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 (http://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\Document; |
19
|
|
|
use League\CommonMark\Block\Element\InlineContainer; |
20
|
|
|
use League\CommonMark\Block\Element\ListBlock; |
21
|
|
|
use League\CommonMark\Block\Element\Paragraph; |
22
|
|
|
use League\CommonMark\Node\NodeWalker; |
23
|
|
|
|
24
|
|
|
class DocParser |
25
|
|
|
{ |
26
|
|
|
/** |
27
|
|
|
* @var Environment |
28
|
|
|
*/ |
29
|
|
|
protected $environment; |
30
|
|
|
|
31
|
|
|
/** |
32
|
|
|
* @var InlineParserEngine |
33
|
|
|
*/ |
34
|
|
|
private $inlineParserEngine; |
35
|
|
|
|
36
|
|
|
/** |
37
|
|
|
* @param Environment $environment |
38
|
|
|
*/ |
39
|
1872 |
|
public function __construct(Environment $environment) |
40
|
|
|
{ |
41
|
1872 |
|
$this->environment = $environment; |
42
|
1872 |
|
$this->inlineParserEngine = new InlineParserEngine($environment); |
43
|
1872 |
|
} |
44
|
|
|
|
45
|
|
|
/** |
46
|
|
|
* @return Environment |
47
|
|
|
*/ |
48
|
1872 |
|
public function getEnvironment() |
49
|
|
|
{ |
50
|
1872 |
|
return $this->environment; |
51
|
|
|
} |
52
|
|
|
|
53
|
|
|
/** |
54
|
|
|
* @param string $input |
55
|
|
|
* |
56
|
|
|
* @return string[] |
57
|
|
|
*/ |
58
|
1872 |
|
private function preProcessInput($input) |
59
|
|
|
{ |
60
|
1872 |
|
$lines = preg_split('/\r\n|\n|\r/', $input); |
61
|
|
|
|
62
|
|
|
// Remove any newline which appears at the very end of the string. |
63
|
|
|
// We've already split the document by newlines, so we can simply drop |
64
|
|
|
// any empty element which appears on the end. |
65
|
1872 |
|
if (end($lines) === '') { |
66
|
1866 |
|
array_pop($lines); |
67
|
1866 |
|
} |
68
|
|
|
|
69
|
1872 |
|
return $lines; |
70
|
|
|
} |
71
|
|
|
|
72
|
|
|
/** |
73
|
|
|
* @param string $input |
74
|
|
|
* |
75
|
|
|
* @return Document |
76
|
|
|
*/ |
77
|
1872 |
|
public function parse($input) |
78
|
3 |
|
{ |
79
|
1872 |
|
$context = new Context(new Document(), $this->getEnvironment()); |
80
|
|
|
|
81
|
1872 |
|
$lines = $this->preProcessInput($input); |
82
|
1872 |
|
foreach ($lines as $line) { |
83
|
1872 |
|
$context->setNextLine($line); |
84
|
1872 |
|
$this->incorporateLine($context); |
85
|
1872 |
|
} |
86
|
|
|
|
87
|
1872 |
|
while ($context->getTip()) { |
88
|
1872 |
|
$context->getTip()->finalize($context, count($lines)); |
89
|
1872 |
|
} |
90
|
|
|
|
91
|
1872 |
|
$this->processInlines($context, $context->getDocument()->walker()); |
92
|
|
|
|
93
|
1872 |
|
$this->processDocument($context); |
94
|
|
|
|
95
|
1872 |
|
return $context->getDocument(); |
96
|
|
|
} |
97
|
|
|
|
98
|
1872 |
|
private function incorporateLine(ContextInterface $context) |
99
|
|
|
{ |
100
|
1872 |
|
$cursor = new Cursor($context->getLine()); |
101
|
1872 |
|
$context->getBlockCloser()->resetTip(); |
102
|
|
|
|
103
|
1872 |
|
$context->setBlocksParsed(false); |
104
|
|
|
|
105
|
1872 |
|
$this->resetContainer($context, $cursor); |
106
|
1872 |
|
$context->getBlockCloser()->setLastMatchedContainer($context->getContainer()); |
107
|
|
|
|
108
|
|
|
// Check to see if we've hit 2nd blank line; if so break out of list: |
109
|
1872 |
|
if ($cursor->isBlank() && $context->getContainer()->endsWithBlankLine()) { |
110
|
39 |
|
$this->breakOutOfLists($context, $context->getContainer()); |
111
|
39 |
|
} |
112
|
|
|
|
113
|
1872 |
|
$this->parseBlocks($context, $cursor); |
114
|
|
|
|
115
|
|
|
// What remains at the offset is a text line. Add the text to the appropriate container. |
116
|
|
|
// First check for a lazy paragraph continuation: |
117
|
1872 |
|
if ($this->isLazyParagraphContinuation($context, $cursor)) { |
118
|
|
|
// lazy paragraph continuation |
119
|
30 |
|
$context->getTip()->addLine($cursor->getRemainder()); |
120
|
|
|
|
121
|
30 |
|
return; |
122
|
|
|
} |
123
|
|
|
|
124
|
|
|
// not a lazy continuation |
125
|
|
|
// finalize any blocks not matched |
126
|
1872 |
|
$context->getBlockCloser()->closeUnmatchedBlocks(); |
127
|
|
|
|
128
|
|
|
// Determine whether the last line is blank, updating parents as needed |
129
|
1872 |
|
$this->setAndPropagateLastLineBlank($context, $cursor); |
130
|
|
|
|
131
|
|
|
// Handle any remaining cursor contents |
132
|
1872 |
|
if ($context->getContainer()->acceptsLines()) { |
133
|
681 |
|
$context->getContainer()->handleRemainingContents($context, $cursor); |
134
|
1872 |
|
} elseif (!$cursor->isBlank()) { |
135
|
|
|
// Create paragraph container for line |
136
|
1560 |
|
$context->addBlock(new Paragraph()); |
137
|
1560 |
|
$cursor->advanceToFirstNonSpace(); |
138
|
1560 |
|
$context->getTip()->addLine($cursor->getRemainder()); |
139
|
1560 |
|
} |
140
|
1872 |
|
} |
141
|
|
|
|
142
|
1872 |
|
private function processDocument(ContextInterface $context) |
143
|
|
|
{ |
144
|
1872 |
|
foreach ($this->getEnvironment()->getDocumentProcessors() as $documentProcessor) { |
145
|
|
|
$documentProcessor->processDocument($context->getDocument()); |
146
|
1872 |
|
} |
147
|
1872 |
|
} |
148
|
|
|
|
149
|
1872 |
|
private function processInlines(ContextInterface $context, NodeWalker $walker) |
150
|
|
|
{ |
151
|
1872 |
|
while (($event = $walker->next()) !== null) { |
152
|
1872 |
|
if (!$event->isEntering()) { |
153
|
1872 |
|
continue; |
154
|
|
|
} |
155
|
|
|
|
156
|
1872 |
|
$node = $event->getNode(); |
157
|
1872 |
|
if ($node instanceof InlineContainer) { |
158
|
1602 |
|
$this->inlineParserEngine->parse($node, $context->getDocument()->getReferenceMap()); |
159
|
1602 |
|
} |
160
|
1872 |
|
} |
161
|
1872 |
|
} |
162
|
|
|
|
163
|
|
|
/** |
164
|
|
|
* Break out of all containing lists, resetting the tip of the |
165
|
|
|
* document to the parent of the highest list, and finalizing |
166
|
|
|
* all the lists. (This is used to implement the "two blank lines |
167
|
|
|
* break out of all lists" feature.) |
168
|
|
|
* |
169
|
|
|
* @param ContextInterface $context |
170
|
|
|
* @param AbstractBlock $block |
171
|
|
|
*/ |
172
|
39 |
|
private function breakOutOfLists(ContextInterface $context, AbstractBlock $block) |
173
|
|
|
{ |
174
|
39 |
|
$b = $block; |
175
|
39 |
|
$lastList = null; |
176
|
|
|
do { |
177
|
39 |
|
if ($b instanceof ListBlock) { |
178
|
24 |
|
$lastList = $b; |
179
|
24 |
|
} |
180
|
39 |
|
$b = $b->parent(); |
181
|
39 |
|
} while ($b); |
182
|
|
|
|
183
|
39 |
|
if ($lastList) { |
184
|
24 |
|
while ($block !== $lastList) { |
185
|
24 |
|
$block->finalize($context, $context->getLineNumber()); |
186
|
24 |
|
$block = $block->parent(); |
187
|
24 |
|
} |
188
|
24 |
|
$lastList->finalize($context, $context->getLineNumber()); |
189
|
24 |
|
} |
190
|
|
|
|
191
|
39 |
|
$context->setContainer($context->getTip()); |
192
|
39 |
|
} |
193
|
|
|
|
194
|
|
|
/** |
195
|
|
|
* Sets the container to the last open child (or its parent) |
196
|
|
|
* |
197
|
|
|
* @param ContextInterface $context |
198
|
|
|
* @param Cursor $cursor |
199
|
|
|
*/ |
200
|
1872 |
|
private function resetContainer(ContextInterface $context, Cursor $cursor) |
201
|
|
|
{ |
202
|
1872 |
|
$context->setContainer($context->getDocument()); |
203
|
|
|
|
204
|
1872 |
|
while ($context->getContainer()->hasChildren()) { |
205
|
1026 |
|
$lastChild = $context->getContainer()->lastChild(); |
206
|
1026 |
|
if (!$lastChild->isOpen()) { |
207
|
399 |
|
break; |
208
|
|
|
} |
209
|
|
|
|
210
|
1017 |
|
$context->setContainer($lastChild); |
211
|
1017 |
|
if (!$context->getContainer()->matchesNextLine($cursor)) { |
212
|
654 |
|
$context->setContainer($context->getContainer()->parent()); // back up to the last matching block |
213
|
654 |
|
break; |
214
|
|
|
} |
215
|
714 |
|
} |
216
|
1872 |
|
} |
217
|
|
|
|
218
|
|
|
/** |
219
|
|
|
* Parse blocks |
220
|
|
|
* |
221
|
|
|
* @param ContextInterface $context |
222
|
|
|
* @param Cursor $cursor |
223
|
|
|
*/ |
224
|
1872 |
|
private function parseBlocks(ContextInterface $context, Cursor $cursor) |
225
|
|
|
{ |
226
|
1872 |
|
while (!$context->getContainer()->isCode() && !$context->getBlocksParsed()) { |
227
|
1872 |
|
$parsed = false; |
228
|
1872 |
|
foreach ($this->environment->getBlockParsers() as $parser) { |
229
|
1872 |
|
if ($parser->parse($context, $cursor)) { |
230
|
747 |
|
$parsed = true; |
231
|
747 |
|
break; |
232
|
|
|
} |
233
|
1872 |
|
} |
234
|
|
|
|
235
|
1872 |
|
if (!$parsed || $context->getContainer()->acceptsLines()) { |
236
|
1848 |
|
$context->setBlocksParsed(true); |
237
|
1848 |
|
} |
238
|
1872 |
|
} |
239
|
1872 |
|
} |
240
|
|
|
|
241
|
|
|
/** |
242
|
|
|
* @param ContextInterface $context |
243
|
|
|
* @param Cursor $cursor |
244
|
|
|
* |
245
|
|
|
* @return bool |
246
|
|
|
*/ |
247
|
1872 |
|
private function isLazyParagraphContinuation(ContextInterface $context, Cursor $cursor) |
248
|
|
|
{ |
249
|
1872 |
|
return !$context->getBlockCloser()->areAllClosed() && |
250
|
1872 |
|
!$cursor->isBlank() && |
251
|
1872 |
|
$context->getTip() instanceof Paragraph && |
252
|
1872 |
|
count($context->getTip()->getStrings()) > 0; |
253
|
|
|
} |
254
|
|
|
|
255
|
|
|
/** |
256
|
|
|
* @param ContextInterface $context |
257
|
|
|
* @param Cursor $cursor |
258
|
|
|
*/ |
259
|
1872 |
|
private function setAndPropagateLastLineBlank(ContextInterface $context, $cursor) |
260
|
|
|
{ |
261
|
1872 |
|
if ($cursor->isBlank() && $lastChild = $context->getContainer()->lastChild()) { |
262
|
426 |
|
if ($lastChild instanceof AbstractBlock) { |
263
|
426 |
|
$lastChild->setLastLineBlank(true); |
264
|
426 |
|
} |
265
|
426 |
|
} |
266
|
|
|
|
267
|
1872 |
|
$container = $context->getContainer(); |
268
|
1872 |
|
$lastLineBlank = $container->shouldLastLineBeBlank($cursor, $context->getLineNumber()); |
269
|
|
|
|
270
|
|
|
// Propagate lastLineBlank up through parents: |
271
|
1872 |
|
while ($container) { |
272
|
1872 |
|
$container->setLastLineBlank($lastLineBlank); |
273
|
1872 |
|
$container = $container->parent(); |
274
|
1872 |
|
} |
275
|
1872 |
|
} |
276
|
|
|
} |
277
|
|
|
|