1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
|
4
|
|
|
namespace Egulias\EmailValidator\Parser; |
5
|
|
|
|
6
|
|
|
use Egulias\EmailValidator\EmailLexer; |
7
|
|
|
use Egulias\EmailValidator\EmailValidator; |
8
|
|
|
|
9
|
|
|
class DomainPart extends Parser |
10
|
|
|
{ |
11
|
|
|
const DOMAIN_MAX_LENGTH = 254; |
12
|
|
|
protected $domainPart = ''; |
13
|
|
|
|
14
|
115 |
|
public function parse($domainPart) |
15
|
|
|
{ |
16
|
115 |
|
$this->lexer->moveNext(); |
17
|
|
|
|
18
|
115 |
|
if ($this->lexer->token['type'] === EmailLexer::S_DOT) { |
19
|
1 |
|
throw new \InvalidArgumentException('ERR_DOT_START'); |
20
|
|
|
} |
21
|
|
|
|
22
|
114 |
|
if ($this->lexer->token['type'] === EmailLexer::S_EMPTY) { |
23
|
5 |
|
throw new \InvalidArgumentException('ERR_NODOMAIN'); |
24
|
|
|
} |
25
|
109 |
|
if ($this->lexer->token['type'] === EmailLexer::S_HYPHEN) { |
26
|
|
|
throw new \InvalidArgumentException('ERR_DOMAINHYPHENEND'); |
27
|
|
|
} |
28
|
|
|
|
29
|
109 |
|
if ($this->lexer->token['type'] === EmailLexer::S_OPENPARENTHESIS) { |
30
|
3 |
|
$this->warnings[] = EmailValidator::DEPREC_COMMENT; |
31
|
3 |
|
$this->parseDomainComments(); |
32
|
1 |
|
} |
33
|
|
|
|
34
|
107 |
|
$domain = $this->doParseDomainPart(); |
35
|
|
|
|
36
|
78 |
|
$prev = $this->lexer->getPrevious(); |
37
|
78 |
|
$length = strlen($domain); |
38
|
|
|
|
39
|
78 |
|
if ($prev['type'] === EmailLexer::S_DOT) { |
40
|
2 |
|
throw new \InvalidArgumentException('ERR_DOT_END'); |
41
|
|
|
} |
42
|
76 |
|
if ($prev['type'] === EmailLexer::S_HYPHEN) { |
43
|
1 |
|
throw new \InvalidArgumentException('ERR_DOMAINHYPHENEND'); |
44
|
|
|
} |
45
|
75 |
|
if ($length > self::DOMAIN_MAX_LENGTH) { |
46
|
4 |
|
$this->warnings[] = EmailValidator::RFC5322_DOMAIN_TOOLONG; |
47
|
4 |
|
} |
48
|
75 |
|
if ($prev['type'] === EmailLexer::S_CR) { |
49
|
|
|
throw new \InvalidArgumentException('ERR_FWS_CRLF_END'); |
50
|
|
|
} |
51
|
75 |
|
$this->domainPart = $domain; |
52
|
75 |
|
} |
53
|
|
|
|
54
|
75 |
|
public function getDomainPart() |
55
|
|
|
{ |
56
|
75 |
|
return $this->domainPart; |
57
|
|
|
} |
58
|
|
|
|
59
|
14 |
|
public function checkIPV6Tag($addressLiteral, $maxGroups = 8) |
60
|
|
|
{ |
61
|
14 |
|
$prev = $this->lexer->getPrevious(); |
62
|
14 |
|
if ($prev['type'] === EmailLexer::S_COLON) { |
63
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_COLONEND; |
64
|
2 |
|
} |
65
|
|
|
|
66
|
14 |
|
$IPv6 = substr($addressLiteral, 5); |
67
|
|
|
//Daniel Marschall's new IPv6 testing strategy |
68
|
14 |
|
$matchesIP = explode(':', $IPv6); |
69
|
14 |
|
$groupCount = count($matchesIP); |
70
|
14 |
|
$colons = strpos($IPv6, '::'); |
71
|
|
|
|
72
|
14 |
|
if (count(preg_grep('/^[0-9A-Fa-f]{0,4}$/', $matchesIP, PREG_GREP_INVERT)) !== 0) { |
73
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_BADCHAR; |
74
|
2 |
|
} |
75
|
|
|
|
76
|
14 |
|
if ($colons === false) { |
77
|
|
|
// We need exactly the right number of groups |
78
|
8 |
|
if ($groupCount !== $maxGroups) { |
79
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_GRPCOUNT; |
80
|
2 |
|
} |
81
|
8 |
|
return; |
82
|
|
|
} |
83
|
|
|
|
84
|
6 |
|
if ($colons !== strrpos($IPv6, '::')) { |
85
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_2X2XCOLON; |
86
|
2 |
|
return; |
87
|
|
|
} |
88
|
|
|
|
89
|
4 |
|
if ($colons === 0 || $colons === (strlen($IPv6) - 2)) { |
90
|
|
|
// RFC 4291 allows :: at the start or end of an address |
91
|
|
|
//with 7 other groups in addition |
92
|
4 |
|
++$maxGroups; |
93
|
4 |
|
} |
94
|
|
|
|
95
|
4 |
|
if ($groupCount > $maxGroups) { |
96
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_MAXGRPS; |
97
|
4 |
|
} elseif ($groupCount === $maxGroups) { |
98
|
2 |
|
$this->warnings[] = EmailValidator::RFC5321_IPV6DEPRECATED; |
99
|
2 |
|
} |
100
|
4 |
|
} |
101
|
|
|
|
102
|
107 |
|
protected function doParseDomainPart() |
103
|
|
|
{ |
104
|
107 |
|
$domain = ''; |
105
|
107 |
|
$openedParenthesis = 0; |
106
|
107 |
|
$openBrackets = false; |
107
|
|
|
do { |
108
|
107 |
|
$prev = $this->lexer->getPrevious(); |
109
|
|
|
|
110
|
107 |
|
$this->checkNotAllowedChars($this->lexer->token); |
111
|
|
|
|
112
|
107 |
|
if ($this->lexer->token['type'] === EmailLexer::S_OPENPARENTHESIS) { |
113
|
4 |
|
$this->parseComments(); |
114
|
4 |
|
$openedParenthesis += $this->getOpenedParenthesis(); |
115
|
4 |
|
$this->lexer->moveNext(); |
116
|
4 |
|
$tmpPrev = $this->lexer->getPrevious(); |
117
|
4 |
|
if ($tmpPrev['type'] === EmailLexer::S_CLOSEPARENTHESIS) { |
118
|
4 |
|
$openedParenthesis--; |
119
|
4 |
|
} |
120
|
4 |
|
} |
121
|
107 |
|
if ($this->lexer->token['type'] === EmailLexer::S_CLOSEPARENTHESIS) { |
122
|
3 |
|
if ($openedParenthesis === 0) { |
123
|
3 |
|
throw new \InvalidArgumentException('ERR_UNOPENEDCOMMENT'); |
124
|
|
|
} else { |
125
|
|
|
$openedParenthesis--; |
126
|
|
|
} |
127
|
|
|
} |
128
|
|
|
|
129
|
106 |
|
$this->checkConsecutiveDots(); |
130
|
106 |
|
$this->checkDomainPartExceptions($prev); |
131
|
|
|
|
132
|
105 |
|
if ($openBrackets = $this->hasBrackets($openBrackets)) { |
133
|
24 |
|
$this->parseDomainLiteral(); |
134
|
22 |
|
} |
135
|
|
|
|
136
|
103 |
|
$this->checkLabelLength($prev); |
137
|
|
|
|
138
|
103 |
|
if ($this->isFWS()) { |
139
|
10 |
|
$this->parseFWS(); |
140
|
8 |
|
} |
141
|
|
|
|
142
|
103 |
|
$domain .= $this->lexer->token['value']; |
143
|
103 |
|
$this->lexer->moveNext(); |
144
|
103 |
|
} while ($this->lexer->token); |
145
|
|
|
|
146
|
78 |
|
return $domain; |
147
|
|
|
} |
148
|
|
|
|
149
|
107 |
|
private function checkNotAllowedChars($token) |
150
|
|
|
{ |
151
|
107 |
|
$notAllowed = array(EmailLexer::S_BACKSLASH => true, EmailLexer::S_SLASH=> true); |
152
|
107 |
|
if (isset($notAllowed[$token['type']])) { |
153
|
8 |
|
throw new \InvalidArgumentException('ERR_DOMAIN_CHAR_NOT_ALLOWED'); |
154
|
|
|
} |
155
|
107 |
|
} |
156
|
|
|
|
157
|
24 |
|
protected function parseDomainLiteral() |
158
|
|
|
{ |
159
|
24 |
|
if ($this->lexer->isNextToken(EmailLexer::S_COLON)) { |
160
|
|
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_COLONSTRT; |
161
|
|
|
} |
162
|
24 |
|
if ($this->lexer->isNextToken(EmailLexer::S_IPV6TAG)) { |
163
|
14 |
|
$lexer = clone $this->lexer; |
164
|
14 |
|
$lexer->moveNext(); |
165
|
14 |
|
if ($lexer->isNextToken(EmailLexer::S_DOUBLECOLON)) { |
166
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_IPV6_COLONSTRT; |
167
|
2 |
|
} |
168
|
14 |
|
} |
169
|
|
|
|
170
|
24 |
|
return $this->doParseDomainLiteral(); |
171
|
|
|
} |
172
|
|
|
|
173
|
24 |
|
protected function doParseDomainLiteral() |
174
|
|
|
{ |
175
|
24 |
|
$IPv6TAG = false; |
176
|
24 |
|
$addressLiteral = ''; |
177
|
|
|
do { |
178
|
24 |
|
if ($this->lexer->token['type'] === EmailLexer::C_NUL) { |
179
|
|
|
throw new \InvalidArgumentException('ERR_EXPECTING_DTEXT'); |
180
|
|
|
} |
181
|
|
|
|
182
|
24 |
|
if ($this->lexer->token['type'] === EmailLexer::INVALID || |
183
|
24 |
|
$this->lexer->token['type'] === EmailLexer::C_DEL || |
184
|
24 |
|
$this->lexer->token['type'] === EmailLexer::S_LF |
185
|
24 |
|
) { |
186
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_DOMLIT_OBSDTEXT; |
187
|
2 |
|
} |
188
|
|
|
|
189
|
24 |
|
if ($this->lexer->isNextTokenAny(array(EmailLexer::S_OPENQBRACKET, EmailLexer::S_OPENBRACKET))) { |
190
|
1 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_DTEXT'); |
191
|
|
|
} |
192
|
|
|
|
193
|
23 |
|
if ($this->lexer->isNextTokenAny( |
194
|
23 |
|
array(EmailLexer::S_HTAB, EmailLexer::S_SP, $this->lexer->token['type'] === EmailLexer::CRLF) |
195
|
23 |
|
)) { |
196
|
|
|
$this->warnings[] = EmailValidator::CFWS_FWS; |
197
|
|
|
$this->parseFWS(); |
198
|
|
|
} |
199
|
|
|
|
200
|
23 |
|
if ($this->lexer->isNextToken(EmailLexer::S_CR)) { |
201
|
1 |
|
throw new \InvalidArgumentException('ERR_CR_NO_LF'); |
202
|
|
|
} |
203
|
22 |
|
if ($this->lexer->token['type'] === EmailLexer::S_BACKSLASH) { |
204
|
|
|
$this->warnings[] = EmailValidator::RFC5322_DOMLIT_OBSDTEXT; |
205
|
|
|
$addressLiteral .= $this->lexer->token['value']; |
206
|
|
|
$this->lexer->moveNext(); |
207
|
|
|
$this->validateQuotedPair(); |
208
|
|
|
} |
209
|
22 |
|
if ($this->lexer->token['type'] === EmailLexer::S_IPV6TAG) { |
210
|
14 |
|
$IPv6TAG = true; |
211
|
14 |
|
} |
212
|
22 |
|
if ($this->lexer->token['type'] === EmailLexer::S_CLOSEQBRACKET) { |
213
|
|
|
break; |
214
|
|
|
} |
215
|
|
|
|
216
|
22 |
|
$addressLiteral .= $this->lexer->token['value']; |
217
|
|
|
|
218
|
22 |
|
} while ($this->lexer->moveNext()); |
219
|
|
|
|
220
|
22 |
|
$addressLiteral = str_replace('[', '', $addressLiteral); |
221
|
22 |
|
$addressLiteral = $this->checkIPV4Tag($addressLiteral); |
222
|
|
|
|
223
|
22 |
|
if (false === $addressLiteral) { |
224
|
2 |
|
return $addressLiteral; |
225
|
|
|
} |
226
|
|
|
|
227
|
20 |
|
if (!$IPv6TAG) { |
228
|
6 |
|
$this->warnings[] = EmailValidator::RFC5322_DOMAINLITERAL; |
229
|
6 |
|
return $addressLiteral; |
230
|
|
|
} |
231
|
|
|
|
232
|
14 |
|
$this->warnings[] = EmailValidator::RFC5321_ADDRESSLITERAL; |
233
|
|
|
|
234
|
14 |
|
$this->checkIPV6Tag($addressLiteral); |
235
|
|
|
|
236
|
14 |
|
return $addressLiteral; |
237
|
|
|
} |
238
|
|
|
|
239
|
22 |
|
protected function checkIPV4Tag($addressLiteral) |
240
|
|
|
{ |
241
|
22 |
|
$matchesIP = array(); |
242
|
|
|
|
243
|
|
|
// Extract IPv4 part from the end of the address-literal (if there is one) |
244
|
22 |
|
if (preg_match( |
245
|
22 |
|
'/\\b(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$/', |
246
|
22 |
|
$addressLiteral, |
247
|
|
|
$matchesIP |
248
|
22 |
|
) > 0 |
249
|
22 |
|
) { |
250
|
4 |
|
$index = strrpos($addressLiteral, $matchesIP[0]); |
251
|
4 |
|
if ($index === 0) { |
252
|
2 |
|
$this->warnings[] = EmailValidator::RFC5321_ADDRESSLITERAL; |
253
|
2 |
|
return false; |
254
|
|
|
} |
255
|
|
|
// Convert IPv4 part to IPv6 format for further testing |
256
|
2 |
|
$addressLiteral = substr($addressLiteral, 0, $index) . '0:0'; |
257
|
2 |
|
} |
258
|
|
|
|
259
|
20 |
|
return $addressLiteral; |
260
|
|
|
} |
261
|
|
|
|
262
|
106 |
|
protected function checkDomainPartExceptions($prev) |
263
|
|
|
{ |
264
|
|
|
$invalidDomainTokens = array( |
265
|
106 |
|
EmailLexer::S_DQUOTE => true, |
266
|
106 |
|
EmailLexer::S_SEMICOLON => true, |
267
|
106 |
|
EmailLexer::S_GREATERTHAN => true, |
268
|
106 |
|
EmailLexer::S_LOWERTHAN => true, |
269
|
106 |
|
); |
270
|
|
|
|
271
|
106 |
|
if (isset($invalidDomainTokens[$this->lexer->token['type']])) { |
272
|
4 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_ATEXT'); |
273
|
|
|
} |
274
|
|
|
|
275
|
106 |
|
if ($this->lexer->token['type'] === EmailLexer::S_COMMA) { |
276
|
1 |
|
throw new \InvalidArgumentException('ERR_COMMA_IN_DOMAIN'); |
277
|
|
|
} |
278
|
|
|
|
279
|
106 |
|
if ($this->lexer->token['type'] === EmailLexer::S_AT) { |
280
|
2 |
|
throw new \InvalidArgumentException('ERR_CONSECUTIVEATS'); |
281
|
|
|
} |
282
|
|
|
|
283
|
105 |
|
if ($this->lexer->token['type'] === EmailLexer::S_OPENQBRACKET && $prev['type'] !== EmailLexer::S_AT) { |
284
|
1 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_ATEXT'); |
285
|
|
|
} |
286
|
|
|
|
287
|
105 |
View Code Duplication |
if ($this->lexer->token['type'] === EmailLexer::S_HYPHEN && $this->lexer->isNextToken(EmailLexer::S_DOT)) { |
|
|
|
|
288
|
1 |
|
throw new \InvalidArgumentException('ERR_DOMAINHYPHENEND'); |
289
|
|
|
} |
290
|
|
|
|
291
|
105 |
View Code Duplication |
if ($this->lexer->token['type'] === EmailLexer::S_BACKSLASH |
|
|
|
|
292
|
105 |
|
&& $this->lexer->isNextToken(EmailLexer::GENERIC)) { |
293
|
|
|
throw new \InvalidArgumentException('ERR_EXPECTING_ATEXT'); |
294
|
|
|
} |
295
|
105 |
|
} |
296
|
|
|
|
297
|
105 |
|
protected function hasBrackets($openBrackets) |
298
|
|
|
{ |
299
|
105 |
|
if ($this->lexer->token['type'] === EmailLexer::S_CLOSEBRACKET && !$openBrackets) { |
300
|
1 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_OPENBRACKET'); |
301
|
|
|
} |
302
|
|
|
|
303
|
105 |
|
if ($this->lexer->token['type'] !== EmailLexer::S_OPENBRACKET) { |
304
|
81 |
|
return false; |
305
|
|
|
} |
306
|
|
|
|
307
|
|
|
try { |
308
|
25 |
|
$this->lexer->find(EmailLexer::S_CLOSEBRACKET); |
309
|
25 |
|
} catch (\RuntimeException $e) { |
310
|
1 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_DOMLIT_CLOSE'); |
311
|
|
|
} |
312
|
|
|
|
313
|
24 |
|
return true; |
314
|
|
|
} |
315
|
|
|
|
316
|
103 |
|
protected function checkLabelLength($prev) |
317
|
|
|
{ |
318
|
103 |
|
if ($this->lexer->token['type'] === EmailLexer::S_DOT && |
319
|
103 |
|
$prev['type'] === EmailLexer::GENERIC && |
320
|
49 |
|
strlen($prev['value']) > 63 |
321
|
103 |
|
) { |
322
|
2 |
|
$this->warnings[] = EmailValidator::RFC5322_LABEL_TOOLONG; |
323
|
2 |
|
} |
324
|
103 |
|
} |
325
|
|
|
|
326
|
3 |
|
protected function parseDomainComments() |
327
|
|
|
{ |
328
|
3 |
|
$this->isUnclosedComment(); |
329
|
2 |
|
while (!$this->lexer->isNextToken(EmailLexer::S_CLOSEPARENTHESIS)) { |
330
|
2 |
|
$this->warnEscaping(); |
331
|
2 |
|
$this->lexer->moveNext(); |
332
|
2 |
|
} |
333
|
|
|
|
334
|
2 |
|
$this->lexer->moveNext(); |
335
|
2 |
|
if ($this->lexer->isNextToken(EmailLexer::S_DOT)) { |
336
|
1 |
|
throw new \InvalidArgumentException('ERR_EXPECTING_ATEXT'); |
337
|
|
|
} |
338
|
1 |
|
} |
339
|
|
|
} |
340
|
|
|
|
Duplicated code is one of the most pungent code smells. If you need to duplicate the same code in three or more different places, we strongly encourage you to look into extracting the code into a single class or operation.
You can also find more detailed suggestions in the “Code” section of your repository.