1
|
|
|
<?php |
2
|
|
|
/** |
3
|
|
|
* This file is part of the ZBateson\MailMimeParser project. |
4
|
|
|
* |
5
|
|
|
* @license http://opensource.org/licenses/bsd-license.php BSD |
6
|
|
|
*/ |
7
|
|
|
namespace ZBateson\MailMimeParser\Message; |
8
|
|
|
|
9
|
|
|
use Psr\Http\Message\StreamInterface; |
10
|
|
|
use ZBateson\MailMimeParser\Message\Part\PartBuilder; |
11
|
|
|
use ZBateson\MailMimeParser\Message\Part\Factory\PartBuilderFactory; |
12
|
|
|
use ZBateson\MailMimeParser\Message\Part\Factory\PartFactoryService; |
13
|
|
|
use GuzzleHttp\Psr7\StreamWrapper; |
14
|
|
|
|
15
|
|
|
/** |
16
|
|
|
* Parses a mail mime message into its component parts. To invoke, call |
17
|
|
|
* MailMimeParser::parse. |
18
|
|
|
* |
19
|
|
|
* @author Zaahid Bateson |
20
|
|
|
*/ |
21
|
|
|
class MessageParser |
22
|
|
|
{ |
23
|
|
|
/** |
24
|
|
|
* @var PartFactoryService service instance used to create MimePartFactory |
25
|
|
|
* objects. |
26
|
|
|
*/ |
27
|
|
|
protected $partFactoryService; |
28
|
|
|
|
29
|
|
|
/** |
30
|
|
|
* @var PartBuilderFactory used to create PartBuilders |
31
|
|
|
*/ |
32
|
|
|
protected $partBuilderFactory; |
33
|
|
|
|
34
|
|
|
/** |
35
|
|
|
* @var int maintains the character length of the last line separator, |
36
|
|
|
* typically 2 for CRLF, to keep track of the correct 'end' position |
37
|
|
|
* for a part because the CRLF before a boundary is considered part of |
38
|
|
|
* the boundary. |
39
|
|
|
*/ |
40
|
|
|
private $lastLineSeparatorLength = 0; |
41
|
|
|
|
42
|
|
|
/** |
43
|
|
|
* Sets up the parser with its dependencies. |
44
|
|
|
* |
45
|
|
|
* @param PartFactoryService $pfs |
46
|
|
|
* @param PartBuilderFactory $pbf |
47
|
|
|
*/ |
48
|
7 |
|
public function __construct( |
49
|
|
|
PartFactoryService $pfs, |
50
|
|
|
PartBuilderFactory $pbf |
51
|
|
|
) { |
52
|
7 |
|
$this->partFactoryService = $pfs; |
53
|
7 |
|
$this->partBuilderFactory = $pbf; |
54
|
7 |
|
} |
55
|
|
|
|
56
|
|
|
/** |
57
|
|
|
* Parses the passed stream into a ZBateson\MailMimeParser\Message object |
58
|
|
|
* and returns it. |
59
|
|
|
* |
60
|
|
|
* @param StreamInterface $stream the stream to parse the message from |
61
|
|
|
* @return \ZBateson\MailMimeParser\Message |
62
|
|
|
*/ |
63
|
7 |
|
public function parse(StreamInterface $stream) |
64
|
|
|
{ |
65
|
7 |
|
$partBuilder = $this->read($stream); |
66
|
7 |
|
return $partBuilder->createMessagePart($stream); |
67
|
|
|
} |
68
|
|
|
|
69
|
|
|
/** |
70
|
|
|
* Ensures the header isn't empty and contains a colon separator character, |
71
|
|
|
* then splits it and calls $partBuilder->addHeader. |
72
|
|
|
* |
73
|
|
|
* @param string $header |
74
|
|
|
* @param PartBuilder $partBuilder |
75
|
|
|
*/ |
76
|
6 |
|
private function addRawHeaderToPart($header, PartBuilder $partBuilder) |
77
|
|
|
{ |
78
|
6 |
|
if ($header !== '' && strpos($header, ':') !== false) { |
79
|
6 |
|
$a = explode(':', $header, 2); |
80
|
6 |
|
$partBuilder->addHeader($a[0], trim($a[1])); |
81
|
|
|
} |
82
|
6 |
|
} |
83
|
|
|
|
84
|
|
|
/** |
85
|
|
|
* Reads a line of up to 4096 characters. If the line is larger than that, |
86
|
|
|
* the remaining characters in the line are read and discarded, and only the |
87
|
|
|
* first 4096 characters are returned. |
88
|
|
|
* |
89
|
|
|
* @param resource $handle |
90
|
|
|
* @return string |
91
|
|
|
*/ |
92
|
6 |
|
private function readLine($handle) |
93
|
|
|
{ |
94
|
6 |
|
$size = 4096; |
95
|
6 |
|
$ret = $line = fgets($handle, $size); |
96
|
6 |
|
while (strlen($line) === $size - 1 && substr($line, -1) !== "\n") { |
97
|
1 |
|
$line = fgets($handle, $size); |
98
|
|
|
} |
99
|
6 |
|
return $ret; |
100
|
|
|
} |
101
|
|
|
|
102
|
|
|
/** |
103
|
|
|
* Reads a line of 2048 characters. If the line is larger than that, the |
104
|
|
|
* remaining characters in the line are read and |
105
|
|
|
* discarded, and only the first part is returned. |
106
|
|
|
* |
107
|
|
|
* This method is identical to readLine, except it calculates the number of |
108
|
|
|
* characters that make up the line's new line characters (e.g. 2 for "\r\n" |
109
|
|
|
* or 1 for "\n"). |
110
|
|
|
* |
111
|
|
|
* @param resource $handle |
112
|
|
|
* @param int $lineSeparatorLength |
113
|
|
|
* @return string |
114
|
|
|
*/ |
115
|
4 |
|
private function readBoundaryLine($handle, &$lineSeparatorLength = 0) |
116
|
|
|
{ |
117
|
4 |
|
$size = 2048; |
118
|
4 |
|
$isCut = false; |
119
|
4 |
|
$line = fgets($handle, $size); |
120
|
4 |
|
while (strlen($line) === $size - 1 && substr($line, -1) !== "\n") { |
121
|
|
|
$line = fgets($handle, $size); |
122
|
|
|
$isCut = true; |
123
|
|
|
} |
124
|
4 |
|
$ret = rtrim($line, "\r\n"); |
125
|
4 |
|
$lineSeparatorLength = strlen($line) - strlen($ret); |
126
|
4 |
|
return ($isCut) ? '' : $ret; |
127
|
|
|
} |
128
|
|
|
|
129
|
|
|
/** |
130
|
|
|
* Reads header lines up to an empty line, adding them to the passed |
131
|
|
|
* $partBuilder. |
132
|
|
|
* |
133
|
|
|
* @param resource $handle the resource handle to read from |
134
|
|
|
* @param PartBuilder $partBuilder the current part to add headers to |
135
|
|
|
*/ |
136
|
6 |
|
protected function readHeaders($handle, PartBuilder $partBuilder) |
137
|
|
|
{ |
138
|
6 |
|
$header = ''; |
139
|
|
|
do { |
140
|
6 |
|
$line = $this->readLine($handle); |
141
|
6 |
|
if (empty($line) || $line[0] !== "\t" && $line[0] !== ' ') { |
142
|
6 |
|
$this->addRawHeaderToPart($header, $partBuilder); |
143
|
6 |
|
$header = ''; |
144
|
|
|
} else { |
145
|
2 |
|
$line = "\r\n" . $line; |
146
|
|
|
} |
147
|
6 |
|
$header .= rtrim($line, "\r\n"); |
148
|
6 |
|
} while ($header !== ''); |
149
|
6 |
|
} |
150
|
|
|
|
151
|
|
|
/** |
152
|
|
|
* Reads lines from the passed $handle, calling |
153
|
|
|
* $partBuilder->setEndBoundaryFound with the passed line until it returns |
154
|
|
|
* true or the stream is at EOF. |
155
|
|
|
* |
156
|
|
|
* setEndBoundaryFound returns true if the passed line matches a boundary |
157
|
|
|
* for the $partBuilder itself or any of its parents. |
158
|
|
|
* |
159
|
|
|
* Once a boundary is found, setStreamPartAndContentEndPos is called with |
160
|
|
|
* the passed $handle's read pos before the boundary and its line separator |
161
|
|
|
* were read. |
162
|
|
|
* |
163
|
|
|
* @param resource $handle |
164
|
|
|
* @param PartBuilder $partBuilder |
165
|
|
|
*/ |
166
|
4 |
|
private function findContentBoundary($handle, PartBuilder $partBuilder) |
167
|
|
|
{ |
168
|
|
|
// last separator before a boundary belongs to the boundary, and is not |
169
|
|
|
// part of the current part |
170
|
4 |
|
while (!feof($handle)) { |
171
|
4 |
|
$endPos = ftell($handle) - $this->lastLineSeparatorLength; |
172
|
4 |
|
$line = $this->readBoundaryLine($handle, $this->lastLineSeparatorLength); |
173
|
4 |
|
if ($line !== '' && $partBuilder->setEndBoundaryFound($line)) { |
174
|
2 |
|
$partBuilder->setStreamPartAndContentEndPos($endPos); |
175
|
2 |
|
return; |
176
|
|
|
} |
177
|
|
|
} |
178
|
4 |
|
$partBuilder->setStreamPartAndContentEndPos(ftell($handle)); |
179
|
4 |
|
$partBuilder->setEof(); |
180
|
4 |
|
} |
181
|
|
|
|
182
|
|
|
/** |
183
|
|
|
* Reads content for a non-mime message. If there are uuencoded attachment |
184
|
|
|
* parts in the message (denoted by 'begin' lines), those parts are read and |
185
|
|
|
* added to the passed $partBuilder as children. |
186
|
|
|
* |
187
|
|
|
* @param resource $handle |
188
|
|
|
* @param PartBuilder $partBuilder |
189
|
|
|
* @return string |
190
|
|
|
*/ |
191
|
3 |
|
protected function readUUEncodedOrPlainTextMessage($handle, PartBuilder $partBuilder) |
192
|
|
|
{ |
193
|
3 |
|
$partBuilder->setStreamContentStartPos(ftell($handle)); |
194
|
3 |
|
$part = $partBuilder; |
195
|
3 |
|
while (!feof($handle)) { |
196
|
2 |
|
$start = ftell($handle); |
197
|
2 |
|
$line = trim($this->readLine($handle)); |
198
|
2 |
|
if (preg_match('/^begin ([0-7]{3}) (.*)$/', $line, $matches)) { |
199
|
1 |
|
$part = $this->partBuilderFactory->newPartBuilder( |
200
|
1 |
|
$this->partFactoryService->getUUEncodedPartFactory() |
201
|
|
|
); |
202
|
1 |
|
$part->setStreamPartStartPos($start); |
203
|
|
|
// 'begin' line is part of the content |
204
|
1 |
|
$part->setStreamContentStartPos($start); |
205
|
1 |
|
$part->setProperty('mode', $matches[1]); |
206
|
1 |
|
$part->setProperty('filename', $matches[2]); |
207
|
1 |
|
$partBuilder->addChild($part); |
208
|
|
|
} |
209
|
2 |
|
$part->setStreamPartAndContentEndPos(ftell($handle)); |
210
|
|
|
} |
211
|
3 |
|
$partBuilder->setStreamPartEndPos(ftell($handle)); |
212
|
3 |
|
} |
213
|
|
|
|
214
|
|
|
/** |
215
|
|
|
* Reads content for a single part of a MIME message. |
216
|
|
|
* |
217
|
|
|
* If the part being read is in turn a multipart part, readPart is called on |
218
|
|
|
* it recursively to read its headers and content. |
219
|
|
|
* |
220
|
|
|
* The start/end positions of the part's content are set on the passed |
221
|
|
|
* $partBuilder, which in turn sets the end position of the part and its |
222
|
|
|
* parents. |
223
|
|
|
* |
224
|
|
|
* @param resource $handle |
225
|
|
|
* @param PartBuilder $partBuilder |
226
|
|
|
*/ |
227
|
4 |
|
private function readPartContent($handle, PartBuilder $partBuilder) |
228
|
|
|
{ |
229
|
4 |
|
$partBuilder->setStreamContentStartPos(ftell($handle)); |
230
|
4 |
|
$this->findContentBoundary($handle, $partBuilder); |
231
|
4 |
|
if ($partBuilder->isMultiPart()) { |
232
|
2 |
|
while (!$partBuilder->isParentBoundaryFound()) { |
233
|
2 |
|
$child = $this->partBuilderFactory->newPartBuilder( |
234
|
2 |
|
$this->partFactoryService->getMimePartFactory() |
235
|
|
|
); |
236
|
2 |
|
$partBuilder->addChild($child); |
237
|
2 |
|
$this->readPart($handle, $child); |
238
|
|
|
} |
239
|
|
|
} |
240
|
4 |
|
} |
241
|
|
|
|
242
|
|
|
/** |
243
|
|
|
* Reads a part and any of its children, into the passed $partBuilder, |
244
|
|
|
* either by calling readUUEncodedOrPlainTextMessage or readPartContent |
245
|
|
|
* after reading headers. |
246
|
|
|
* |
247
|
|
|
* @param resource $handle |
248
|
|
|
* @param PartBuilder $partBuilder |
249
|
|
|
*/ |
250
|
7 |
|
protected function readPart($handle, PartBuilder $partBuilder) |
251
|
|
|
{ |
252
|
7 |
|
$partBuilder->setStreamPartStartPos(ftell($handle)); |
253
|
|
|
|
254
|
7 |
|
if ($partBuilder->canHaveHeaders()) { |
255
|
6 |
|
$this->readHeaders($handle, $partBuilder); |
256
|
6 |
|
$this->lastLineSeparatorLength = 0; |
257
|
|
|
} |
258
|
7 |
|
if ($partBuilder->getParent() === null && !$partBuilder->isMime()) { |
259
|
3 |
|
$this->readUUEncodedOrPlainTextMessage($handle, $partBuilder); |
260
|
|
|
} else { |
261
|
4 |
|
$this->readPartContent($handle, $partBuilder); |
262
|
|
|
} |
263
|
7 |
|
} |
264
|
|
|
|
265
|
|
|
/** |
266
|
|
|
* Reads the message from the passed stream and returns a PartBuilder |
267
|
|
|
* representing it. |
268
|
|
|
* |
269
|
|
|
* @param StreamInterface $stream |
270
|
|
|
* @return PartBuilder |
271
|
|
|
*/ |
272
|
7 |
|
protected function read(StreamInterface $stream) |
273
|
|
|
{ |
274
|
7 |
|
$partBuilder = $this->partBuilderFactory->newPartBuilder( |
275
|
7 |
|
$this->partFactoryService->getMessageFactory() |
276
|
|
|
); |
277
|
|
|
// the remaining parts use a resource handle for better performance... |
278
|
|
|
// it seems fgets does much better than Psr7\readline (not specifically |
279
|
|
|
// measured, but difference in running tests is big) |
280
|
7 |
|
$this->readPart(StreamWrapper::getResource($stream), $partBuilder); |
281
|
7 |
|
return $partBuilder; |
282
|
|
|
} |
283
|
|
|
} |
284
|
|
|
|