1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
declare(strict_types=1); |
4
|
|
|
|
5
|
|
|
/* |
6
|
|
|
* This file is part of the league/commonmark package. |
7
|
|
|
* |
8
|
|
|
* (c) Colin O'Dell <[email protected]> |
9
|
|
|
* |
10
|
|
|
* Original code based on the CommonMark JS reference parser (https://bitly.com/commonmark-js) |
11
|
|
|
* - (c) John MacFarlane |
12
|
|
|
* |
13
|
|
|
* For the full copyright and license information, please view the LICENSE |
14
|
|
|
* file that was distributed with this source code. |
15
|
|
|
*/ |
16
|
|
|
|
17
|
|
|
namespace League\CommonMark\Parser; |
18
|
|
|
|
19
|
|
|
use League\CommonMark\Environment\EnvironmentInterface; |
20
|
|
|
use League\CommonMark\Event\DocumentParsedEvent; |
21
|
|
|
use League\CommonMark\Event\DocumentPreParsedEvent; |
22
|
|
|
use League\CommonMark\Input\MarkdownInput; |
23
|
|
|
use League\CommonMark\Node\Block\Document; |
24
|
|
|
use League\CommonMark\Node\Block\Paragraph; |
25
|
|
|
use League\CommonMark\Parser\Block\BlockContinueParserInterface; |
26
|
|
|
use League\CommonMark\Parser\Block\BlockStart; |
27
|
|
|
use League\CommonMark\Parser\Block\BlockStartParserInterface; |
28
|
|
|
use League\CommonMark\Parser\Block\DocumentBlockParser; |
29
|
|
|
use League\CommonMark\Parser\Block\ParagraphParser; |
30
|
|
|
use League\CommonMark\Reference\ReferenceInterface; |
31
|
|
|
use League\CommonMark\Reference\ReferenceMap; |
32
|
|
|
use League\CommonMark\Util\RegexHelper; |
33
|
|
|
|
34
|
|
|
final class MarkdownParser implements MarkdownParserInterface |
35
|
|
|
{ |
36
|
|
|
/** |
37
|
|
|
* @var EnvironmentInterface |
38
|
|
|
* |
39
|
|
|
* @psalm-readonly |
40
|
|
|
*/ |
41
|
|
|
private $environment; |
42
|
|
|
|
43
|
|
|
/** |
44
|
|
|
* @var int|float |
45
|
|
|
* |
46
|
|
|
* @psalm-readonly |
47
|
|
|
*/ |
48
|
|
|
private $maxNestingLevel; |
49
|
|
|
|
50
|
|
|
/** |
51
|
|
|
* @var ReferenceMap |
52
|
|
|
* |
53
|
|
|
* @psalm-readonly-allow-private-mutation |
54
|
|
|
*/ |
55
|
|
|
private $referenceMap; |
56
|
|
|
|
57
|
|
|
/** |
58
|
|
|
* @var int |
59
|
|
|
* |
60
|
|
|
* @psalm-readonly-allow-private-mutation |
61
|
|
|
*/ |
62
|
|
|
private $lineNumber = 0; |
63
|
|
|
|
64
|
|
|
/** |
65
|
|
|
* @var Cursor |
66
|
|
|
* |
67
|
|
|
* @psalm-readonly-allow-private-mutation |
68
|
|
|
*/ |
69
|
|
|
private $cursor; |
70
|
|
|
|
71
|
|
|
/** |
72
|
|
|
* @var array<int, BlockContinueParserInterface> |
73
|
|
|
* |
74
|
|
|
* @psalm-readonly-allow-private-mutation |
75
|
|
|
*/ |
76
|
|
|
private $allBlockParsers = []; |
77
|
|
|
|
78
|
|
|
/** |
79
|
|
|
* @var array<int, BlockContinueParserInterface> |
80
|
|
|
* |
81
|
|
|
* @psalm-readonly-allow-private-mutation |
82
|
|
|
*/ |
83
|
|
|
private $activeBlockParsers = []; |
84
|
|
|
|
85
|
2511 |
|
public function __construct(EnvironmentInterface $environment) |
86
|
|
|
{ |
87
|
2511 |
|
$this->environment = $environment; |
88
|
2511 |
|
$this->maxNestingLevel = $environment->getConfig('max_nesting_level', \INF); |
89
|
2511 |
|
} |
90
|
|
|
|
91
|
2502 |
|
private function initialize(): void |
92
|
|
|
{ |
93
|
2502 |
|
$this->referenceMap = new ReferenceMap(); |
94
|
2502 |
|
$this->lineNumber = 0; |
95
|
2502 |
|
$this->allBlockParsers = []; |
96
|
2502 |
|
$this->activeBlockParsers = []; |
97
|
2502 |
|
} |
98
|
|
|
|
99
|
|
|
/** |
100
|
|
|
* @throws \RuntimeException |
101
|
|
|
*/ |
102
|
2502 |
|
public function parse(string $input): Document |
103
|
|
|
{ |
104
|
2502 |
|
$this->initialize(); |
105
|
|
|
|
106
|
2502 |
|
$documentParser = new DocumentBlockParser($this->referenceMap); |
107
|
2502 |
|
$this->activateBlockParser($documentParser); |
108
|
|
|
|
109
|
2502 |
|
$preParsedEvent = new DocumentPreParsedEvent($documentParser->getBlock(), new MarkdownInput($input)); |
110
|
2496 |
|
$this->environment->dispatch($preParsedEvent); |
111
|
2496 |
|
$markdown = $preParsedEvent->getMarkdown(); |
112
|
|
|
|
113
|
2496 |
|
foreach ($markdown->getLines() as $line) { |
114
|
2496 |
|
++$this->lineNumber; |
115
|
2496 |
|
$this->incorporateLine($line); |
116
|
|
|
} |
117
|
|
|
|
118
|
2496 |
|
$this->finalizeBlocks($this->getActiveBlockParsers(), $this->lineNumber); |
119
|
2496 |
|
$this->processInlines(); |
120
|
|
|
|
121
|
2496 |
|
$this->environment->dispatch(new DocumentParsedEvent($documentParser->getBlock())); |
122
|
|
|
|
123
|
2493 |
|
return $documentParser->getBlock(); |
124
|
|
|
} |
125
|
|
|
|
126
|
|
|
/** |
127
|
|
|
* Analyze a line of text and update the document appropriately. We parse markdown text by calling this on each |
128
|
|
|
* line of input, then finalizing the document. |
129
|
|
|
*/ |
130
|
2496 |
|
private function incorporateLine(string $line): void |
131
|
|
|
{ |
132
|
2496 |
|
$this->cursor = new Cursor($line); |
133
|
|
|
|
134
|
2496 |
|
$matches = 1; |
135
|
2496 |
|
foreach ($this->getActiveBlockParsers(1) as $blockParser) { |
136
|
|
|
\assert($blockParser instanceof BlockContinueParserInterface); |
137
|
1302 |
|
$blockContinue = $blockParser->tryContinue(clone $this->cursor, $this->getActiveBlockParser()); |
138
|
1302 |
|
if ($blockContinue === null) { |
139
|
828 |
|
break; |
140
|
|
|
} |
141
|
|
|
|
142
|
900 |
|
if ($blockContinue->isFinalize()) { |
143
|
87 |
|
$this->finalizeAndClose($blockParser, $this->lineNumber); |
144
|
|
|
|
145
|
87 |
|
return; |
146
|
|
|
} |
147
|
|
|
|
148
|
894 |
|
if (($state = $blockContinue->getCursorState()) !== null) { |
149
|
894 |
|
$this->cursor->restoreState($state); |
|
|
|
|
150
|
|
|
} |
151
|
|
|
|
152
|
894 |
|
$matches++; |
153
|
|
|
} |
154
|
|
|
|
155
|
2496 |
|
$unmatchedBlockParsers = $this->getActiveBlockParsers($matches); |
156
|
2496 |
|
$lastMatchedBlockParser = $this->getActiveBlockParsers()[$matches - 1]; |
157
|
2496 |
|
$blockParser = $lastMatchedBlockParser; |
158
|
2496 |
|
$allClosed = empty($unmatchedBlockParsers); |
159
|
|
|
|
160
|
|
|
// Unless last matched container is a code block, try new container starts |
161
|
2496 |
|
$tryBlockStarts = $blockParser->getBlock() instanceof Paragraph || $blockParser->isContainer(); |
162
|
2496 |
|
while ($tryBlockStarts) { |
163
|
|
|
// this is a little performance optimization |
164
|
2496 |
|
if ($this->cursor->isBlank()) { |
165
|
630 |
|
$this->cursor->advanceToEnd(); |
166
|
630 |
|
break; |
167
|
|
|
} |
168
|
|
|
|
169
|
2496 |
|
if (! $this->cursor->isIndented() && RegexHelper::isLetter($this->cursor->getNextNonSpaceCharacter())) { |
170
|
1023 |
|
$this->cursor->advanceToNextNonSpaceOrTab(); |
171
|
1023 |
|
break; |
172
|
|
|
} |
173
|
|
|
|
174
|
2112 |
|
if ($blockParser->getBlock()->getDepth() >= $this->maxNestingLevel) { |
175
|
3 |
|
break; |
176
|
|
|
} |
177
|
|
|
|
178
|
2112 |
|
$blockStart = $this->findBlockStart($blockParser); |
179
|
2112 |
|
if ($blockStart === null) { |
180
|
1299 |
|
$this->cursor->advanceToNextNonSpaceOrTab(); |
181
|
1299 |
|
break; |
182
|
|
|
} |
183
|
|
|
|
184
|
954 |
|
if (($state = $blockStart->getCursorState()) !== null) { |
185
|
954 |
|
$this->cursor->restoreState($state); |
|
|
|
|
186
|
|
|
} |
187
|
|
|
|
188
|
954 |
|
if (! $allClosed) { |
189
|
210 |
|
$this->finalizeBlocks($unmatchedBlockParsers, $this->lineNumber - 1); |
190
|
210 |
|
$allClosed = true; |
191
|
|
|
} |
192
|
|
|
|
193
|
954 |
|
if ($blockStart->isReplaceActiveBlockParser()) { |
194
|
132 |
|
$this->prepareActiveBlockParserForReplacement(); |
195
|
|
|
} |
196
|
|
|
|
197
|
954 |
|
foreach ($blockStart->getBlockParsers() as $newBlockParser) { |
198
|
954 |
|
$blockParser = $this->addChild($newBlockParser); |
199
|
954 |
|
$tryBlockStarts = $newBlockParser->isContainer(); |
200
|
|
|
} |
201
|
|
|
} |
202
|
|
|
|
203
|
|
|
// What remains ath the offset is a text line. Add the text to the appropriate block. |
204
|
|
|
|
205
|
|
|
// First check for a lazy paragraph continuation: |
206
|
2496 |
|
if (! $allClosed && ! $this->cursor->isBlank() && $this->getActiveBlockParser()->canHaveLazyContinuationLines()) { |
207
|
48 |
|
$this->getActiveBlockParser()->addLine($this->cursor->getRemainder()); |
208
|
|
|
} else { |
209
|
|
|
// finalize any blocks not matched |
210
|
2496 |
|
if (! $allClosed) { |
211
|
669 |
|
$this->finalizeBlocks($unmatchedBlockParsers, $this->lineNumber); |
212
|
|
|
} |
213
|
|
|
|
214
|
2496 |
|
if (! $blockParser->isContainer()) { |
215
|
981 |
|
$this->getActiveBlockParser()->addLine($this->cursor->getRemainder()); |
216
|
2181 |
|
} elseif (! $this->cursor->isBlank()) { |
217
|
2133 |
|
$this->addChild(new ParagraphParser()); |
218
|
2133 |
|
$this->getActiveBlockParser()->addLine($this->cursor->getRemainder()); |
219
|
|
|
} |
220
|
|
|
} |
221
|
2496 |
|
} |
222
|
|
|
|
223
|
2112 |
|
private function findBlockStart(BlockContinueParserInterface $lastMatchedBlockParser): ?BlockStart |
224
|
|
|
{ |
225
|
2112 |
|
$matchedBlockParser = new MarkdownParserState($this->getActiveBlockParser(), $lastMatchedBlockParser); |
226
|
|
|
|
227
|
2112 |
|
foreach ($this->environment->getBlockStartParsers() as $blockStartParser) { |
228
|
|
|
\assert($blockStartParser instanceof BlockStartParserInterface); |
229
|
2109 |
|
if (($result = $blockStartParser->tryStart(clone $this->cursor, $matchedBlockParser)) !== null) { |
230
|
954 |
|
return $result; |
231
|
|
|
} |
232
|
|
|
} |
233
|
|
|
|
234
|
1299 |
|
return null; |
235
|
|
|
} |
236
|
|
|
|
237
|
|
|
/** |
238
|
|
|
* @param array<int, BlockContinueParserInterface> $blockParsers |
239
|
|
|
*/ |
240
|
2496 |
|
private function finalizeBlocks(array $blockParsers, int $endLineNumber): void |
241
|
|
|
{ |
242
|
2496 |
|
foreach (\array_reverse($blockParsers) as $blockParser) { |
243
|
2496 |
|
$this->finalizeAndClose($blockParser, $endLineNumber); |
244
|
|
|
} |
245
|
2496 |
|
} |
246
|
|
|
|
247
|
|
|
/** |
248
|
|
|
* Finalize a block. Close it and do any necessary postprocessing, e.g. creating string_content from strings, |
249
|
|
|
* setting the 'tight' or 'loose' status of a list, and parsing the beginnings of paragraphs for reference |
250
|
|
|
* definitions. |
251
|
|
|
*/ |
252
|
2496 |
|
private function finalizeAndClose(BlockContinueParserInterface $blockParser, int $endLineNumber): void |
253
|
|
|
{ |
254
|
2496 |
|
if ($this->getActiveBlockParser() === $blockParser) { |
255
|
2496 |
|
$this->deactivateBlockParser(); |
256
|
|
|
} |
257
|
|
|
|
258
|
2496 |
|
if ($blockParser instanceof ParagraphParser) { |
259
|
2031 |
|
$this->updateReferenceMap($blockParser->getReferences()); |
260
|
|
|
} |
261
|
|
|
|
262
|
2496 |
|
$blockParser->getBlock()->setEndLine($endLineNumber); |
263
|
2496 |
|
$blockParser->closeBlock(); |
264
|
2496 |
|
} |
265
|
|
|
|
266
|
|
|
/** |
267
|
|
|
* Walk through a block & children recursively, parsing string content into inline content where appropriate. |
268
|
|
|
*/ |
269
|
2496 |
|
private function processInlines(): void |
270
|
|
|
{ |
271
|
2496 |
|
$p = new InlineParserEngine($this->environment, $this->referenceMap); |
272
|
|
|
|
273
|
2496 |
|
foreach ($this->allBlockParsers as $blockParser) { |
274
|
|
|
\assert($blockParser instanceof BlockContinueParserInterface); |
275
|
2496 |
|
$blockParser->parseInlines($p); |
276
|
|
|
} |
277
|
2496 |
|
} |
278
|
|
|
|
279
|
|
|
/** |
280
|
|
|
* Add block of type tag as a child of the tip. If the tip can't accept children, close and finalize it and try |
281
|
|
|
* its parent, and so on til we find a block that can accept children. |
282
|
|
|
*/ |
283
|
2496 |
|
private function addChild(BlockContinueParserInterface $blockParser): BlockContinueParserInterface |
284
|
|
|
{ |
285
|
2496 |
|
$blockParser->getBlock()->setStartLine($this->lineNumber); |
286
|
|
|
|
287
|
2496 |
|
while (! $this->getActiveBlockParser()->canContain($blockParser->getBlock())) { |
288
|
123 |
|
$this->finalizeAndClose($this->getActiveBlockParser(), $this->lineNumber - 1); |
289
|
|
|
} |
290
|
|
|
|
291
|
2496 |
|
$this->getActiveBlockParser()->getBlock()->appendChild($blockParser->getBlock()); |
292
|
2496 |
|
$this->activateBlockParser($blockParser); |
293
|
|
|
|
294
|
2496 |
|
return $blockParser; |
295
|
|
|
} |
296
|
|
|
|
297
|
2502 |
|
private function activateBlockParser(BlockContinueParserInterface $blockParser): void |
298
|
|
|
{ |
299
|
2502 |
|
$this->activeBlockParsers[] = $blockParser; |
300
|
2502 |
|
$this->allBlockParsers[] = $blockParser; |
301
|
2502 |
|
} |
302
|
|
|
|
303
|
2496 |
|
private function deactivateBlockParser(): BlockContinueParserInterface |
304
|
|
|
{ |
305
|
2496 |
|
$popped = \array_pop($this->activeBlockParsers); |
306
|
2496 |
|
if ($popped === null) { |
307
|
|
|
throw new \RuntimeException('The last block parser should not be deactivated'); |
308
|
|
|
} |
309
|
|
|
|
310
|
2496 |
|
return $popped; |
311
|
|
|
} |
312
|
|
|
|
313
|
132 |
|
private function prepareActiveBlockParserForReplacement(): void |
314
|
|
|
{ |
315
|
132 |
|
$old = $this->deactivateBlockParser(); |
316
|
132 |
|
$key = \array_search($old, $this->allBlockParsers, true); |
317
|
132 |
|
unset($this->allBlockParsers[$key]); |
318
|
|
|
|
319
|
132 |
|
if ($old instanceof ParagraphParser) { |
320
|
132 |
|
$this->updateReferenceMap($old->getReferences()); |
321
|
|
|
} |
322
|
|
|
|
323
|
132 |
|
$old->getBlock()->detach(); |
324
|
132 |
|
} |
325
|
|
|
|
326
|
|
|
/** |
327
|
|
|
* @param ReferenceInterface[] $references |
328
|
|
|
*/ |
329
|
2133 |
|
private function updateReferenceMap(iterable $references): void |
330
|
|
|
{ |
331
|
2133 |
|
foreach ($references as $reference) { |
332
|
237 |
|
if (! $this->referenceMap->contains($reference->getLabel())) { |
333
|
237 |
|
$this->referenceMap->add($reference); |
334
|
|
|
} |
335
|
|
|
} |
336
|
2133 |
|
} |
337
|
|
|
|
338
|
|
|
/** |
339
|
|
|
* @return array<int, BlockContinueParserInterface> |
|
|
|
|
340
|
|
|
*/ |
341
|
2496 |
|
private function getActiveBlockParsers(?int $offset = 0): array |
342
|
|
|
{ |
343
|
2496 |
|
if (\is_int($offset)) { |
344
|
2496 |
|
return \array_slice($this->activeBlockParsers, $offset); |
345
|
|
|
} |
346
|
|
|
|
347
|
|
|
return $this->activeBlockParsers; |
348
|
|
|
} |
349
|
|
|
|
350
|
2496 |
|
public function getActiveBlockParser(): BlockContinueParserInterface |
351
|
|
|
{ |
352
|
2496 |
|
$active = \end($this->activeBlockParsers); |
353
|
2496 |
|
if ($active === false) { |
354
|
|
|
throw new \RuntimeException('No active block parsers are available'); |
355
|
|
|
} |
356
|
|
|
|
357
|
2496 |
|
return $active; |
358
|
|
|
} |
359
|
|
|
} |
360
|
|
|
|
PHP Analyzer performs a side-effects analysis of your code. A side-effect is basically anything that might be visible after the scope of the method is left.
Let’s take a look at an example:
If we look at the
getEmail()
method, we can see that it has no side-effect. Whether you call this method or not, no future calls to other methods are affected by this. As such code as the following is useless:On the hand, if we look at the
setEmail()
, this method _has_ side-effects. In the following case, we could not remove the method call: