summaryrefslogtreecommitdiff
path: root/kmime/Mainpage.dox
blob: 7f47e488025872b5ae7235a02f487005abf37128 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
/**

\mainpage The KMime Library

\section introduction Introduction

KMime is a library for handling mail messages and newsgroup articles. Both mail messages and
newsgroup articles are based on the same standard called MIME, which stands for
<b>Multipurpose Internet Mail Extensions</b>. In this document, the term \c message is used to
refer to both mail messages and newsgroup articles.

KMime deals solely with the in-memory representation of messages, topics such a transport or storage
of messages are handled by other libraries, for example by
<a href="http://api.kde.org/4.x-api/kdepimlibs-apidocs/mailtransport/html/index.html">the mailtransport library</a>
or by <a href="http://api.kde.org/4.x-api/kdepimlibs-apidocs/kimap/html/index.html">the KIMAP library</a>.
Similary, this library does not deal with displaying messages or advanced composing, for those there
are the <a href="http://api.kde.org/4.x-api/kdepim-apidocs/messageviewer/html/index.html">messageviewer</a>
and the <a href="http://websvn.kde.org/trunk/KDE/kdepim/messagecomposer/">messagecomposer</a>
components in the KDEPIM module.

KMime's main function is to parse, modify and assemble messages in-memory. In a
\ref string-broken-down "later section", <i>parsing</i> and <i>assembling</i> is actually explained.
KMime provides high-level classes that make these tasks easy.

MIME is defined by various RFCs, see the \ref rfcs "RFC section" for a list of them.

\section structure Structure of this document

This document will first give an \ref mime-intro "introduction to the MIME specification", as it is
essential to understand the basics of the structure of MIME messages for using this library.
The introduction here is aimed at users of the library, it gives a broad overview with examples and
omits some details. Developers who wish to modifiy KMime should read the
\ref rfcs "corresponding RFCs" as well, but this is not necessary for library users.

After the introduction to the MIME format, the two ways of representing a message in memory are
discussed, the \ref string-broken-down "string representation and the broken down representation".

This is followed by a section giving an
\ref classes-overview "overview of the most important KMime classes".

The last sections give a list of \ref rfcs "relevant RFCs" and provide links for
\ref links "further reading".

\section mime-intro Structure of MIME messages

\subsection history A brief history of the MIME standard

The MIME standard is quite new (1993), email and usenet existed way before the MIME standard came into
existence. Because of this, the MIME standard has to keep backwards compatibility. The email
standard before MIME lacked many capabilities like encodings other than ASCII or attachments. These
and other things were later added by MIME. The standard for messages before MIME is defined in
<a href="http://tools.ietf.org/html/rfc5322">RFC 5322</a>. In <a href="http://tools.ietf.org/html/rfc2045">RFC 2045</a>
to <a href="http://tools.ietf.org/html/rfc2049">RFC 2049</a>, several backward-compatible extensions
to the basic message format are defined, adding support for attachments, different encodings and many
others.

Actually, there is an even older standard, defined in <a href="http://tools.ietf.org/html/rfc733">RFC 733</a>
(<i>Standard for the format of ARPA network text messages</i>, introduced in 1977).
This standard is now obsoleted by RFC 5322, but backwards compatibilty is in some cases supported, as
there are still messages in this format around.

Since pre-MIME messages had no way to handle attachments, attachments were sometimes added to the message
text in an <a href="http://en.wikipedia.org/wiki/Uuencoding">uuencoded</a> form. Although this is also
obsolete, reading uuencoded attachments is still supported by KMime.

After MIME was introduced, people realized that there is no way to have the filename of attachments
encoded in anything different than ASCII. Thus, <a href="http://tools.ietf.org/html/rfc2231">RFC 2231</a>
was introduced to allow abitrary encodings for parameter values, such as the attachment filename.

\subsection examples MIME by examples

In the following sections, MIME message examples are shown, examined and explained, starting with
a simple message and proceeding to more interesting examples.
You can get additional examples by simply viewing the source of your own messages in your mail client,
or by having a look at the examples in the \ref rfcs "various RFCs".

\subsubsection simple-mail A simple message

\verbatim
Subject: First Mail
From: John Doe <john.doe@domain.com>
Date: Sun, 21 Feb 2010 19:16:11 +0100
MIME-Version: 1.0

Hello World!
\endverbatim
The above example features a very simple message. The two main parts of this message are the \b header
and the \b body, which are seperated by an empty line. The body contains the actual message content,
and the header contains metadata about the message itself. The header consists of several <b>header fields</b>,
each of them in their own line. Header fields are made up from the <b>header field name</b>, followed by a colon, followed
by the <b>header field body</b>.

The \b MIME-Version header field is mandatory for MIME messages. \b Subject,
\b From and \b Date are important header fields, they are usually displayed in the message list of a
mail client. The \c Subject header field can be anything, it does not have a special structure. It is a
so-called \b unstructured header field. In contrast, the \c From and the \c Date header fields have
to follow a special structure, they must be formed in a way that machines can parse. They are \b structured
header fields. For example, a mail client needs to understand
the \c Date header field so that it can sort the messages by date in the message list.
The exact details of how the header field bodies of structured header fields should be
formed are specified in an RFC.

In this example, the \c From header contains a single email address. More precisly, a single email address is called
a \b mailbox, which is made up of the <b>display name</b> (John Doe) and the <b>address specification</b> (john.doe@domain.com),
which is enclosed in angle brackets. The \c addr-spec consists of the user name, the <b>local part</b>,
and the \b domain name.

Many header fields can contain multiple email addresses, for example the \c To field for messages with
multiple recipients can have a comma-seperated list of mailboxes.
A list of mailboxes, together with a display name for the list, forms a \b group, and multiple groups can form an
<b>address list</b>. This is however rarely used, you'll most often see a simple list of plain mailboxes.

There are many more possible header fields than shown in this example, and the header can even contain
abitrary header fields, which usually are prefixed with \c X-, like \c X-Face.

\subsubsection encodings Encodings and charsets

\verbatim
From: John Doe <john.doe@domain.com>
Date: Mon, 22 Feb 2010 00:42:45 +0100
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Gr=FCezi Welt!
\endverbatim

The above shows a message that is using a different \b charset than the standard \b US-ASCII charset. The
message body contains the string "Grüezi Welt!", which is \b encoded in a special way.

The \b content-type of this message is \b text/plain, which means that the message is simple text. Later,
other content types will be introduced, such as \b text/html. If there is no \c Content-Type header
field, it is assumed that the content-type is \c text/plain.

Before MIME was introduced, all messages were limited to the US-ASCII charset. Only the
lower 127 values of the bytes were allowed to be used, the so-called \b 7-bit range. Writing a message in
another charset or using letters from the upper 127 byte values was not allowed.

\par Charset Encoding

When talking about charsets, it is important to understand how strings of text are converted to
byte arrays, and the other way around. A message is nothing else than a big array of bytes.
The bytes that form the body of the message somehow need to be interpreted as a text string. Interpreting
a byte array as a text string is called \b decoding the text. Converting a text string to a byte array is called
\b encoding the text. A \b codec (<b>co</b>der-<b>dec</b>oder) is a utility that can encode and decode text.
In Qt, the class for text strings is QString, and the class for byte arrays is QByteArray. The base class
of all codecs is QTextCodec.

With the US-ASCII charset, encoding and decoding text is easy, one just has to look at an <a href="http://en.wikipedia.org/wiki/ASCII_table">
ASCII table</a> to be able to convert text strings to byte arrays and byte arrays to text strings. For
example, the letter 'A' is represented by a single byte with the value of 65. When encountering a byte
with the value 84, we can look that up in the table and see that it represents the letter 'T'.
With the US-ASCII charset, each letter is represented by exactly one byte, which is very convenient.
Even better, all letters commonly used in English text have byte values below 127, so the 7-bit limit
of messages is no problem for text encoded with the US-ASCII charset.
Another example: The string "Hello World!" is represented by the following byte array:<br>
<code>48 65 6C 6C 6F 20 57 6F 72 6C 64</code><br>
Note that the byte values are written in hexadecimal form here, not in decimal as earlier.

Now, what if we want to write a message that contains German umlauts or Chinese letters? Those
are not in the ASCII table, therefore a different charset has to be used. There is a wealth of charsets
to chose from. Not all charsets can handle all letters, for example the
<a href="http://en.wikipedia.org/wiki/ISO-8859-1#ISO-8859-1">ISO-8859-1</a> charset can handle
German umlauts, but can not handle Chinese or Arabic letters. The <a href="http://en.wikipedia.org/wiki/Unicode">
Unicode standard</a> is an attempt to introduce charsets that can handle all known letters in the
world, in all languages. Unicode actually has several charsets, for example <a href="http://en.wikipedia.org/wiki/UTF-8">UTF-8</a>
and <a href="http://en.wikipedia.org/wiki/UTF-16">UTF-16</a>. In an ideal world, everyone would be using
Unicode charsets, but for historic and legacy reasons, other charsets are still much in use.

Charsets other than US-ASCII don't generally have as nice properties: A single letter can be represented
by multiple bytes, and generally the byte values are not in the 7-bit range. Pay attention to the UTF-8
charset: At first glance, it looks exactly like the US-ASCII charset, common latin letters like A - Z
are encoded with the same byte values as with US-ASCII. However, letters other than A - Z are suddenly
encoded with two or even more bytes. In general, one letter can be encoded in an abitrary number of bytes, depending
on the charset. One can \b not rely on the <code>1 letter == 1 byte</code> assumption.

Now, what should be done when the text string "Grüezi Welt!" should be sent in the body of a message?
The first step is to chose a charset that can represent all letters. This already excludes US-ASCII.
Once a charset is chosen, the text string is encoded into a byte array.
"Grüezi Welt!" encoded with the ISO-8859-1 charset produces the following byte array:<br>
<code>47 72 FC 65 7A 69 20 57 65 6C 74 21</code><br>
The letter 'ü' here is encoded using a single byte with the value <code>FC</code>.
The same string encoded with UTF-8 looks slightly different:<br>
<code>47 72 C3 BC 65 7A 69 20 57 65 6C 74 21</code><br>
Here, the letter 'ü' is encoded with two bytes, <code>C3 BC</code>. Still, one can see the similarity
between the two charsets for the other letters.

You can try this out yourself: Open your favorite text editor and enter some text with non-latin
letters. Then save the file and view it in a hex editor to see how the text was converted to a
byte array. Make sure to try out setting different charsets in your text editor.

At this point, the text string is sucessfully converted to a byte array, using e.g. the ISO-8859-1
charset. To indicate which charset was used, a \b Content-Type header field has to be added, with the correct
\b charset parameter. In our example above, that was done. If the charset parameter of the \c Content-Type,
or even the complete \c Content-Type header field is left out, the receiver can not know how to interpret
the byte array! In these cases, the byte array is usually decoded incorrectly, and the text strings contain
wrong letters or lots of questionmarks. There is even a special term for such wrongly decoded text,
<a href="http://en.wikipedia.org/wiki/Mojibake">Mojibake</a>. It is important to always know what charset
your byte array is encoded with, otherwise an attempt at decoding the byte array into a text string will fail and produce
Mojibake. <b>There is no such thing as plain text!</b> If there is no \c Content-Type header field in
a message, the message body should be interpreted as US-ASCII.

To learn more about charsets and encodings, read
<a href="http://www.joelonsoftware.com/articles/Unicode.html">The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)</a>
and <a href="http://www.cs.tut.fi/~jkorpela/chars.html">A tutorial on character code issues</a>. Especially
the first article should really be read, as the name indicates.

\par Content Transfer Encoding

Now, we can't use the byte array that was just created in a message. The string encoded with ISO-8859-1
has the byte value <code>FC</code> for the letter 'ü', which is decimal value 252. However, as said earlier,
messages are only valid when all bytes are in the 7-bit range, i.e. have byte value below 127.
So what should we do for byte values that are greater than 127, how can they be added to messages? The solution
for this is to use a <b>content transfer encoding</b> (CTE). A content transfer encoding takes a byte
array as input and transforms it. The output is another byte array, but one which only uses byte values
in the 7-bit range. One such content transfer encoding is <b>quoted-printable</b> (QTP), which is used in the
above example. Quoted-printable is easy to understand: When encountering a byte that has a value greater
than 127, it is simply replaced by a '=', followed by the hexadecimal code of the byte value, represented
as letters and digits encoded with ASCII. This means
that a byte with the value 252 is replaced with the ASCII string <code>=FC</code>, since <code>FC</code>
is the hexadecimal value of 252. The ASCII string <code>=FC</code> itself is now three bytes big,
<code>3D 46 43</code>. Therefore, the quoted-printable encoding replaces each byte outside of the 7-bit
range with 3 new bytes. Decoding quoted-printable encoding is also easy: Each time a byte with the value
<code>3D</code>, which is the letter '=' in ASCII, is encountered, the next two following bytes are interpreted
as the hex value of the resulting byte. The quoted-printable encoding was invented to make reading the
byte array easy for humans.

The quoted-printable encoding is not a good choice when the input byte array contains lots of bytes
outside the 7-bit range, as the resulting byte array will be three times as big in the worst case,
which is a waste of space. Therefore another content transfer encoding was introduced, \b Base64.
The details of the base64 encoding are too much to write about here; refer to the
<a href="http://en.wikipedia.org/wiki/Base64">Wikipedia article</a>
or the <a href="http://tools.ietf.org/html/rfc2045#section-6.8">RFC</a> for details.
As an example, the ISO-8859-1 encoded text string "Grüezi Welt!" is, after encoding it with base64,
represented by the following ASCII string: <code>R3L8ZXppIFdlbHQh</code>.
To express the same in byte arrays: The byte array <code>47 72 FC 65 7A 69 20 57 65 6C 74 21</code>
is, after encoding it with base64,
represented by the byte array <code>52 33 4C 38 5A 58 70 70 49 46 64 6C 62 48 51 68</code>

There are two other content transfer encodings besides quoted printable and base64: \b 7-bit and
\b 8-bit. 7-bit is just a marker to indicate that no content transfer encoding is used. This is the
case when the byte array is already completley in the 7-bit range, for example when writing English
text using the US-ASCII charset. 8-bit is also a marker to indicate that no content transfer encoding
was used. This time, not because it was not necessary, but because of a special exception, byte values
outside of the 7-bit range are allowed. For example, some SMTP servers support the
<a href="http://tools.ietf.org/html/rfc1652">8BITMIME</a> extension, which indicates that they accept
bytes outside of the 7-bit range. In this case, one can simply use the byte arrays as-is, without using
any content transfer encoding. Creating messages with 8-bit content transfer encoding is currently not
supported by KMime. The advantage of 8-bit is that there is no overhead in size, unlike with
base64 or even quoted-printable.

When using one of the 4 contents transfer encodings, i.e. quoted-printable, base64, 7-bit or 8-bit, this
has to be indicated in the header field \b Content-Transfer-Encoding. If the header field is left out,
it is assumed that the content transfer encoding is 7-bit. The example above uses quoted-printable.

\verbatim
From: John Doe <john.doe@domain.com>
Date: Mon, 22 Feb 2010 00:42:45 +0100
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: base64

R3L8ZXppIFdlbHQh
\endverbatim
The same example, this time encoded with the base64 content transfer encoding.

\verbatim
From: John Doe <john.doe@domain.com>
Date: Mon, 22 Feb 2010 00:42:45 +0100
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: base64

R3LDvGV6aSBXZWx0IQ==
\endverbatim
Again the same example, this time using UTF-8 as the charset.

\verbatim
From: John Doe <john.doe@domain.com>
Date: Mon, 22 Feb 2010 00:42:45 +0100
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Gr=C3=BCezi Welt!
\endverbatim
The example with a combination of UTF-8 and quoted-printable CTE. As said somewhere above, with the
UTF-8 encoding, the letter 'ü' is represented by the two bytes <code>C3 BC</code>.

\verbatim
From: John Doe <john.doe@domain.com>
Date: Mon, 22 Feb 2010 00:42:45 +0100
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: 7-bit

Hello World
\endverbatim
A different example, showing 7-bit content transfer encoding. Although the UTF-8 charset has lots
of letters that are represented by bytes outside of the 7-bit range, the string "Hello World" can
be fully represented in the 7-bit range here, even with UTF-8.

In the \ref links "further reading" section, you will find links to web applications that demonstrate
encodings and charsets.

\par Conclusion

When adding a text string to the body of a message, it needs to be encoded twice: First, the encoding of the charset
needs to be applied, which transforms the text string into a byte array. Afterwards, the content transfer
encoding has to be applied, which transforms the byte array from the first step into a byte array that
only has bytes in the 7-bit range.

When decoding, the same has to be done, in reverse: One first has decode the byte array with the content transfer encoding, to get a byte
array that has all 256 possible byte values. Afterwards, the resulting byte array needs to be decoded
with the correct charset, to transform it into a text string. For those two decoding steps, one has to
look at the \c Content-Type and the \c Content-Transfer-Encoding header fields to find the correct
charset and CTE for decoding.

It is important to always keep the charset and the content transfer encoding in mind. Byte arrays and
strings are not to be confused. Byte arrays that are encoded with a CTE are not to be confused with
byte arrays that are \b not encoded with a CTE.

This section showed how to use different charsets in the <i>body</i> of a message. The next section will
show what to do when another charset is needed in one of the <i>header</i> field bodies.

\subsubsection header-encoding Encoding in Header Fields

In the last section, we discussed how to use different charsets in the body of a message. But what if
a different charset needs to be added to one of the header fields? For example one might want to write
a mail to a mailbox with the display name "András Manţia" and with the subject "Grüezi!".

The header fields are limited to characters in the 7-bit range, and are interpreted as US-ASCII.
That means the header field names, such as "From: ", are all encoded in US-ASCII. The header field
bodies, such as the "1.0" of \c MIME-Version, are also encoded with US-ASCII. This is mandated by
<a href="http://tools.ietf.org/html/rfc5322#section-2">the RFC</a>.

The \c Content-Type and the \c Content-Transfer-Encoding header fields only apply to the message body,
they have no meaning for other header fields.

This means that any letter in a different charset has to be encoded in some way to statisfy the RFC.
Letters with a different charset are only allowed in some of the header field bodies, the header field
names always have to be in US-ASCII.

\verbatim
From: Thomas McGuire <thomas@domain.com>
Subject: =?iso-8859-1?q?Gr=FCezi!?=
Date: Mon, 22 Feb 2010 14:34:01 +0100
MIME-Version: 1.0
To: =?utf-8?q?Andr=C3=A1s?= =?utf-8?q?_Man=C5=A3ia?= <andras@domain.com>
Content-Type: Text/Plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

bla bla bla
\endverbatim

The above example shows how text that is encoded with a different charset than US-ASCII is handled
in the message header. This can be seen in the bodies of the \c Subject header field and the \c To header field.
In this example, the body of the message is unimportant, it is just "bla bla bla" in US-ASCII.
The way the header field bodies are encoded is sometimes referred to as a <b>RFC2047 string</b> or as an <b>encoded word</b>, which has
the origin in the <a href="http://tools.ietf.org/html/rfc2047">RFC</a> where this encoding scheme is defined.
RFC2047 strings are only allowed in some of the header fields, like \c Subject and in the display name
of mailboxes in header fields like \c From and \c To. In other header fields, such as \c Date and
\c MIME-Version, they are not allowed, but they wouldn't make much sense there anyway, since those are
structured header fields with a clearly defined structure.

RFC2047 strings start with "=?" and end with "?=". Between those markers, they consists of three parts:
\li The charset, such as "iso-8859-1"
\li The encoding, which is "q" or "b"
\li The encoded text

These three parts are sperated with a '?'. Encoding the third part, the text, is very similar to how
text strings in the message body are encoded: First, the text string is encoded to a byte array using
the charset encoding. Afterwards, the second encoding is used on the result, to ensure that all resulting
bytes are within the 7-bit range.

The <i>second encoding</i> here is almost identical to the content transfer encoding. There are two
possible encodings, \b b and \b q. The \c b encoding is the same as the base64 encoding of the content
transfer encoding. The \c q encoding is very similar to the quoted-printable encoding of the content
transfer encoding, but with some little differences that are described in
<a href="http://tools.ietf.org/html/rfc2047#section-4.2">the RFC</a>.

Let's examine the subject of the message, <code>=?iso-8859-1?q?Gr=FCezi!?=</code>, in detail:<br>
The first part of the RFC2027 string is the charset, so it is ISO-8859-1 in this case. The second part
is the encoding, which is the \c q encoding here. The last part is the encoded text, which is
<code>Gr=FCezi!</code>. As with the quoted-printable encoding, "=FC" is the encoding for the byte with
the value <code>FC</code>, which in the ISO-8859-1 charset is the letter 'ü'. The complete decoded
text is therefore "Grüezi!".

Each RFC2047 string in the header can use a different charset: In this example, the \c Subject uses ISO-8859-1,
\c To uses UTF-8 and the message body uses US-ASCII.

In the \c To header field, two RFC2047 strings are used. A single, bigger, RFC2047 string for the whole
display name could also have been used. In this case, the second RFC2047 string starts with an underscore,
which is decoded as a space in the \c q encoding. The space between the two RFC2047 strings is ignored,
it is just used to seperate the two encoded words.

There are some restriction on RFC2047 strings: They are not allowed to be longer than 75 characters,
which means two or more encoded words have to be used for long text strings. Also, there are some
restrictions on where RFC2047 strings are allowed; most importantly, the address specification must no
not be encoded, to be backwards compatible. For further details, refer to the RFC.

\subsubsection multipart-mixed Messages with attachments

Until now, we only looked at messages that had a single text part as the message body. In this section,
we'll examine messages with attachments.

\verbatim
From: frank@domain.com
To: greg@domain.com
Subject: Nice Photo
Date: Sun, 28 Feb 2010 19:57:00 +0100
MIME-Version: 1.0
Content-Type: Multipart/Mixed;
  boundary="Boundary-00=_8xriL5W6LSj00Ly"

--Boundary-00=_8xriL5W6LSj00Ly
Content-Type: Text/Plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi Greg,

attached you'll find a nice photo.

--Boundary-00=_8xriL5W6LSj00Ly
Content-Type: image/jpeg;
  name="test.jpeg"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
  filename="test.jpeg"

/9j/4AAQSkZJRgABAQAAAQABAAD/4Q3XRXhpZgAASUkqAAgAAAAHAAsAAgAPAAAAYgAAAAABBAAB
[SNIP 800 lines]
ze5CdSH2Z8yTatHSV2veW0rKzeq30//Z

--Boundary-00=_8xriL5W6LSj00Ly--

\endverbatim
<i>Note: Since the image in this message would be really big, most of it is omitted / snipped here.</i>

The above example consists of two parts: A normal text part and an image attachment. Messages that
consist of multiple parts are called \b multipart messages. The top-level content-type therefore is
<b>multipart/mixed</b>. \c Mixed simply means that the following parts have no relation to each other,
it is just a random mixture of parts. Later, we will look at other types, such as \c multipart/alternative
or \c multipart/related. A \b part is sometimes also called \b node, \b content or <b>MIME part</b>.

Each MIME part of the message is seperated by a \b boundary, and that boundary
is specified in the top-level content-type header as a parameter. In the message body, the boundary
is prefixed with \c "--", and the last boundary is suffixed with \c "--", so that the end of the message can
be detected. When creating a message, care must be taken that the boundary appears nowhere else in the
message, for example in the text part, as the parser would get confused by this.

A MIME part begins right after the boundary. It consists of a <b>MIME header</b> and a <b>MIME body</b>, which
are seperated by an empty line. The MIME header should not be confused with the message header: The
message header contains metadata about the whole message, like subject and date. The MIME header only
contains metadata about the specific MIME part, like the content type of the MIME part. MIME header
field names always start with \c "Content-".
The example above shows the three most important MIME header fields, usually those are the only ones
used. The top-level header of a message actually mixes the message metadata and the MIME metadata into one header: In this
example, the header contains the \c Date header field, which is an ordinary header field, and it contains
the \c Content-Type header field, which is a MIME header field.

MIME parts can be nested, and therefore form a tree. The above example has the following tree:
\verbatim
multipart/mixed
|- text/plain
\- image/jpeg
\endverbatim
The \c text/plain node is therefore a \b child of the \c multipart/mixed mode. The \c multipart/mixed node
is a \b parent of the other two nodes. The \c image/jpeg node is a \b sibling of the \c text/plain node.
\c Multipart nodes are the only nodes that have children, other nodes are \b leaf nodes.
The body of a multipart node consists of all complete child nodes (MIME header and MIME body), seperated
by the boundary.

Each MIME part can have a different content transfer encoding. In the above example, the text part has
a \c 7bit CTE, while the image part has a \c base64 CTE. The multipart/mixed node does not specifiy
a CTE, multipart nodes always have \c 7bit as the CTE. This is because the body of multipart nodes can
only consist of bytes in the 7 bit range: The boundary is 7 bit, the MIME headers are 7 bit, and the
MIME bodies are already ancoded with the CTE of the child MIME part, and are therefore also 7 bit. This means
no CTE for multipart nodes is necessary.

The MIME part for the image does not specify a charset parameter in the content type header field. This
is because the body of that MIME part will not be interpreted as a text string, therefore the byte array
does not need to be decoded to a string. Instead, the byte array is interpreted as an image, by an image
renderer. The message viewer application passes the MIME part body as a byte array to the image renderer.
The content type consists of a <b>media type</b> and a <b>subtype</b>. For example, the content type
\c "text/html" has the media type "text" and the subtype "html". Only nodes that have the media type "text"
need to specify a charset, as those nodes are the only nodes of which the body is interpreted as a text string.

The only header field not yet encountered in previous sections is the \b Content-Disposition header field,
which is defined in a <a href="http://tools.ietf.org/html/rfc2183">seperate RFC</a>. It describes how
the message viewer application should display the MIME part. In the case of the image part, is should
be presented as an attachment. The \b filename parameter tells the message viewer application which filename
should be used by default when the user saves the attachment to disk.

The content type header field for the image MIME part has a \b name parameter, which is similar to the
\c filename parameter of the \c Content-Disposition header field. The difference is that \c name refers
to the name of the complete MIME part, whereas \c filename refers to the name of the attachment. The
\c name paramter of the \c Content-Type header field in this case is superfluous and only exists for
backwards compatibility, and can be ignored;
the \c filename parameter of the \c Content-Disposition header field should be prefered when it is present.

\verbatim
From: Thomas McGuire <thomas@domain.com>
To: sebastian@domain.com
Subject: Help with SPARQL
Date: Sun, 28 Feb 2010 21:57:51 +0100
MIME-Version: 1.0
Content-Type: Multipart/Mixed;
  boundary="Boundary-00=_PjtiLU2PvHpvp/R"

--Boundary-00=_PjtiLU2PvHpvp/R
Content-Type: Text/Plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi Sebastian,

I have a problem with a SPARQL query, can you help me debug this? Attached is
the query and a screenshot showing the result.

--Boundary-00=_PjtiLU2PvHpvp/R
Content-Type: text/plain;
  charset="UTF-8";
  name="query.txt"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
  filename="query.txt"

prefix nco:<http://www.semanticdesktop.org/ontologies/2007/03/22/nco#>

SELECT ?person
WHERE {
  ?person a nco:PersonContact .
  ?person nco:birthDate ?birthDate .
}"
--Boundary-00=_PjtiLU2PvHpvp/R
Content-Type: image/png;
  name="screenshot.png"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
  filename="screenshot.png"

AAAAyAAAAAEBBAABAAAAyAAAAA0BAgATAAAAcQAAABIBAwABAAAAAQAAADEBAgAPAAAAhAAAAGmH
[SNIP]
YXJlLmpwZWcAZGlnaUthbS0w

--Boundary-00=_PjtiLU2PvHpvp/R--
\endverbatim
The above example message consists of three MIME parts: The main text part and two attachments.
One attachment has the media type \c text, therefore a charset parameter is necessary to correctly
display it. The MIME tree looks like this:
\verbatim
multipart/mixed
|- text/plain
|- text/plain
\- image/jpeg
\endverbatim

\subsubsection multipart-alternative HTML Messages

\verbatim
From: Thomas McGuire <thomas@domain.com>
Subject: HTML test
Date: Thu, 4 Mar 2010 13:59:18 +0100
MIME-Version: 1.0
Content-Type: multipart/alternative;
  boundary="Boundary-01=_m66jLd2/vZrH5oe"
Content-Transfer-Encoding: 7bit

--Boundary-01=_m66jLd2/vZrH5oe
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hello World

--Boundary-01=_m66jLd2/vZrH5oe
Content-Type: text/html;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN" "http://www.w3.org/TR/REC-html40/strict.dtd">
<html>
  <head></head>
  <body>
    Hello <b>World</b>
  </body>
</html>
--Boundary-01=_m66jLd2/vZrH5oe--
\endverbatim

The above example is a simple HTML message, it consists of a plain text and a HTML part, which are
in a \b multipart/alternative container. The message has the following structure:
\verbatim
multipart/alternative
|- text/plain
\- text/html
\endverbatim
The HTML part and the plain text part have the identical content, except that the HTML part contains
additional markup, in this case for displaying the word \c World in bold. Since those parts are in a
multipart/alternative container, the message viewer application can freely chose which part it displays.
Some users might prefer reading the message in HTML format, some might prefer reading the message
in plain text format.

Of course, a HTML message could also consist only of a single \c text/html, without the multipart/alternative
container and therefore without an alternative plain text part. However, people prefering the plain
text version wouldn't like this, especially if their mail client has no HTML engine and they would see
the HTML source including all tags only. Therefore, HTML messages should always include an alternative plain text part.

HTML messages can of course also contain attachments. In this case, the message contains both a
multipart/alternative and a multipart/mixed node, for example with the following structure, for a HTML
message that has an image attachment:
\verbatim
multipart/mixed
|- multipart/alternative
|  |- text/plain
|  \- text/html
\- image/png
\endverbatim

The message itself would look like this:
\verbatim
From: Thomas McGuire <thomas@domain.com>
Subject: HTML message with an attachment
Date: Thu, 4 Mar 2010 15:20:26 +0100
MIME-Version: 1.0
Content-Type: Multipart/Mixed;
  boundary="Boundary-00=_qG8jLwWCwkUfJV1"

--Boundary-00=_qG8jLwWCwkUfJV1
Content-Type: multipart/alternative;
  boundary="Boundary-01=_qG8jLfs1FRmlOhl"
Content-Transfer-Encoding: 7bit

--Boundary-01=_qG8jLfs1FRmlOhl
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hello World

--Boundary-01=_qG8jLfs1FRmlOhl
Content-Type: text/html;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN" "http://www.w3.org/TR/REC-html40/strict.dtd">
<html>
  <head></head>
  <body>
    Hello <b>World</b>
  </body>
</html>
--Boundary-01=_qG8jLfs1FRmlOhl--

--Boundary-00=_qG8jLwWCwkUfJV1
Content-Type: image/png;
  name="test.png"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
  filename="test.png"

iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/9hAAAACXBIWXMAAA8SAAAPEgEhm/IzAAAC
[SNIP]
eFkXsFgBMG4fJhYlx+iyB3cLpNZwYr/iP7teTwNYa7DZAAAAAElFTkSuQmCC

--Boundary-00=_qG8jLwWCwkUfJV1--

\endverbatim

\subsubsection multipart-related HTML Messages with Inline Images

HTML has support for showing images, with the \c img tag. Such an image is shown at the place where
the \c img tag occurs, which is called an <b>inline image</b>. Note that inline images are different
from images that are just normal attachments: Normal attachments are always shown at the beginning or
at the end of the message, while inline images are shown in-place. In HTML, the \c img tag points to an
image file that is either a file on disk or an URL to an image on the Internet. To make inline images
work with MIME messages, a different mechanism is needed, since the image is not a file on disk or on
the Internet, but a MIME part somewhere in the same message. As specified in
<a href="http://tools.ietf.org/html/rfc2557">RFC 2557</a>, the way this can be done is by refering
to a \b Content-ID in the \c img tag, and marking the MIME part that is the image with that content
ID as well.

An example will probably be more clear than this explaination:
\verbatim
From: Thomas McGuire <thomas@domain.com>
Subject: Inine Image Test
Date: Thu, 4 Mar 2010 16:54:53 +0100
MIME-Version: 1.0
Content-Type: multipart/related;
  boundary="Boundary-02=_Nf9jLpJ2aGp5RQK"
Content-Transfer-Encoding: 7bit

--Boundary-02=_Nf9jLpJ2aGp5RQK
Content-Type: multipart/alternative;
  boundary="Boundary-01=_Nf9jLZ6aPhm3WrN"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

--Boundary-01=_Nf9jLZ6aPhm3WrN
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Text before image

Text after image

--Boundary-01=_Nf9jLZ6aPhm3WrN
Content-Type: text/html;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN" "http://www.w3.org/TR/REC-html40/strict.dtd">
<html>
  <head></head>
  <body>
    Text before image<br>
    <img src="cid:547730348@KDE" /><br>
    Text after image
  </body>
</html>
--Boundary-01=_Nf9jLZ6aPhm3WrN--

--Boundary-02=_Nf9jLpJ2aGp5RQK
Content-Type: image/png;
  name="test.png"
Content-Transfer-Encoding: base64
Content-Id: <547730348@KDE>

iVBORw0KGgoAAAANSUhEUgAAAMgAAADICAIAAAAiOjnJAAAACXBIWXMAAA7EAAAOxAGVKw4bAAAg
[SNIP]
AABJRU5ErkJggg==
--Boundary-02=_Nf9jLpJ2aGp5RQK--
\endverbatim

The first thing you'll notice in this example probably is that it has a \b multipart/related node with
the following structure:
\verbatim
multipart/related
|- multipart/alternative
|  |- text/plain
|  \- text/html
\- image/png
\endverbatim

When the HTML part has inline image, the HTML part and its image part both have to be children of a
multipart/related container, like in this example.
In this case, the \c img tag has the source \c cid:547730348@KDE, which is a placeholder that refers
to the Content-Id header of another part. The image part contains exactly that value in its \c Content-Id
header, and therefore a message viewer application can connect both.

The plain text part can not have inline images, therefore its text might seem a bit confusing.

HTML messages with inline images can of course also have attachments, in which the message structure
becomes a mix of multipart/related, multipart/alternative and multipart/mixed. The following example
shows the structure of a message with two inline images and one \c .tar.gz attachment:
\verbatim
multipart/mixed
|- multipart/related
|  |- multipart/alternative
|  |  |- text/plain
|  |  \- text/html
|  |- image/png
|  \- image/png
\- application/x-compressed-tar
\endverbatim

The structure of MIME messages can get arbitrarily complex, the above is just one relativley simply example.
The nesting of multipart nodes can get much deeper, there is no restriction on nesting levels.

\subsubsection encapsulated Encapsulated messages

Encapsulated messages are messages which are attachments to another message. The most common example
is a forwareded mail, like in this example:
\verbatim
From: Frank <frank@domain.com>
To: Bob <bob@domain.com>
Subject: Fwd: Blub
MIME-Version: 1.0
Content-Type: Multipart/Mixed;
  boundary="Boundary-00=_sX+jLVPkV1bLFdZ"

--Boundary-00=_sX+jLVPkV1bLFdZ
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi Bob,

hereby I forward you an interesting message from Greg.

--Boundary-00=_sX+jLVPkV1bLFdZ
Content-Type: message/rfc822;
  name="forwarded message"
Content-Transfer-Encoding: 7bit
Content-Description: Forwarded Message
Content-Disposition: inline

From: Greg <greg@domain.com>
To: Frank <frank@domain.com>
Subject: Blub
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Bla Bla Bla

--Boundary-00=_sX+jLVPkV1bLFdZ--
\endverbatim

\verbatim
multipart/mixed
|- text/plain
\- message/rfc822
   \- text/plain
\endverbatim

The attached message is treated like any other attachment, and therefore the top-level content type
is multipart/mixed.
The most interesting part is the \c message/rfc822 MIME part. As usual, it has some MIME headers, like
\c Content-Type or \c Content-Disposition, followed by the MIME body. The MIME body in this case is
the attached message. Since it is a message, it consists of a header and a body itself.
Therefore, the \c message/rfc822 MIME part appears to have two headers; in reality, it is the normal
MIME header and the message header of the encapsulated message. The message header and the message body
are both in the MIME body of the \c message/rfc822 MIME part.

\subsubsection crypto Signed and Encryped Messages

MIME messages can be cryptographically signed and/or encrypted. The format for those messages is
defined in <a href="http://tools.ietf.org/html/rfc1847">RFC 1847</a>, which specifies two new
multipart subtypes, \b multipart/signed and \b multipart/encrypted. The crypto format of these new
security multiparts is defined in additional RFCs; the most common formats are
<a href="http://tools.ietf.org/html/rfc3156">OpenPGP</a> and
<a href="http://tools.ietf.org/html/rfc2633">S/MIME</a>. Both formats use the principle of
<a href="http://en.wikipedia.org/wiki/Public-key_cryptography">public-key cryptography</a>. OpenPGP
uses \b keys, and S/MIME uses \b certificates. For easier text flow, only the term \c key will be used
for both keys and certificates in the text below.

Security multiparts only sign or encrypt a specifc MIME part. The consequence is that the message headers
can not be signed or encrypted. Also this means that it is possible to sign or encrypt only some of
the MIME parts of a message, while leaving other MIME parts unsigned or unencrypted. Furthermore, it
is possible to sign or encrypt different MIME parts with different crypto formats. As you can see,
security multiparts are very flexible.

Security multiparts are not supported by KMime. However, it is possible for applications to use KMime
when providing support for crypto messages. For example, the
<a href="http://api.kde.org/4.x-api/kdepim-apidocs/messageviewer/html/index.html">messageviewer</a>
component in KDEPIM supports signed and encrypted MIME parts, and the
<a href="http://websvn.kde.org/trunk/KDE/kdepim/messagecomposer/">messagecomposer</a> library can create
such messages.

Signed MIME parts are signed with the private key of the sender, everybody who has the
public key of the sender can verifiy the signature. Encrypted MIME parts are encrypted with the public
key of the receiver, and only the receiver, who is the sole person possessing the private key, can decrypt
it. Sending an encrypted message to multiple recipients therefore means that the message has to be sent
multiple times, once for each receiver, as each message needs to be encrypted with a different key.

\par Signed MIME parts

A multipart/signed MIME part has exactly two children: The first child is the content that is signed,
and the second child is the signature.

\verbatim
From: Thomas McGuire <thomas@domain.com>
Subject: My Subject
Date: Mon, 15 Mar 2010 12:20:16 +0100
MIME-Version: 1.0
Content-Type: multipart/signed;
  boundary="nextPart2567247.O5e8xBmMpa";
  protocol="application/pgp-signature";
  micalg=pgp-sha1
Content-Transfer-Encoding: 7bit

--nextPart2567247.O5e8xBmMpa
Content-Type: Text/Plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit

Simple message

--nextPart2567247.O5e8xBmMpa
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)

iEYEABECAAYFAkueF/UACgkQKglv3sO8a1MdTACgnBEP6ZUal931Vwu7PyiXT1bn
Zr0Anj4bAI9JhHEDiwA/iwrWGfSC+Nlz
=d2ol
-----END PGP SIGNATURE-----
--nextPart2567247.O5e8xBmMpa--
\endverbatim

\verbatim
multipart/signed
|- text/plain
\- application/pgp-signature
\endverbatim
The example here uses the OpenPGP format to sign a simply plain text message. Here, the text/plain
MIME part is signed, and the application/pgp-signature MIME part contains the signature data, which in
this case is ASCII-armored.

As said above, it is possible to sign only some MIME parts. A message which has a image/jpeg attachment
that is signed, but a main text part is not signed, has the following MIME structure:
\verbatim
multipart/mixed
|- text/plain
\- multipart/signed
   |- image/jpeg
   \- application/pgp-signature
\endverbatim

It is possible to sign multipart parts as well. Consider the above example that has a plain text part
and an image attachment. Those two parts can be signed together, with the following structure:
\verbatim
multipart/signed
|- multipart/mixed
|  |- text/plain
|  \- image/jpeg
\- application/pgp-signature
\endverbatim

Signed messages in the S/MIME format use a different content type for the signature data, like here:
\verbatim
multipart/signed
|- text/plain
\- application/x-pkcs7-signature
\endverbatim

\par Encrypted MIME parts

Multipart/encrypted MIME parts also have exactly two children: The first child contains metadata about
the encrypted data, such as a version number. The second child then contains the actual encrypted data.

\verbatim
From: someone@domain.com
To: Thomas McGuire <thomas@domain.com>
Subject: Encrypted message
Date: Mon, 15 Mar 2010 12:50:16 +0100
MIME-Version: 1.0
Content-Type: multipart/encrypted;
  boundary="nextPart2726747.j47xUGTWKg";
  protocol="application/pgp-encrypted"
Content-Transfer-Encoding: 7bit

--nextPart2726747.j47xUGTWKg
Content-Type: application/pgp-encrypted
Content-Disposition: attachment

Version: 1
--nextPart2726747.j47xUGTWKg
Content-Type: application/octet-stream
Content-Disposition: inline; filename="msg.asc"

-----BEGIN PGP MESSAGE-----
Version: GnuPG v2.0.14 (GNU/Linux)

hQIOA8p5rdC5CBNfEAf+NZVzVq48C1r5opOOiWV96+FUzIWuMQ6u8fzFgI7YVyCn
[SNIP]
=reNr
--nextPart2726747.j47xUGTWKg--
-----END PGP MESSAGE-----
\endverbatim

\verbatim
multipart/encrypted
|- application/pgp-encrypted
\- application/octet-stream
\endverbatim

The encrypted data is contained in the \c application/octet-stream MIME part. Without decrypting
the data, it is unknown what the original content type of the encrypted MIME data is! The encrypted
data could be a simple text/plain MIME part, an image attachment, or a multipart part. The encrypted
data contains both the MIME header and the MIME body of the original MIME part, as the header is needed
to know the content type of the data. The data could as well by of content type multipart/signed, in
which case the message would be both signed and encrypted.

\par Inline cryto formats

Although using the security multiparts \c multipart/signed and \c multipart/encrypted is the recommended
standard, there are other possibilities to sign or encrypt a message. The most common methods are
<b>Inline OpenPGP</b> and <b>S/MIME Opaque</b>.

For inline OpenPGP messages, the crypto data is contained inlined in the actual MIME part. For example,
a message with a signed text/plain part might look like this:
\verbatim
From: someone@domain.com
To: someoneelse@domain.com
Subject: Inline OpenPGP test
MIME-Version: 1.0
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Inline OpenPGP signed example.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)

iEYEARECAAYFAkueJ2EACgkQKglv3sO8a1MS3QCfcsYnJG7uYQxzxz6J5cPF7lHz
WIoAn3PjVPlWibu02dfdFObwd2eJ1jAW
=p3uO
-----END PGP SIGNATURE-----

\endverbatim

Encrypted inline OpenPGP works in a similar way. Opaque S/MIME messages are also similar: For signed
MIME parts, both the signature and the signed data are contained in a single MIME part with a content
type of \c application/pkcs7-mime.

As security multiparts are preferred over inline OpenPGP and over opaque S/MIME, I won't go into more
detail here.

\subsubsection misc Miscellaneous Points about Messages

\par Line Breaks

Each line in a MIME message has to end with a \b CRLF, which is a carriage return followed by a
newline, which is the escape sequence<code>\\r\\n</code>. CR and LF may not appear in other places in
a MIME message. Special care needs to be taken with encoded linebreaks in binary data, and with
distinguishing soft and hard line breaks when converting between different content transfer encodings.
For more details, have a look at the RFCs.

While the official format is to have a CRLF at the end of each line, KMime only expects a single LF
for its in-memory storage. Therefore, when loading a message from disk or from a server into KMime, the CRLFs need
to be converted to LFs first, for example with KMime::CRLFtoLF(). The opposite needs to be done when
storing a KMime message somewhere.

Lines should not be longer than 78 characters and may not be longer than 998 characters.

\par Header Folding and CFWS

Header fields can span multiple lines, which was already shown in some of the examples above where
the parameters of the header field value were in the next line. The header field is said to be
\b folded in this case. In general, header fields can be folded whenever whitespace (\b WS) occurs.

Header field values can contain \b comments; these comments are semantically invisible and have no
meaning. Comments are surrouned by parentheses.
\verbatim
Date: Thu, 13
      Feb 1969 23:32 -0330 (Newfoundland Time)
\endverbatim

This example shows a folded header that also has a comment (<i>Newfoundland Time</i>). The date header is a structured header
field, and therefore it has to obey to a defined syntax; however, adding comments and whitespace is
allowed almost anywhere, and they are ignored when parsing the message. Comments and whitespace where
folding is allowed is sometimes referred to as \b CFWS. Any occurence of CFWS is semantically regarded
as a single space.

\section string-broken-down The two in-memory representations of messages

There are two representations of messages in memory. The first is called <b>string representation</b>
and the other one is called <b>broken-down representation</b>.

String representation is somehow misnamed,
a better term would be <c>byte array representation</c>. The string representation is just a big array of
bytes in memory, and those bytes make up the encoded mail. The string representation is what is stored
on disk or what is received from an IMAP server, for example.

With the broken-down representation, the mail is <i>broken down</i> into smaller structures. For example,
instead of having a single byte array for all headers, the broken-down structure has a list of individual headers,
and each header in that list is again broken down into a structure. While the string representation
is just an array of 7 bit characters that might be encoded, the broken-down representations contain the
decoded text strings.

As an example, conside the byte array
\verbatim
"Hugo Maier" <hugo.maier@mailer.domain>
\endverbatim

Although this is just a bunch of 7 bit characters, a human immediatley recognizes the broken-down structure and
sees that the display name is "Hugo Maier" and that the localpart of the email address is "hugo.maier".
To illustrate, the broken-down structure could be stored in a structure like this:
\verbatim
struct Mailbox
{
  QString displayName;
  QByteArray addressSpec;
};
\endverbatim
The address spec actually could be broken down further into a localpart and a domain.
The process of converting the string representation to a broken-down representation is called \b parsing, and
the reverse is called \b assembling.
Parsing a message is necessary when wanting to access or modify the broken-down structure. For example, when sending a mail,
the address spec of a mailbox needs to be passed to the SMTP server, which means that the recipient headers need to
be parsed in order to access that information. Another example is the message list in an mail application, where the
broken-down structure of a mail is needed
to display information like subject, sender and date in the list.
On the other hand, assembling a message is for example done in the composer of a mail application, where the mail information
is available in a broken-down form in the composer window, and is then assembled into a final MIME message that is then sent with SMTP.

Parsing is often quite tricky, you should always use the methods from KMime instead of writing parsing
routines yourself. Even the simple mailbox example above is in pratice difficult to parse, as many things like comments
and escaped characters need to be taken into consideration.
The same is true for assembling: In the above case, one could be tempted to assemble the mailbox by simply
writting code like this:
\verbatim
QByteArray stringRepresentation = '"' + displayName + "\" <" + addressSpec + ">";
\endverbatim
However, just like with parsing, you shouldn't be doing assembling yourself. In the above case, for example,
the display name might contain non-ASCII characters, and RFC2047 encoding would need to be applied. So use
KMime for assembling in all cases.

When parsing a message and assembling it afterwards, the result might not be the same as the original byte
array. For example, comments in header fields are ignored during parsing and not stored in the broken-down
structure, therefore the assembled message will also not contain comments.

Messages in memory are usually stored in a broken-down structure so that it is easy to to access and
manipulate the message. On disk and on servers, messages are stored in string representation.

\section classes-overview Overview of KMime classes

KMime has basically two sets of classes: Classes for headers and classes for MIME
parts. A MIME part is represented by \c KMime::Content. A Content can be parsed from a string representation
and also be assembled from the broken-down representation again. If parsed, it has a list of sub-contents (in case of multipart contents) and a
list of headers. If the Content is not parsed, it stores the headers and the body in a byte array, which can be accessed
with head() and body().
There is also a class \c KMime::Message, which basically is a thin wrapper around Content for the top-level
MIME part. Message also contains convenience methods to access the message headers.

For headers, there is a class hierachy, with \c KMime::Headers::Base as the base class, and
\c KMime::Headers::Generics::Structured and \c KMime::Headers::Generics::Unstructured in the next levels. Unstructured is
for headers that don't have a defined structure, like Subject, whereas Structured headers have a
specific structure, like Date. The header classes have methods to parse headers, like from7BitString(),
and to assemble them, like as7BitString(). Once a header is parsed, the classes provide access to the
broken-down structures, for example the Date header has a method dateTime().
The parsing in from7BitString() is usually handled by a protected parse() function, which in turn call
parsing functions for different types, like parseAddressList() or parseAddrSpec() from the \c KMime::HeaderParsing
namespace.

When modifing messages, the message is first parsed into a broken-down representation. This broken-down
representation can then be accessed and modified with the appropriate functions. After changing the broken-down
structure, it needs to be assembled again to get the modified string representation.

KMime also comes with some codes for handling base64 and quoted-printable encoding, with \c KMime::Codec
as the base class.

\section rfcs RFCs

\li <a href="http://tools.ietf.org/html/rfc5322">RFC 5322</a>: Internet Message Format
\li <a href="http://tools.ietf.org/html/rfc5536">RFC 5536</a>: Netnews Article Format
\li <a href="http://tools.ietf.org/html/rfc2045">RFC 2045</a>: Multipurpose Internet Mail Extensions (MIME), Part 1: Format of Internet Message Bodies
\li <a href="http://tools.ietf.org/html/rfc2046">RFC 2046</a>: Multipurpose Internet Mail Extensions (MIME), Part 2: Media Types
\li <a href="http://tools.ietf.org/html/rfc2047">RFC 2047</a>: Multipurpose Internet Mail Extensions (MIME), Part 3: Message Header Extensions for Non-ASCII Text
\li <a href="http://tools.ietf.org/html/rfc2048">RFC 2048</a>: Multipurpose Internet Mail Extensions (MIME), Part 4: Registration Procedures
\li <a href="http://tools.ietf.org/html/rfc2049">RFC 2049</a>: Multipurpose Internet Mail Extensions (MIME), Part 5: Conformance Criteria and Examples
\li <a href="http://tools.ietf.org/html/rfc2231">RFC 2231</a>: MIME Parameter Value and Encoded Word Extensions: Character Sets, Languages, and Continuations
\li <a href="http://tools.ietf.org/html/rfc2183">RFC 2183</a>: Communicating Presentation Information in Internet Message: The Content-Disposition Header Field
\li <a href="http://tools.ietf.org/html/rfc2557">RFC 2557</a>: MIME Encapsulation of Aggregate Documents, such as HTML (MHTML)
\li <a href="http://tools.ietf.org/html/rfc1847">RFC 1847</a>: Security Multiparts for MIME: Multipart/Signed and Multipart/Encrypted
\li <a href="http://tools.ietf.org/html/rfc3851">RFC 3851</a>: S/MIME Version 3 Message Specification
\li <a href="http://tools.ietf.org/html/rfc3156">RFC 3156</a>: MIME Security with OpenPGP
\li <a href="http://tools.ietf.org/html/rfc2298">RFC 2298</a>: An Extensible Message Format for Message Disposition Notifications
\li <a href="http://tools.ietf.org/html/rfc2646">RFC 2646</a>: The Text/Plain Format Parameter (not supported by KMime)

\section links Further Reading
\li <a href="http://en.wikipedia.org/wiki/MIME">Wikipedia article on MIME</a>\n
\li <a href="http://www.joelonsoftware.com/articles/Unicode.html">The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)</a>
\li <a href="http://www.cs.tut.fi/~jkorpela/chars.html">A tutorial on character code issues</a>
\li <a href="http://www.motobit.com/util/base64-decoder-encoder.asp">Online Base64 encoder and decoder</a>
\li <a href="http://www.motobit.com/util/quoted-printable-encoder.asp">Online quoted-printable encoder</a>
\li <a href="http://www.motobit.com/util/quoted-printable-decoder.asp">Online quoted-printable decoder</a>
\li <a href="http://www.motobit.com/util/charset-codepage-conversion.asp">Online charset converter</a>
\li <a href="http://en.wikipedia.org/wiki/Public-key_cryptography">Wikipedia article on public-key cryptography</a>

\authors

The major authors of this library are:
\li Christian Gebauer
\li Volker Krause \<vkrause@kde.org\>
\li Marc Mutz \<mutz@kde.org\>
\li Christian Thurner \<cthurner@freepage.de\>
\li Tom Albers \<tomalbers@kde.nl\>
\li Thomas McGuire \<mcguire@kde.org\>

This document was written by:\n
\li Thomas McGuire \<mcguire@kde.org\>

\maintainers
\li Volker Krause \<vkrause@kde.org\>
\li Marc Mutz \<mutz@kde.org\>

\licenses
\lgpl

*/

// DOXYGEN_PROJECTNAME=KMIME Library