1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
/* |
4
|
|
|
* This file is part of the league/commonmark package. |
5
|
|
|
* |
6
|
|
|
* (c) Colin O'Dell <[email protected]> |
7
|
|
|
* |
8
|
|
|
* Original code based on the CommonMark JS reference parser (https://bitly.com/commonmark-js) |
9
|
|
|
* - (c) John MacFarlane |
10
|
|
|
* |
11
|
|
|
* For the full copyright and license information, please view the LICENSE |
12
|
|
|
* file that was distributed with this source code. |
13
|
|
|
*/ |
14
|
|
|
|
15
|
|
|
namespace League\CommonMark; |
16
|
|
|
|
17
|
|
|
use League\CommonMark\Block\Element\AbstractBlock; |
18
|
|
|
use League\CommonMark\Block\Element\AbstractStringContainerBlock; |
19
|
|
|
use League\CommonMark\Block\Element\Document; |
20
|
|
|
use League\CommonMark\Block\Element\Paragraph; |
21
|
|
|
use League\CommonMark\Block\Element\StringContainerInterface; |
22
|
|
|
use League\CommonMark\Event\DocumentParsedEvent; |
23
|
|
|
use League\CommonMark\Event\DocumentPreParsedEvent; |
24
|
|
|
use League\CommonMark\Exception\UnexpectedEncodingException; |
25
|
|
|
|
26
|
|
|
final class DocParser implements DocParserInterface |
27
|
|
|
{ |
28
|
|
|
/** |
29
|
|
|
* @var EnvironmentInterface |
30
|
|
|
*/ |
31
|
|
|
private $environment; |
32
|
|
|
|
33
|
|
|
/** |
34
|
|
|
* @var InlineParserEngine |
35
|
|
|
*/ |
36
|
|
|
private $inlineParserEngine; |
37
|
|
|
|
38
|
|
|
/** |
39
|
|
|
* @var int|float |
40
|
|
|
*/ |
41
|
|
|
private $maxNestingLevel; |
42
|
|
|
|
43
|
|
|
/** |
44
|
|
|
* @param EnvironmentInterface $environment |
45
|
|
|
*/ |
46
|
2472 |
|
public function __construct(EnvironmentInterface $environment) |
47
|
|
|
{ |
48
|
2472 |
|
$this->environment = $environment; |
49
|
2472 |
|
$this->inlineParserEngine = new InlineParserEngine($environment); |
50
|
2472 |
|
$this->maxNestingLevel = $environment->getConfig('max_nesting_level', \INF); |
51
|
2472 |
|
} |
52
|
|
|
|
53
|
|
|
/** |
54
|
|
|
* @param string $input |
55
|
|
|
* |
56
|
|
|
* @return string[] |
57
|
|
|
*/ |
58
|
2457 |
|
private function preProcessInput(string $input): array |
59
|
|
|
{ |
60
|
|
|
/** @var string[] $lines */ |
61
|
2457 |
|
$lines = \preg_split('/\r\n|\n|\r/', $input); |
62
|
|
|
|
63
|
|
|
// Remove any newline which appears at the very end of the string. |
64
|
|
|
// We've already split the document by newlines, so we can simply drop |
65
|
|
|
// any empty element which appears on the end. |
66
|
2457 |
|
if (\end($lines) === '') { |
67
|
2187 |
|
\array_pop($lines); |
68
|
|
|
} |
69
|
|
|
|
70
|
2457 |
|
return $lines; |
71
|
|
|
} |
72
|
|
|
|
73
|
|
|
/** |
74
|
|
|
* @param string $input |
75
|
|
|
* |
76
|
|
|
* @throws \RuntimeException |
77
|
|
|
* |
78
|
|
|
* @return Document |
79
|
|
|
*/ |
80
|
2463 |
|
public function parse(string $input): Document |
81
|
|
|
{ |
82
|
2463 |
|
$document = new Document(); |
83
|
|
|
|
84
|
2463 |
|
$this->environment->dispatch(new DocumentPreParsedEvent($document)); |
85
|
|
|
|
86
|
2463 |
|
$context = new Context($document, $this->environment); |
87
|
|
|
|
88
|
2463 |
|
$this->assertValidUTF8($input); |
89
|
2457 |
|
$lines = $this->preProcessInput($input); |
90
|
2457 |
|
foreach ($lines as $line) { |
91
|
2457 |
|
$context->setNextLine($line); |
92
|
2457 |
|
$this->incorporateLine($context); |
93
|
|
|
} |
94
|
|
|
|
95
|
2457 |
|
$lineCount = \count($lines); |
96
|
2457 |
|
while ($tip = $context->getTip()) { |
97
|
2457 |
|
$tip->finalize($context, $lineCount); |
98
|
|
|
} |
99
|
|
|
|
100
|
2457 |
|
$this->processInlines($context); |
101
|
|
|
|
102
|
2457 |
|
$this->environment->dispatch(new DocumentParsedEvent($document)); |
103
|
|
|
|
104
|
2454 |
|
return $document; |
105
|
|
|
} |
106
|
|
|
|
107
|
2457 |
|
private function incorporateLine(ContextInterface $context) |
108
|
|
|
{ |
109
|
2457 |
|
$context->getBlockCloser()->resetTip(); |
110
|
2457 |
|
$context->setBlocksParsed(false); |
111
|
|
|
|
112
|
2457 |
|
$cursor = new Cursor($context->getLine()); |
113
|
|
|
|
114
|
2457 |
|
$this->resetContainer($context, $cursor); |
115
|
2457 |
|
$context->getBlockCloser()->setLastMatchedContainer($context->getContainer()); |
116
|
|
|
|
117
|
2457 |
|
$this->parseBlocks($context, $cursor); |
118
|
|
|
|
119
|
|
|
// What remains at the offset is a text line. Add the text to the appropriate container. |
120
|
|
|
// First check for a lazy paragraph continuation: |
121
|
2457 |
|
if ($this->handleLazyParagraphContinuation($context, $cursor)) { |
122
|
36 |
|
return; |
123
|
|
|
} |
124
|
|
|
|
125
|
|
|
// not a lazy continuation |
126
|
|
|
// finalize any blocks not matched |
127
|
2457 |
|
$context->getBlockCloser()->closeUnmatchedBlocks(); |
128
|
|
|
|
129
|
|
|
// Determine whether the last line is blank, updating parents as needed |
130
|
2457 |
|
$this->setAndPropagateLastLineBlank($context, $cursor); |
131
|
|
|
|
132
|
|
|
// Handle any remaining cursor contents |
133
|
2457 |
|
if ($context->getContainer() instanceof StringContainerInterface) { |
134
|
888 |
|
$context->getContainer()->handleRemainingContents($context, $cursor); |
135
|
2187 |
|
} elseif (!$cursor->isBlank()) { |
136
|
|
|
// Create paragraph container for line |
137
|
2106 |
|
$p = new Paragraph(); |
138
|
2106 |
|
$context->addBlock($p); |
139
|
2106 |
|
$cursor->advanceToNextNonSpaceOrTab(); |
140
|
2106 |
|
$p->addLine($cursor->getRemainder()); |
141
|
|
|
} |
142
|
2457 |
|
} |
143
|
|
|
|
144
|
2457 |
|
private function processInlines(ContextInterface $context) |
145
|
|
|
{ |
146
|
2457 |
|
$walker = $context->getDocument()->walker(); |
147
|
|
|
|
148
|
2457 |
|
while ($event = $walker->next()) { |
149
|
2457 |
|
if (!$event->isEntering()) { |
150
|
2457 |
|
continue; |
151
|
|
|
} |
152
|
|
|
|
153
|
2457 |
|
$node = $event->getNode(); |
154
|
2457 |
|
if ($node instanceof AbstractStringContainerBlock) { |
155
|
2415 |
|
$this->inlineParserEngine->parse($node, $context->getDocument()->getReferenceMap()); |
156
|
|
|
} |
157
|
|
|
} |
158
|
2457 |
|
} |
159
|
|
|
|
160
|
|
|
/** |
161
|
|
|
* Sets the container to the last open child (or its parent) |
162
|
|
|
* |
163
|
|
|
* @param ContextInterface $context |
164
|
|
|
* @param Cursor $cursor |
165
|
|
|
*/ |
166
|
2457 |
|
private function resetContainer(ContextInterface $context, Cursor $cursor) |
167
|
|
|
{ |
168
|
2457 |
|
$container = $context->getDocument(); |
169
|
|
|
|
170
|
2457 |
|
while ($lastChild = $container->lastChild()) { |
171
|
1269 |
|
if (!($lastChild instanceof AbstractBlock)) { |
172
|
|
|
break; |
173
|
|
|
} |
174
|
|
|
|
175
|
1269 |
|
if (!$lastChild->isOpen()) { |
176
|
513 |
|
break; |
177
|
|
|
} |
178
|
|
|
|
179
|
1260 |
|
$container = $lastChild; |
180
|
1260 |
|
if (!$container->matchesNextLine($cursor)) { |
181
|
822 |
|
$container = $container->parent(); // back up to the last matching block |
182
|
822 |
|
break; |
183
|
|
|
} |
184
|
|
|
} |
185
|
|
|
|
186
|
2457 |
|
$context->setContainer($container); |
187
|
2457 |
|
} |
188
|
|
|
|
189
|
|
|
/** |
190
|
|
|
* Parse blocks |
191
|
|
|
* |
192
|
|
|
* @param ContextInterface $context |
193
|
|
|
* @param Cursor $cursor |
194
|
|
|
*/ |
195
|
2457 |
|
private function parseBlocks(ContextInterface $context, Cursor $cursor) |
196
|
|
|
{ |
197
|
2457 |
|
while (!$context->getContainer()->isCode() && !$context->getBlocksParsed()) { |
198
|
2457 |
|
$parsed = false; |
199
|
2457 |
|
foreach ($this->environment->getBlockParsers() as $parser) { |
200
|
2457 |
|
if ($parser->parse($context, $cursor)) { |
201
|
942 |
|
$parsed = true; |
202
|
1447 |
|
break; |
203
|
|
|
} |
204
|
|
|
} |
205
|
|
|
|
206
|
2457 |
|
if (!$parsed || $context->getContainer() instanceof StringContainerInterface || (($tip = $context->getTip()) && $tip->getDepth() >= $this->maxNestingLevel)) { |
207
|
2430 |
|
$context->setBlocksParsed(true); |
208
|
2430 |
|
break; |
209
|
|
|
} |
210
|
|
|
} |
211
|
2457 |
|
} |
212
|
|
|
|
213
|
|
|
/** |
214
|
|
|
* @param ContextInterface $context |
215
|
|
|
* @param Cursor $cursor |
216
|
|
|
* |
217
|
|
|
* @return bool |
218
|
|
|
*/ |
219
|
2457 |
|
private function handleLazyParagraphContinuation(ContextInterface $context, Cursor $cursor): bool |
220
|
|
|
{ |
221
|
2457 |
|
$tip = $context->getTip(); |
222
|
|
|
|
223
|
2457 |
|
if ($tip instanceof Paragraph && |
224
|
2457 |
|
!$context->getBlockCloser()->areAllClosed() && |
225
|
2457 |
|
!$cursor->isBlank() && |
226
|
2457 |
|
\count($tip->getStrings()) > 0) { |
227
|
|
|
|
228
|
|
|
// lazy paragraph continuation |
229
|
36 |
|
$tip->addLine($cursor->getRemainder()); |
230
|
|
|
|
231
|
36 |
|
return true; |
232
|
|
|
} |
233
|
|
|
|
234
|
2457 |
|
return false; |
235
|
|
|
} |
236
|
|
|
|
237
|
|
|
/** |
238
|
|
|
* @param ContextInterface $context |
239
|
|
|
* @param Cursor $cursor |
240
|
|
|
*/ |
241
|
2457 |
|
private function setAndPropagateLastLineBlank(ContextInterface $context, Cursor $cursor) |
242
|
|
|
{ |
243
|
2457 |
|
$container = $context->getContainer(); |
244
|
|
|
|
245
|
2457 |
|
if ($cursor->isBlank() && $lastChild = $container->lastChild()) { |
246
|
579 |
|
if ($lastChild instanceof AbstractBlock) { |
247
|
579 |
|
$lastChild->setLastLineBlank(true); |
248
|
|
|
} |
249
|
|
|
} |
250
|
|
|
|
251
|
2457 |
|
$lastLineBlank = $container->shouldLastLineBeBlank($cursor, $context->getLineNumber()); |
252
|
|
|
|
253
|
|
|
// Propagate lastLineBlank up through parents: |
254
|
2457 |
|
while ($container instanceof AbstractBlock && $container->endsWithBlankLine() !== $lastLineBlank) { |
255
|
729 |
|
$container->setLastLineBlank($lastLineBlank); |
256
|
729 |
|
$container = $container->parent(); |
257
|
|
|
} |
258
|
2457 |
|
} |
259
|
|
|
|
260
|
2463 |
|
private function assertValidUTF8(string $input) |
261
|
|
|
{ |
262
|
2463 |
|
if (!\mb_check_encoding($input, 'UTF-8')) { |
263
|
6 |
|
throw new UnexpectedEncodingException('Unexpected encoding - UTF-8 or ASCII was expected'); |
264
|
|
|
} |
265
|
2457 |
|
} |
266
|
|
|
} |
267
|
|
|
|