ruby-msg 1.5.0 → 1.5.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (7) hide show
  1. data/lib/mapi.rb +1 -1
  2. data/lib/mapi/msg.rb +0 -1
  3. data/lib/mapi/property_set.rb +29 -10
  4. data/lib/mapi/rtf.rb +103 -1
  5. data/test/test_msg.rb +4 -6
  6. metadata +87 -79
  7. data/lib/rtf.rb +0 -109
data/lib/mapi.rb CHANGED
@@ -2,7 +2,7 @@ require 'mapi/types'
2
2
  require 'mapi/property_set'
3
3
 
4
4
  module Mapi
5
- VERSION = '1.5.0'
5
+ VERSION = '1.5.1'
6
6
 
7
7
  #
8
8
  # Mapi::Item is the base class used for all mapi objects, and is purely a
data/lib/mapi/msg.rb CHANGED
@@ -1,4 +1,3 @@
1
- require 'rubygems'
2
1
  require 'ole/storage'
3
2
  require 'mapi'
4
3
  require 'mapi/rtf'
data/lib/mapi/property_set.rb CHANGED
@@ -1,7 +1,6 @@
1
1
  require 'yaml'
2
2
  require 'mapi/types'
3
3
  require 'mapi/rtf'
4
- require 'rtf'
5
4
 
6
5
  module Mapi
7
6
  #
@@ -241,7 +240,7 @@ module Mapi
241
240
  # last resort
242
241
  if !@body or @body.strip.empty?
243
242
  Log.warn 'creating text body from rtf'
244
- @body = (::RTF::Converter.rtf2text body_rtf rescue nil)
243
+ @body = (RTF::Converter.rtf2text body_rtf rescue nil)
245
244
  end
246
245
  @body
247
246
  end
@@ -249,18 +248,38 @@ module Mapi
249
248
  # for providing rtf decompression
250
249
  def body_rtf
251
250
  return @body_rtf if defined?(@body_rtf)
252
- @body_rtf = (RTF.rtfdecompr rtf_compressed.read)# rescue nil)
251
+ @body_rtf = nil
252
+ if self[:rtf_compressed]
253
+ begin
254
+ @body_rtf = RTF.rtfdecompr self[:rtf_compressed].read
255
+ rescue
256
+ Log.warn 'unable to decompress rtf'
257
+ end
258
+ end
259
+ @body_rtf
253
260
  end
254
261
 
255
- # for providing rtf to html conversion
262
+ # for providing rtf to html extraction or conversion
256
263
  def body_html
257
264
  return @body_html if defined?(@body_html)
258
- @body_html = (self[:body_html].read rescue nil)
259
- @body_html = (RTF.rtf2html body_rtf rescue nil) if !@body_html or @body_html.strip.empty?
260
- # last resort
261
- if !@body_html or @body_html.strip.empty?
262
- Log.warn 'creating html body from rtf'
263
- @body_html = (::RTF::Converter.rtf2text body_rtf, :html rescue nil)
265
+ @body_html = self[:body_html]
266
+ # sometimes body_html is a stream, and sometimes a string
267
+ @body_html = @body_html.read if @body_html.respond_to?(:read)
268
+ @body_html = nil if @body_html.to_s.strip.empty?
269
+ if body_rtf and !@body_html
270
+ begin
271
+ @body_html = RTF.rtf2html body_rtf
272
+ rescue
273
+ Log.warn 'unable to extract html from rtf'
274
+ end
275
+ if !@body_html
276
+ Log.warn 'creating html body from rtf'
277
+ begin
278
+ @body_html = RTF::Converter.rtf2text body_rtf, :html
279
+ rescue
280
+ Log.warn 'unable to convert rtf to html'
281
+ end
282
+ end
264
283
  end
265
284
  @body_html
266
285
  end
data/lib/mapi/rtf.rb CHANGED
@@ -1,6 +1,5 @@
1
1
  require 'stringio'
2
2
  require 'strscan'
3
- require 'rtf'
4
3
 
5
4
  class StringIO # :nodoc:
6
5
  begin
@@ -20,6 +19,109 @@ module Mapi
20
19
  # Both were ported from their original C versions for simplicity's sake.
21
20
  #
22
21
  module RTF
22
+ class Tokenizer
23
+ def self.process io
24
+ while true do
25
+ case c = io.getc
26
+ when ?{; yield :open_group
27
+ when ?}; yield :close_group
28
+ when ?\\
29
+ case c = io.getc
30
+ when ?{, ?}, ?\\; yield :text, c.chr
31
+ when ?'; yield :text, [io.read(2)].pack('H*')
32
+ when ?a..?z, ?A..?Z
33
+ # read control word
34
+ str = c.chr
35
+ str << c while c = io.read(1) and c =~ /[a-zA-Z]/
36
+ neg = 1
37
+ neg = -1 and c = io.read(1) if c == '-'
38
+ num = if c =~ /[0-9]/
39
+ num = c
40
+ num << c while c = io.read(1) and c =~ /[0-9]/
41
+ num.to_i * neg
42
+ end
43
+ raise "invalid rtf stream" if neg == -1 and !num # ???? \blahblah- some text
44
+ io.seek(-1, IO::SEEK_CUR) if c != ' '
45
+ yield :control_word, str, num
46
+ when nil
47
+ raise "invalid rtf stream" # \EOF
48
+ else
49
+ # other kind of control symbol
50
+ yield :control_symbol, c.chr
51
+ end
52
+ when nil
53
+ return
54
+ when ?\r, ?\n
55
+ # ignore
56
+ else yield :text, c.chr
57
+ end
58
+ end
59
+ end
60
+ end
61
+
62
+ class Converter
63
+ # this is pretty crap, its just to ensure there is always something readable if
64
+ # there is an rtf only body, with no html encapsulation.
65
+ def self.rtf2text str, format=:text
66
+ group = 0
67
+ text = ''
68
+ text << "<html>\n<body>" if format == :html
69
+ group_type = []
70
+ group_tags = []
71
+ RTF::Tokenizer.process(StringIO.new(str)) do |a, b, c|
72
+ add_text = ''
73
+ case a
74
+ when :open_group; group += 1; group_type[group] = nil; group_tags[group] = []
75
+ when :close_group; group_tags[group].reverse.each { |t| text << "</#{t}>" }; group -= 1;
76
+ when :control_word; # ignore
77
+ group_type[group] ||= b
78
+ # maybe change this to use utf8 where possible
79
+ add_text = if b == 'par' || b == 'line' || b == 'page'; "\n"
80
+ elsif b == 'tab' || b == 'cell'; "\t"
81
+ elsif b == 'endash' || b == 'emdash'; "-"
82
+ elsif b == 'emspace' || b == 'enspace' || b == 'qmspace'; " "
83
+ elsif b == 'ldblquote'; '"'
84
+ else ''
85
+ end
86
+ if b == 'b' || b == 'i' and format == :html
87
+ close = c == 0 ? '/' : ''
88
+ text << "<#{close}#{b}>"
89
+ if c == 0
90
+ group_tags[group].delete b
91
+ else
92
+ group_tags[group] << b
93
+ end
94
+ end
95
+ # lot of other ones belong in here.\
96
+ =begin
97
+ \bullet Bullet character.
98
+ \lquote Left single quotation mark.
99
+ \rquote Right single quotation mark.
100
+ \ldblquote Left double quotation mark.
101
+ \rdblquote
102
+ =end
103
+ when :control_symbol; # ignore
104
+ group_type[group] ||= b
105
+ add_text = ' ' if b == '~' # non-breakable space
106
+ add_text = '-' if b == '_' # non-breakable hypen
107
+ when :text
108
+ add_text = b if group <= 1 or group_type[group] == 'rtlch' && !group_type[0...group].include?('*')
109
+ end
110
+ if format == :html
111
+ text << add_text.gsub(/([<>&"'])/) do
112
+ ent = { '<' => 'lt', '>' => 'gt', '&' => 'amp', '"' => 'quot', "'" => 'apos' }[$1]
113
+ "&#{ent};"
114
+ end
115
+ text << '<br>' if add_text == "\n"
116
+ else
117
+ text << add_text
118
+ end
119
+ end
120
+ text << "</body>\n</html>\n" if format == :html
121
+ text
122
+ end
123
+ end
124
+
23
125
  RTF_PREBUF =
24
126
  "{\\rtf1\\ansi\\mac\\deff0\\deftab720{\\fonttbl;}" \
25
127
  "{\\f0\\fnil \\froman \\fswiss \\fmodern \\fscript " \
data/test/test_msg.rb CHANGED
@@ -5,6 +5,7 @@ $: << "#{TEST_DIR}/../lib"
5
5
 
6
6
  require 'test/unit'
7
7
  require 'mapi/msg'
8
+ require 'mapi/convert'
8
9
 
9
10
  class TestMsg < Test::Unit::TestCase
10
11
  def test_blammo
@@ -17,12 +18,9 @@ class TestMsg < Test::Unit::TestCase
17
18
  assert_equal 66, msg.properties.raw.length
18
19
  # this is unique named properties
19
20
  assert_equal 48, msg.properties.to_h.length
20
- # get the named property keys
21
- keys = msg.properties.raw.keys.select { |key| String === key.code }
22
- assert_equal '55555555-5555-5555-c000-000000000046', keys[0].guid.format
23
- assert_equal 'Yippee555', msg.properties[keys[0]]
24
- assert_equal '66666666-6666-6666-c000-000000000046', keys[1].guid.format
25
- assert_equal 'Yippee666', msg.properties[keys[1]]
21
+ # test accessing the named property keys - same name but different namespace
22
+ assert_equal 'Yippee555', msg.props['Name4', Ole::Types::Clsid.parse('55555555-5555-5555-c000-000000000046')]
23
+ assert_equal 'Yippee666', msg.props['Name4', Ole::Types::Clsid.parse('66666666-6666-6666-c000-000000000046')]
26
24
  end
27
25
  end
28
26
  end
metadata CHANGED
@@ -1,116 +1,124 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: ruby-msg
3
3
  version: !ruby/object:Gem::Version
4
+ hash: 1
4
5
  prerelease: false
5
6
  segments:
6
- - 1
7
- - 5
8
- - 0
9
- version: 1.5.0
7
+ - 1
8
+ - 5
9
+ - 1
10
+ version: 1.5.1
10
11
  platform: ruby
11
12
  authors:
12
- - Charles Lowe
13
+ - Charles Lowe
13
14
  autorequire:
14
15
  bindir: bin
15
16
  cert_chain: []
16
17
 
17
- date: 2011-05-18 00:00:00 -04:00
18
+ date: 2012-07-03 00:00:00 +10:00
18
19
  default_executable:
19
20
  dependencies:
20
- - !ruby/object:Gem::Dependency
21
- name: ruby-ole
22
- prerelease: false
23
- requirement: &id001 !ruby/object:Gem::Requirement
24
- requirements:
25
- - - ">="
26
- - !ruby/object:Gem::Version
27
- segments:
28
- - 1
29
- - 2
30
- - 8
31
- version: 1.2.8
32
- type: :runtime
33
- version_requirements: *id001
34
- - !ruby/object:Gem::Dependency
35
- name: vpim
36
- prerelease: false
37
- requirement: &id002 !ruby/object:Gem::Requirement
38
- requirements:
39
- - - ">="
40
- - !ruby/object:Gem::Version
41
- segments:
42
- - 0
43
- - 360
44
- version: "0.360"
45
- type: :runtime
46
- version_requirements: *id002
21
+ - !ruby/object:Gem::Dependency
22
+ name: ruby-ole
23
+ prerelease: false
24
+ requirement: &id001 !ruby/object:Gem::Requirement
25
+ none: false
26
+ requirements:
27
+ - - ">="
28
+ - !ruby/object:Gem::Version
29
+ hash: 15
30
+ segments:
31
+ - 1
32
+ - 2
33
+ - 8
34
+ version: 1.2.8
35
+ type: :runtime
36
+ version_requirements: *id001
37
+ - !ruby/object:Gem::Dependency
38
+ name: vpim
39
+ prerelease: false
40
+ requirement: &id002 !ruby/object:Gem::Requirement
41
+ none: false
42
+ requirements:
43
+ - - ">="
44
+ - !ruby/object:Gem::Version
45
+ hash: 731
46
+ segments:
47
+ - 0
48
+ - 360
49
+ version: "0.360"
50
+ type: :runtime
51
+ version_requirements: *id002
47
52
  description: A library for reading and converting Outlook msg and pst files (mapi message stores).
48
53
  email: aquasync@gmail.com
49
54
  executables:
50
- - mapitool
55
+ - mapitool
51
56
  extensions: []
52
57
 
53
58
  extra_rdoc_files:
54
- - README
59
+ - README
55
60
  files:
56
- - data/mapitags.yaml
57
- - data/types.yaml
58
- - data/named_map.yaml
59
- - Rakefile
60
- - README
61
- - FIXES
62
- - bin/mapitool
63
- - lib/rtf.rb
64
- - lib/mapi.rb
65
- - lib/mime.rb
66
- - lib/orderedhash.rb
67
- - lib/mapi/rtf.rb
68
- - lib/mapi/property_set.rb
69
- - lib/mapi/pst.rb
70
- - lib/mapi/convert.rb
71
- - lib/mapi/types.rb
72
- - lib/mapi/msg.rb
73
- - lib/mapi/convert/contact.rb
74
- - lib/mapi/convert/note-mime.rb
75
- - lib/mapi/convert/note-tmail.rb
76
- - test/test_property_set.rb
77
- - test/test_convert_note.rb
78
- - test/test_mime.rb
79
- - test/test_convert_contact.rb
80
- - test/test_types.rb
81
- - test/test_msg.rb
61
+ - data/mapitags.yaml
62
+ - data/types.yaml
63
+ - data/named_map.yaml
64
+ - Rakefile
65
+ - README
66
+ - FIXES
67
+ - bin/mapitool
68
+ - lib/mapi.rb
69
+ - lib/mime.rb
70
+ - lib/orderedhash.rb
71
+ - lib/mapi/rtf.rb
72
+ - lib/mapi/property_set.rb
73
+ - lib/mapi/convert/contact.rb
74
+ - lib/mapi/convert/note-mime.rb
75
+ - lib/mapi/convert/note-tmail.rb
76
+ - lib/mapi/pst.rb
77
+ - lib/mapi/convert.rb
78
+ - lib/mapi/types.rb
79
+ - lib/mapi/msg.rb
80
+ - test/test_property_set.rb
81
+ - test/test_convert_note.rb
82
+ - test/test_mime.rb
83
+ - test/test_convert_contact.rb
84
+ - test/test_types.rb
85
+ - test/test_msg.rb
82
86
  has_rdoc: true
83
87
  homepage: http://code.google.com/p/ruby-msg
84
88
  licenses: []
85
89
 
86
90
  post_install_message:
87
91
  rdoc_options:
88
- - --main
89
- - README
90
- - --title
91
- - ruby-msg documentation
92
- - --tab-width
93
- - "2"
92
+ - --main
93
+ - README
94
+ - --title
95
+ - ruby-msg documentation
96
+ - --tab-width
97
+ - "2"
94
98
  require_paths:
95
- - lib
99
+ - lib
96
100
  required_ruby_version: !ruby/object:Gem::Requirement
101
+ none: false
97
102
  requirements:
98
- - - ">="
99
- - !ruby/object:Gem::Version
100
- segments:
101
- - 0
102
- version: "0"
103
+ - - ">="
104
+ - !ruby/object:Gem::Version
105
+ hash: 3
106
+ segments:
107
+ - 0
108
+ version: "0"
103
109
  required_rubygems_version: !ruby/object:Gem::Requirement
110
+ none: false
104
111
  requirements:
105
- - - ">="
106
- - !ruby/object:Gem::Version
107
- segments:
108
- - 0
109
- version: "0"
112
+ - - ">="
113
+ - !ruby/object:Gem::Version
114
+ hash: 3
115
+ segments:
116
+ - 0
117
+ version: "0"
110
118
  requirements: []
111
119
 
112
120
  rubyforge_project: ruby-msg
113
- rubygems_version: 1.3.6
121
+ rubygems_version: 1.3.7
114
122
  signing_key:
115
123
  specification_version: 3
116
124
  summary: Ruby Msg library.
data/lib/rtf.rb DELETED
@@ -1,109 +0,0 @@
1
- require 'stringio'
2
-
3
- # this file is pretty crap, its just to ensure there is always something readable if
4
- # there is an rtf only body, with no html encapsulation.
5
-
6
- module RTF
7
- class Tokenizer
8
- def self.process io
9
- while true do
10
- case c = io.getc
11
- when ?{; yield :open_group
12
- when ?}; yield :close_group
13
- when ?\\
14
- case c = io.getc
15
- when ?{, ?}, ?\\; yield :text, c.chr
16
- when ?'; yield :text, [io.read(2)].pack('H*')
17
- when ?a..?z, ?A..?Z
18
- # read control word
19
- str = c.chr
20
- str << c while c = io.read(1) and c =~ /[a-zA-Z]/
21
- neg = 1
22
- neg = -1 and c = io.read(1) if c == '-'
23
- num = if c =~ /[0-9]/
24
- num = c
25
- num << c while c = io.read(1) and c =~ /[0-9]/
26
- num.to_i * neg
27
- end
28
- raise "invalid rtf stream" if neg == -1 and !num # ???? \blahblah- some text
29
- io.seek(-1, IO::SEEK_CUR) if c != ' '
30
- yield :control_word, str, num
31
- when nil
32
- raise "invalid rtf stream" # \EOF
33
- else
34
- # other kind of control symbol
35
- yield :control_symbol, c.chr
36
- end
37
- when nil
38
- return
39
- when ?\r, ?\n
40
- # ignore
41
- else yield :text, c.chr
42
- end
43
- end
44
- end
45
- end
46
-
47
- class Converter
48
- # crappy
49
- def self.rtf2text str, format=:text
50
- group = 0
51
- text = ''
52
- text << "<html>\n<body>" if format == :html
53
- group_type = []
54
- group_tags = []
55
- RTF::Tokenizer.process(StringIO.new(str)) do |a, b, c|
56
- add_text = ''
57
- case a
58
- when :open_group; group += 1; group_type[group] = nil; group_tags[group] = []
59
- when :close_group; group_tags[group].reverse.each { |t| text << "</#{t}>" }; group -= 1;
60
- when :control_word; # ignore
61
- group_type[group] ||= b
62
- # maybe change this to use utf8 where possible
63
- add_text = if b == 'par' || b == 'line' || b == 'page'; "\n"
64
- elsif b == 'tab' || b == 'cell'; "\t"
65
- elsif b == 'endash' || b == 'emdash'; "-"
66
- elsif b == 'emspace' || b == 'enspace' || b == 'qmspace'; " "
67
- elsif b == 'ldblquote'; '"'
68
- else ''
69
- end
70
- if b == 'b' || b == 'i' and format == :html
71
- close = c == 0 ? '/' : ''
72
- text << "<#{close}#{b}>"
73
- if c == 0
74
- group_tags[group].delete b
75
- else
76
- group_tags[group] << b
77
- end
78
- end
79
- # lot of other ones belong in here.\
80
- =begin
81
- \bullet Bullet character.
82
- \lquote Left single quotation mark.
83
- \rquote Right single quotation mark.
84
- \ldblquote Left double quotation mark.
85
- \rdblquote
86
- =end
87
- when :control_symbol; # ignore
88
- group_type[group] ||= b
89
- add_text = ' ' if b == '~' # non-breakable space
90
- add_text = '-' if b == '_' # non-breakable hypen
91
- when :text
92
- add_text = b if group <= 1 or group_type[group] == 'rtlch' && !group_type[0...group].include?('*')
93
- end
94
- if format == :html
95
- text << add_text.gsub(/([<>&"'])/) do
96
- ent = { '<' => 'lt', '>' => 'gt', '&' => 'amp', '"' => 'quot', "'" => 'apos' }[$1]
97
- "&#{ent};"
98
- end
99
- text << '<br>' if add_text == "\n"
100
- else
101
- text << add_text
102
- end
103
- end
104
- text << "</body>\n</html>\n" if format == :html
105
- text
106
- end
107
- end
108
- end
109
-