1
|
|
|
<?php |
2
|
|
|
namespace Thunder\Shortcode\Parser; |
3
|
|
|
|
4
|
|
|
use Thunder\Shortcode\Shortcode\ParsedShortcode; |
5
|
|
|
use Thunder\Shortcode\Shortcode\Shortcode; |
6
|
|
|
use Thunder\Shortcode\Syntax\CommonSyntax; |
7
|
|
|
use Thunder\Shortcode\Syntax\SyntaxInterface; |
8
|
|
|
use Thunder\Shortcode\Utility\RegexBuilderUtility; |
9
|
|
|
|
10
|
|
|
/** |
11
|
|
|
* @author Tomasz Kowalczyk <[email protected]> |
12
|
|
|
*/ |
13
|
|
|
final class RegularParser implements ParserInterface |
14
|
|
|
{ |
15
|
|
|
private $lexerRegex; |
16
|
|
|
private $nameRegex; |
17
|
|
|
private $tokens; |
18
|
|
|
private $tokensCount; |
19
|
|
|
private $position; |
20
|
|
|
/** @var int[] */ |
21
|
|
|
private $backtracks; |
22
|
|
|
private $lastBacktrack; |
23
|
|
|
private $tokenMap; |
24
|
|
|
|
25
|
|
|
const TOKEN_OPEN = 1; |
26
|
|
|
const TOKEN_CLOSE = 2; |
27
|
|
|
const TOKEN_MARKER = 3; |
28
|
|
|
const TOKEN_SEPARATOR = 4; |
29
|
|
|
const TOKEN_DELIMITER = 5; |
30
|
|
|
const TOKEN_STRING = 6; |
31
|
|
|
const TOKEN_WS = 7; |
32
|
|
|
|
33
|
14 |
|
public function __construct(SyntaxInterface $syntax = null) |
34
|
|
|
{ |
35
|
14 |
|
$this->lexerRegex = $this->prepareLexer($syntax ?: new CommonSyntax()); |
36
|
14 |
|
$this->nameRegex = '~^'.RegexBuilderUtility::buildNameRegex().'$~us'; |
37
|
14 |
|
} |
38
|
|
|
|
39
|
|
|
/** |
40
|
|
|
* @param string $text |
41
|
|
|
* |
42
|
|
|
* @return ParsedShortcode[] |
43
|
|
|
*/ |
44
|
57 |
|
public function parse($text) |
45
|
|
|
{ |
46
|
57 |
|
$this->tokens = $this->tokenize($text); |
47
|
57 |
|
$this->backtracks = array(); |
48
|
57 |
|
$this->lastBacktrack = 0; |
49
|
57 |
|
$this->position = 0; |
50
|
57 |
|
$this->tokensCount = \count($this->tokens); |
51
|
|
|
|
52
|
57 |
|
$shortcodes = array(); |
53
|
57 |
|
while($this->position < $this->tokensCount) { |
54
|
56 |
|
while($this->position < $this->tokensCount && false === $this->lookahead(self::TOKEN_OPEN)) { |
55
|
24 |
|
$this->position++; |
56
|
|
|
} |
57
|
56 |
|
$names = array(); |
58
|
56 |
|
$this->beginBacktrack(); |
59
|
56 |
|
$matches = $this->shortcode($names); |
60
|
56 |
|
if(\is_array($matches)) { |
61
|
48 |
|
foreach($matches as $shortcode) { |
62
|
48 |
|
$shortcodes[] = $shortcode; |
63
|
|
|
} |
64
|
|
|
} |
65
|
|
|
} |
66
|
|
|
|
67
|
57 |
|
return $shortcodes; |
68
|
|
|
} |
69
|
|
|
|
70
|
48 |
|
private function getObject($name, $parameters, $bbCode, $offset, $content, $text) |
71
|
|
|
{ |
72
|
48 |
|
return new ParsedShortcode(new Shortcode($name, $parameters, $content, $bbCode), $text, $offset); |
73
|
|
|
} |
74
|
|
|
|
75
|
|
|
/* --- RULES ----------------------------------------------------------- */ |
76
|
|
|
|
77
|
56 |
|
private function shortcode(array &$names) |
78
|
|
|
{ |
79
|
56 |
|
$name = null; |
|
|
|
|
80
|
56 |
|
$offset = null; |
|
|
|
|
81
|
|
|
|
82
|
56 |
|
if(!$this->match(self::TOKEN_OPEN, false)) { return false; } |
83
|
56 |
|
$offset = $this->tokens[$this->position - 1][2]; |
84
|
56 |
|
$this->match(self::TOKEN_WS, false); |
85
|
56 |
|
if('' === $name = $this->match(self::TOKEN_STRING, false)) { return false; } |
86
|
53 |
|
if($this->lookahead(self::TOKEN_STRING)) { return false; } |
87
|
53 |
|
if(1 !== preg_match($this->nameRegex, $name, $matches)) { return false; } |
88
|
52 |
|
$this->match(self::TOKEN_WS, false); |
89
|
|
|
// bbCode |
90
|
52 |
|
$bbCode = $this->match(self::TOKEN_SEPARATOR, true) ? $this->value() : null; |
91
|
52 |
|
if(false === $bbCode) { return false; } |
92
|
|
|
// parameters |
93
|
51 |
|
if(false === ($parameters = $this->parameters())) { return false; } |
94
|
|
|
|
95
|
|
|
// self-closing |
96
|
49 |
|
if($this->match(self::TOKEN_MARKER, true)) { |
97
|
16 |
|
if(!$this->match(self::TOKEN_CLOSE, false)) { return false; } |
98
|
|
|
|
99
|
15 |
|
return array($this->getObject($name, $parameters, $bbCode, $offset, null, $this->getBacktrack())); |
100
|
|
|
} |
101
|
|
|
|
102
|
|
|
// just-closed or with-content |
103
|
37 |
|
if(!$this->match(self::TOKEN_CLOSE, false)) { return false; } |
104
|
37 |
|
$this->beginBacktrack(); |
105
|
37 |
|
$names[] = $name; |
106
|
37 |
|
list($content, $shortcodes, $closingName) = $this->content($names); |
107
|
37 |
|
if(null !== $closingName && $closingName !== $name) { |
108
|
6 |
|
array_pop($names); |
109
|
6 |
|
array_pop($this->backtracks); |
110
|
6 |
|
array_pop($this->backtracks); |
111
|
|
|
|
112
|
6 |
|
return $closingName; |
113
|
|
|
} |
114
|
37 |
|
if(false === $content || $closingName !== $name) { |
115
|
24 |
|
$this->backtrack(false); |
116
|
24 |
|
$text = $this->backtrack(false); |
117
|
|
|
|
118
|
24 |
|
return array_merge(array($this->getObject($name, $parameters, $bbCode, $offset, null, $text)), $shortcodes); |
119
|
|
|
} |
120
|
19 |
|
$content = $this->getBacktrack(); |
121
|
19 |
|
if(!$this->close($names)) { return false; } |
122
|
|
|
|
123
|
19 |
|
return array($this->getObject($name, $parameters, $bbCode, $offset, $content, $this->getBacktrack())); |
124
|
|
|
} |
125
|
|
|
|
126
|
37 |
|
private function content(array &$names) |
127
|
|
|
{ |
128
|
37 |
|
$content = ''; |
129
|
37 |
|
$shortcodes = array(); |
130
|
37 |
|
$closingName = null; |
131
|
|
|
|
132
|
37 |
|
while($this->position < $this->tokensCount) { |
133
|
28 |
View Code Duplication |
while($this->position < $this->tokensCount && false === $this->lookahead(self::TOKEN_OPEN)) { |
|
|
|
|
134
|
24 |
|
$content .= $this->match(null, true); |
135
|
|
|
} |
136
|
|
|
|
137
|
28 |
|
$this->beginBacktrack(); |
138
|
28 |
|
$matchedShortcodes = $this->shortcode($names); |
139
|
28 |
|
if(\is_string($matchedShortcodes)) { |
140
|
6 |
|
$closingName = $matchedShortcodes; |
141
|
6 |
|
break; |
142
|
|
|
} |
143
|
28 |
|
if(\is_array($matchedShortcodes)) { |
144
|
15 |
|
foreach($matchedShortcodes as $matchedShortcode) { |
145
|
15 |
|
$shortcodes[] = $matchedShortcode; |
146
|
|
|
} |
147
|
15 |
|
continue; |
148
|
|
|
} |
149
|
22 |
|
$this->backtrack(); |
150
|
|
|
|
151
|
22 |
|
$this->beginBacktrack(); |
152
|
22 |
|
if(false !== ($closingName = $this->close($names))) { |
153
|
19 |
|
if(null === $content) { $content = ''; } |
154
|
19 |
|
$this->backtrack(); |
155
|
19 |
|
$shortcodes = array(); |
156
|
19 |
|
break; |
157
|
|
|
} |
158
|
8 |
|
$closingName = null; |
159
|
8 |
|
$this->backtrack(); |
160
|
|
|
|
161
|
8 |
|
$content .= $this->match(null, false); |
162
|
|
|
} |
163
|
|
|
|
164
|
37 |
|
return array($this->position < $this->tokensCount ? $content : false, $shortcodes, $closingName); |
165
|
|
|
} |
166
|
|
|
|
167
|
22 |
|
private function close(array &$names) |
168
|
|
|
{ |
169
|
22 |
|
if(!$this->match(self::TOKEN_OPEN, true)) { return false; } |
170
|
20 |
|
if(!$this->match(self::TOKEN_MARKER, true)) { return false; } |
171
|
20 |
|
if(!$closingName = $this->match(self::TOKEN_STRING, true)) { return false; } |
172
|
20 |
|
if(!$this->match(self::TOKEN_CLOSE, false)) { return false; } |
173
|
|
|
|
174
|
20 |
|
return \in_array($closingName, $names, true) ? $closingName : false; |
175
|
|
|
} |
176
|
|
|
|
177
|
51 |
|
private function parameters() |
178
|
|
|
{ |
179
|
51 |
|
$parameters = array(); |
180
|
|
|
|
181
|
51 |
|
while(true) { |
182
|
51 |
|
$name = null; |
|
|
|
|
183
|
|
|
|
184
|
51 |
|
$this->match(self::TOKEN_WS, false); |
185
|
51 |
|
if($this->lookahead(self::TOKEN_MARKER) || $this->lookahead(self::TOKEN_CLOSE)) { break; } |
186
|
27 |
|
if(!$name = $this->match(self::TOKEN_STRING, true)) { return false; } |
187
|
26 |
|
if(!$this->match(self::TOKEN_SEPARATOR, true)) { $parameters[$name] = null; continue; } |
188
|
25 |
|
if(false === ($value = $this->value())) { return false; } |
189
|
24 |
|
$this->match(self::TOKEN_WS, false); |
190
|
|
|
|
191
|
24 |
|
$parameters[$name] = $value; |
192
|
|
|
} |
193
|
|
|
|
194
|
49 |
|
return $parameters; |
195
|
|
|
} |
196
|
|
|
|
197
|
27 |
|
private function value() |
198
|
|
|
{ |
199
|
27 |
|
$value = ''; |
200
|
|
|
|
201
|
27 |
|
if($this->match(self::TOKEN_DELIMITER, false)) { |
202
|
19 |
View Code Duplication |
while($this->position < $this->tokensCount && false === $this->lookahead(self::TOKEN_DELIMITER)) { |
|
|
|
|
203
|
19 |
|
$value .= $this->match(null, false); |
204
|
|
|
} |
205
|
|
|
|
206
|
19 |
|
return $this->match(self::TOKEN_DELIMITER, false) ? $value : false; |
207
|
|
|
} |
208
|
|
|
|
209
|
14 |
|
if($tmp = $this->match(self::TOKEN_STRING, false)) { |
210
|
13 |
|
$value .= $tmp; |
211
|
13 |
|
while($tmp = $this->match(self::TOKEN_STRING, false)) { |
212
|
2 |
|
$value .= $tmp; |
213
|
|
|
} |
214
|
|
|
|
215
|
13 |
|
return $value; |
216
|
|
|
} |
217
|
|
|
|
218
|
1 |
|
return false; |
219
|
|
|
} |
220
|
|
|
|
221
|
|
|
/* --- PARSER ---------------------------------------------------------- */ |
222
|
|
|
|
223
|
56 |
|
private function beginBacktrack() |
224
|
|
|
{ |
225
|
56 |
|
$this->backtracks[] = $this->position; |
226
|
56 |
|
$this->lastBacktrack = $this->position; |
227
|
56 |
|
} |
228
|
|
|
|
229
|
30 |
|
private function getBacktrack() |
230
|
|
|
{ |
231
|
30 |
|
$position = array_pop($this->backtracks); |
232
|
30 |
|
$backtrack = ''; |
233
|
30 |
View Code Duplication |
for($i = $position; $i < $this->position; $i++) { |
|
|
|
|
234
|
30 |
|
$backtrack .= $this->tokens[$i][1]; |
235
|
|
|
} |
236
|
|
|
|
237
|
30 |
|
return $backtrack; |
238
|
|
|
} |
239
|
|
|
|
240
|
37 |
|
private function backtrack($modifyPosition = true) |
241
|
|
|
{ |
242
|
37 |
|
$position = array_pop($this->backtracks); |
243
|
37 |
|
if($modifyPosition) { |
244
|
22 |
|
$this->position = $position; |
245
|
|
|
} |
246
|
|
|
|
247
|
37 |
|
$backtrack = ''; |
248
|
37 |
View Code Duplication |
for($i = $position; $i < $this->lastBacktrack; $i++) { |
|
|
|
|
249
|
24 |
|
$backtrack .= $this->tokens[$i][1]; |
250
|
|
|
} |
251
|
37 |
|
$this->lastBacktrack = $position; |
252
|
|
|
|
253
|
37 |
|
return $backtrack; |
254
|
|
|
} |
255
|
|
|
|
256
|
56 |
|
private function lookahead($type) |
257
|
|
|
{ |
258
|
56 |
|
return $this->position < $this->tokensCount && (empty($type) || $this->tokens[$this->position][0] === $type); |
259
|
|
|
} |
260
|
|
|
|
261
|
56 |
|
private function match($type, $ws) |
262
|
|
|
{ |
263
|
56 |
|
if($this->position >= $this->tokensCount) { |
264
|
20 |
|
return ''; |
265
|
|
|
} |
266
|
|
|
|
267
|
56 |
|
$token = $this->tokens[$this->position]; |
268
|
56 |
|
if(!empty($type) && $token[0] !== $type) { |
269
|
56 |
|
return ''; |
270
|
|
|
} |
271
|
|
|
|
272
|
56 |
|
$this->position++; |
273
|
56 |
|
if($ws && $this->position < $this->tokensCount && $this->tokens[$this->position][0] === self::TOKEN_WS) { |
274
|
17 |
|
$this->position++; |
275
|
|
|
} |
276
|
|
|
|
277
|
56 |
|
return $token[1]; |
278
|
|
|
} |
279
|
|
|
|
280
|
|
|
/* --- LEXER ----------------------------------------------------------- */ |
281
|
|
|
|
282
|
57 |
|
private function tokenize($text) |
283
|
|
|
{ |
284
|
57 |
|
preg_match_all($this->lexerRegex, $text, $matches, PREG_OFFSET_CAPTURE); |
285
|
57 |
|
if(preg_last_error() !== PREG_NO_ERROR) { |
286
|
|
|
throw new \RuntimeException(sprintf('PCRE failure `%s`.', preg_last_error())); |
287
|
|
|
} |
288
|
|
|
|
289
|
57 |
|
$tokens = array(); |
290
|
57 |
|
$position = 0; |
291
|
57 |
|
foreach($matches[0] as $match) { |
292
|
56 |
|
$type = isset($this->tokenMap[$match[0]]) |
293
|
56 |
|
? $this->tokenMap[$match[0]] |
294
|
56 |
|
: (ctype_space($match[0]) ? self::TOKEN_WS : self::TOKEN_STRING); |
295
|
56 |
|
$tokens[] = array($type, $match[0], $position); |
296
|
56 |
|
$position += mb_strlen($match[0], 'utf-8'); // FIXME match[1] |
|
|
|
|
297
|
|
|
} |
298
|
|
|
|
299
|
57 |
|
return $tokens; |
300
|
|
|
} |
301
|
|
|
|
302
|
14 |
|
private function prepareLexer(SyntaxInterface $syntax) |
303
|
|
|
{ |
304
|
14 |
|
$this->tokenMap = array( |
305
|
14 |
|
$syntax->getOpeningTag() => self::TOKEN_OPEN, |
306
|
14 |
|
$syntax->getClosingTag() => self::TOKEN_CLOSE, |
307
|
14 |
|
$syntax->getClosingTagMarker() => self::TOKEN_MARKER, |
308
|
14 |
|
$syntax->getParameterValueSeparator() => self::TOKEN_SEPARATOR, |
309
|
14 |
|
$syntax->getParameterValueDelimiter() => self::TOKEN_DELIMITER, |
310
|
|
|
); |
311
|
|
|
|
312
|
14 |
|
$quote = function($text) { |
313
|
14 |
|
return preg_replace('/(.)/us', '\\\\$0', $text); |
314
|
14 |
|
}; |
315
|
14 |
|
$symbols = array_map($quote, array_keys($this->tokenMap)); |
316
|
|
|
|
317
|
14 |
|
return '~('.implode('|', $symbols).'|\s+|\\\\.|[\w-]+|.)~us'; |
318
|
|
|
} |
319
|
|
|
} |
320
|
|
|
|
This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently.
Both the
$myVar
assignment in line 1 and the$higher
assignment in line 2 are dead. The first because$myVar
is never used and the second because$higher
is always overwritten for every possible time line.