirent, path
-					@pos = 0
-					# FIXME: hack, and probably not really desired
-					@entries = %w[. ..] + @dirent.children.map(&:name)
-				end
-
-				def each(&block)
-					@entries.each(&block)
-				end
-
-				def close
-				end
-
-				def read
-					@entries[@pos]
-				ensure
-					@pos += 1 if @pos < @entries.length
-				end
-
-				def pos= pos
-					@pos = [[0, pos].max, @entries.length].min
-				end
-
-				def rewind
-					@pos = 0
-				end
-
-				alias tell :pos
-				alias seek :pos=
-			end
-		end
-	end
-end
-