1
|
|
|
<?php // phpcs:ignore WordPress.Files.FileName.NotHyphenatedLowercase |
2
|
|
|
/** |
3
|
|
|
* Parser for a keepachangelog.com format changelog. |
4
|
|
|
* |
5
|
|
|
* @package automattic/jetpack-changelogger |
6
|
|
|
*/ |
7
|
|
|
|
8
|
|
|
// phpcs:disable WordPress.WP.AlternativeFunctions, WordPress.NamingConventions.ValidFunctionName.MethodNameInvalid, WordPress.NamingConventions.ValidVariableName |
9
|
|
|
|
10
|
|
|
namespace Automattic\Jetpack\Changelog; |
11
|
|
|
|
12
|
|
|
use DateTime; |
13
|
|
|
use DateTimeZone; |
14
|
|
|
use InvalidArgumentException; |
15
|
|
|
|
16
|
|
|
/** |
17
|
|
|
* Parser for a keepachangelog.com format changelog. |
18
|
|
|
*/ |
19
|
|
|
class KeepAChangelogParser extends Parser { |
20
|
|
|
|
21
|
|
|
/** |
22
|
|
|
* Bullet for changes. |
23
|
|
|
* |
24
|
|
|
* @var string |
25
|
|
|
*/ |
26
|
|
|
private $bullet = '-'; |
27
|
|
|
|
28
|
|
|
/** |
29
|
|
|
* Output date format. |
30
|
|
|
* |
31
|
|
|
* @var string |
32
|
|
|
*/ |
33
|
|
|
private $dateFormat = 'Y-m-d'; |
34
|
|
|
|
35
|
|
|
/** |
36
|
|
|
* If true, try to parse authors from entries. |
37
|
|
|
* |
38
|
|
|
* @var bool |
39
|
|
|
*/ |
40
|
|
|
private $parseAuthors = false; |
41
|
|
|
|
42
|
|
|
/** |
43
|
|
|
* Constructor. |
44
|
|
|
* |
45
|
|
|
* @param array $config Configuration. |
46
|
|
|
* - bullet: (string) Bullet for changes. Default '-'. |
47
|
|
|
* - dateFormat: (string) Date format to use in output. Default 'Y-m-d'. |
48
|
|
|
* - parseAuthors: (bool) Try to parse authors out of change entries. Default false. |
49
|
|
|
*/ |
50
|
|
|
public function __construct( array $config = array() ) { |
51
|
|
|
if ( ! empty( $config['bullet'] ) ) { |
52
|
|
|
$this->bullet = $config['bullet']; |
53
|
|
|
} |
54
|
|
|
if ( ! empty( $config['dateFormat'] ) ) { |
55
|
|
|
$this->dateFormat = $config['dateFormat']; |
56
|
|
|
} |
57
|
|
|
$this->parseAuthors = ! empty( $config['parseAuthors'] ); |
58
|
|
|
} |
59
|
|
|
|
60
|
|
|
/** |
61
|
|
|
* Test if there's a link at the end of a string. |
62
|
|
|
* |
63
|
|
|
* @param string $s String. |
64
|
|
|
* @return array|null Match data. |
65
|
|
|
*/ |
66
|
|
|
private function endsInLink( $s ) { |
67
|
|
|
if ( preg_match( '/^\[([^]]+)\]: *(\S+(?: +(?:"(?:[^"]|\\\\.)*"|\'(?:[^\']|\\\\.)*\'|\((?:[^()]|\\\\.)*\)))?)\s*\Z/m', $s, $m ) ) { |
68
|
|
|
return array( |
69
|
|
|
'match' => $m[0], |
70
|
|
|
'id' => $m[1], |
71
|
|
|
'link' => $m[2], |
72
|
|
|
); |
73
|
|
|
} |
74
|
|
|
return null; |
75
|
|
|
} |
76
|
|
|
|
77
|
|
|
/** |
78
|
|
|
* Split a string in two at the first occurrence of a substring. |
79
|
|
|
* |
80
|
|
|
* @param string $haystack String to split. |
81
|
|
|
* @param string[] ...$needles Strings to split on. Earliest match in $haystack wins. |
82
|
|
|
* @return string[] Two elements: The part before $needles and the part after, both trimmed. |
83
|
|
|
*/ |
|
|
|
|
84
|
|
|
private function split( $haystack, ...$needles ) { |
85
|
|
|
$i = false; |
86
|
|
|
foreach ( $needles as $needle ) { |
87
|
|
|
$j = strpos( $haystack, $needle ); |
88
|
|
|
if ( false !== $j ) { |
89
|
|
|
$i = false === $i ? $j : min( $i, $j ); |
90
|
|
|
} |
91
|
|
|
} |
92
|
|
|
if ( false === $i ) { |
93
|
|
|
return array( trim( $haystack, "\n" ), '' ); |
94
|
|
|
} |
95
|
|
|
return array( |
96
|
|
|
trim( substr( $haystack, 0, $i ), "\n" ), |
97
|
|
|
trim( substr( $haystack, $i ), "\n" ), |
98
|
|
|
); |
99
|
|
|
} |
100
|
|
|
|
101
|
|
|
/** |
102
|
|
|
* Parse changelog data into a Changelog object. |
103
|
|
|
* |
104
|
|
|
* This does not handle all markdown! In particular, it makes the following assumptions: |
105
|
|
|
* |
106
|
|
|
* - All level-2 ATX headings with no indentation are changelog entry headings. |
107
|
|
|
* - Changelog entry headings consist of either a bare version number or a version number as |
108
|
|
|
* link text with no destination or title, followed by a spaced ASCII hyphen, followed by a timestamp. |
109
|
|
|
* - All level-3 ATX headings with no indentation are changelog entry subheadings. |
110
|
|
|
* - All change entries are formatted as lists starting with the configured bullet followed by a space, |
111
|
|
|
* and do not make use of lazy continuation. Indentation of continued |
112
|
|
|
* lines is equal to the length of the bullet plus the space. |
113
|
|
|
* - All link definitions come at the end of the document, with no intervening blank lines or |
114
|
|
|
* other content, and are not indented and do not contain newlines. Link definitions for |
115
|
|
|
* changelog entries have no titles. |
116
|
|
|
* |
117
|
|
|
* @param string $changelog Changelog contents. |
118
|
|
|
* @return Changelog |
119
|
|
|
* @throws InvalidArgumentException If the changelog data cannot be parsed. |
120
|
|
|
*/ |
121
|
|
|
public function parse( $changelog ) { |
122
|
|
|
$ret = new Changelog(); |
123
|
|
|
|
124
|
|
|
$bullet = $this->bullet . ' '; |
125
|
|
|
$len = strlen( $bullet ); |
126
|
|
|
$indent = str_repeat( ' ', $len ); |
127
|
|
|
|
128
|
|
|
// Fix newlines and expand tabs. |
129
|
|
|
$changelog = strtr( $changelog, array( "\r\n" => "\n" ) ); |
130
|
|
|
$changelog = strtr( $changelog, array( "\r" => "\n" ) ); |
131
|
|
|
while ( strpos( $changelog, "\t" ) !== false ) { |
132
|
|
|
$changelog = preg_replace_callback( |
133
|
|
|
'/^([^\t\n]*)\t/m', |
134
|
|
|
function ( $m ) { |
135
|
|
|
return $m[1] . str_repeat( ' ', 4 - ( mb_strlen( $m[1] ) % 4 ) ); |
136
|
|
|
}, |
137
|
|
|
$changelog |
138
|
|
|
); |
139
|
|
|
} |
140
|
|
|
|
141
|
|
|
// Extract link definitions. |
142
|
|
|
$links = array(); |
143
|
|
|
$usedlinks = array(); |
144
|
|
|
while ( ( $m = $this->endsInLink( $changelog ) ) ) { // phpcs:ignore WordPress.CodeAnalysis.AssignmentInCondition.FoundInWhileCondition |
145
|
|
|
$links[ $m['id'] ] = $m['link']; |
146
|
|
|
$usedlinks[ $m['id'] ] = false; |
147
|
|
|
$changelog = (string) substr( $changelog, 0, -strlen( $m['match'] ) ); |
148
|
|
|
} |
149
|
|
|
$links = array_reverse( $links ); |
150
|
|
|
|
151
|
|
|
// Everything up to the first level-2 ATX heading is the prologue. |
152
|
|
|
list( $prologue, $changelog ) = $this->split( "\n$changelog", "\n## " ); |
153
|
|
|
$ret->setPrologue( $prologue ); |
154
|
|
|
|
155
|
|
|
// Entries make up the rest of the document. |
156
|
|
|
$entries = array(); |
157
|
|
|
while ( '' !== $changelog ) { |
158
|
|
|
// Extract the first entry from the changelog file, then extract the heading from it. |
159
|
|
|
list( $content, $changelog ) = $this->split( $changelog, "\n## " ); |
160
|
|
|
list( $heading, $content ) = $this->split( $content, "\n" ); |
161
|
|
|
|
162
|
|
|
// Parse the heading and create a ChangelogEntry for it. |
163
|
|
|
if ( ! preg_match( '/^## +(\[?[^] ]+\]?) - (.+?) *$/', $heading, $m ) ) { |
164
|
|
|
throw new InvalidArgumentException( "Invalid heading: $heading" ); |
165
|
|
|
} |
166
|
|
|
$link = null; |
167
|
|
|
$version = $m[1]; |
168
|
|
|
$timestamp = $m[2]; |
169
|
|
|
if ( '[' === $version[0] && ']' === substr( $version, -1 ) ) { |
170
|
|
|
$version = substr( $version, 1, -1 ); |
171
|
|
|
if ( ! isset( $links[ $version ] ) ) { |
172
|
|
|
throw new InvalidArgumentException( "Heading seems to have a linked version, but link was not found: $heading" ); |
173
|
|
|
} |
174
|
|
|
$link = $links[ $version ]; |
175
|
|
|
$usedlinks[ $version ] = true; |
176
|
|
|
} |
177
|
|
|
try { |
178
|
|
|
$timestamp = new DateTime( $timestamp, new DateTimeZone( 'UTC' ) ); |
179
|
|
|
} catch ( \Exception $ex ) { |
180
|
|
|
throw new InvalidArgumentException( "Heading has an invalid timestamp: $heading", 0, $ex ); |
181
|
|
|
} |
182
|
|
|
if ( strtotime( $m[2], 0 ) !== strtotime( $m[2], 1000000000 ) ) { |
183
|
|
|
throw new InvalidArgumentException( "Heading has a relative timestamp: $heading" ); |
184
|
|
|
} |
185
|
|
|
$entry = $this->newChangelogEntry( |
186
|
|
|
$version, |
187
|
|
|
array( |
188
|
|
|
'link' => $link, |
189
|
|
|
'timestamp' => $timestamp, |
190
|
|
|
) |
191
|
|
|
); |
192
|
|
|
$entries[] = $entry; |
193
|
|
|
|
194
|
|
|
// Extract the prologue, if any. |
195
|
|
|
list( $prologue, $content ) = $this->split( "\n$content", "\n### ", "\n$bullet" ); |
196
|
|
|
$entry->setPrologue( $prologue ); |
197
|
|
|
|
198
|
|
|
if ( '' === $content ) { |
199
|
|
|
// Huh, no changes. |
200
|
|
|
continue; |
201
|
|
|
} |
202
|
|
|
|
203
|
|
|
// Inject an empty heading if necessary so the change parsing can be more straightforward. |
204
|
|
|
if ( '#' !== $content[0] ) { |
205
|
|
|
$content = "### \n$content"; |
206
|
|
|
} |
207
|
|
|
|
208
|
|
|
// Now parse all the subheadings and changes. |
209
|
|
|
while ( '' !== $content ) { |
210
|
|
|
list( $section, $content ) = $this->split( $content, "\n### " ); |
211
|
|
|
list( $subheading, $section ) = $this->split( $section, "\n" ); |
212
|
|
|
$subheading = trim( substr( $subheading, 4 ) ); |
213
|
|
|
$changes = array(); |
214
|
|
|
$cur = ''; |
215
|
|
|
$section = explode( "\n", $section ); |
216
|
|
|
while ( $section ) { |
|
|
|
|
217
|
|
|
$line = array_shift( $section ); |
218
|
|
|
$prefix = substr( $line, 0, $len ); |
219
|
|
|
if ( $prefix === $bullet ) { |
220
|
|
|
$cur = trim( $cur ); |
221
|
|
|
if ( '' !== $cur ) { |
222
|
|
|
$changes[] = $cur; |
223
|
|
|
} |
224
|
|
|
$cur = substr( $line, $len ) . "\n"; |
225
|
|
|
} elseif ( $prefix === $indent ) { |
226
|
|
|
$cur .= substr( $line, $len ) . "\n"; |
227
|
|
|
} elseif ( '' === $line ) { |
228
|
|
|
$cur .= "\n"; |
229
|
|
|
} else { |
230
|
|
|
// If there are no more subsections and the rest of the lines don't contain |
231
|
|
|
// bullets, assume it's an epilogue. Otherwise, assume it's an error. |
232
|
|
|
$section = $line . "\n" . implode( "\n", $section ); |
233
|
|
|
if ( '' === $content && strpos( $section, "\n$bullet" ) === false ) { |
234
|
|
|
$entry->setEpilogue( $section ); |
235
|
|
|
break; |
236
|
|
|
} else { |
237
|
|
|
throw new InvalidArgumentException( "Malformatted changes list near \"$line\"" ); |
238
|
|
|
} |
239
|
|
|
} |
240
|
|
|
} |
241
|
|
|
$cur = trim( $cur ); |
242
|
|
|
if ( '' !== $cur ) { |
243
|
|
|
$changes[] = $cur; |
244
|
|
|
} |
245
|
|
|
foreach ( $changes as $change ) { |
246
|
|
|
$author = ''; |
247
|
|
|
if ( $this->parseAuthors && preg_match( '/ \(([^()\n]+)\)$/', $change, $m ) ) { |
248
|
|
|
$author = $m[1]; |
249
|
|
|
$change = substr( $change, 0, -strlen( $m[0] ) ); |
250
|
|
|
} |
251
|
|
|
$entry->appendChange( |
252
|
|
|
$this->newChangeEntry( |
253
|
|
|
array( |
254
|
|
|
'subheading' => $subheading, |
255
|
|
|
'author' => $author, |
256
|
|
|
'content' => $change, |
257
|
|
|
'timestamp' => $timestamp, |
258
|
|
|
) |
259
|
|
|
) |
260
|
|
|
); |
261
|
|
|
} |
262
|
|
|
} |
263
|
|
|
} |
264
|
|
|
$ret->setEntries( $entries ); |
265
|
|
|
|
266
|
|
|
// Append any unused links to the epilogue. |
267
|
|
|
$epilogue = $ret->getEpilogue(); |
268
|
|
|
foreach ( $links as $id => $content ) { |
269
|
|
|
if ( empty( $usedlinks[ $id ] ) ) { |
270
|
|
|
$epilogue .= "\n[$id]: $content"; |
271
|
|
|
} |
272
|
|
|
} |
273
|
|
|
$ret->setEpilogue( $epilogue ); |
274
|
|
|
|
275
|
|
|
return $ret; |
276
|
|
|
} |
277
|
|
|
|
278
|
|
|
/** |
279
|
|
|
* Write a Changelog object to a string. |
280
|
|
|
* |
281
|
|
|
* @param Changelog $changelog Changelog object. |
282
|
|
|
* @return string |
283
|
|
|
*/ |
284
|
|
|
public function format( Changelog $changelog ) { |
285
|
|
|
$ret = ''; |
286
|
|
|
|
287
|
|
|
$bullet = $this->bullet . ' '; |
288
|
|
|
$indent = str_repeat( ' ', strlen( $bullet ) ); |
289
|
|
|
|
290
|
|
|
$prologue = trim( $changelog->getPrologue() ); |
291
|
|
|
if ( '' !== $prologue ) { |
292
|
|
|
$ret .= "$prologue\n\n"; |
293
|
|
|
} |
294
|
|
|
|
295
|
|
|
$links = array(); |
296
|
|
|
foreach ( $changelog->getEntries() as $entry ) { |
297
|
|
|
$ret .= '## '; |
298
|
|
|
if ( $entry->getLink() !== null ) { |
299
|
|
|
$links[ $entry->getVersion() ] = $entry->getLink(); |
300
|
|
|
$ret .= "[{$entry->getVersion()}]"; |
301
|
|
|
} else { |
302
|
|
|
$ret .= $entry->getVersion(); |
303
|
|
|
} |
304
|
|
|
$ret .= ' - ' . $entry->getTimestamp()->format( $this->dateFormat ) . "\n"; |
305
|
|
|
|
306
|
|
|
$prologue = trim( $entry->getPrologue() ); |
307
|
|
|
if ( '' !== $prologue ) { |
308
|
|
|
$ret .= "\n$prologue\n\n"; |
309
|
|
|
} |
310
|
|
|
|
311
|
|
|
foreach ( $entry->getChangesBySubheading() as $heading => $changes ) { |
312
|
|
|
if ( '' !== $heading ) { |
|
|
|
|
313
|
|
|
$heading = "### $heading\n"; |
314
|
|
|
} else { |
315
|
|
|
$heading = substr( $ret, -2 ) === "\n\n" ? '' : "\n"; |
316
|
|
|
} |
317
|
|
|
foreach ( $changes as $change ) { |
|
|
|
|
318
|
|
|
$text = trim( $change->getContent() ); |
319
|
|
|
if ( '' !== $text ) { |
320
|
|
|
if ( $change->getAuthor() !== '' ) { |
321
|
|
|
$text .= " ({$change->getAuthor()})"; |
322
|
|
|
} |
323
|
|
|
$ret .= $heading . $bullet . str_replace( "\n", "\n$indent", $text ) . "\n"; |
324
|
|
|
$heading = ''; |
325
|
|
|
} |
326
|
|
|
} |
327
|
|
|
$ret .= "\n"; |
328
|
|
|
} |
329
|
|
|
|
330
|
|
|
$epilogue = trim( $entry->getEpilogue() ); |
331
|
|
|
if ( '' !== $epilogue ) { |
332
|
|
|
$ret = trim( $ret ) . "\n\n$epilogue\n"; |
333
|
|
|
} |
334
|
|
|
$ret = trim( $ret ) . "\n\n"; |
335
|
|
|
} |
336
|
|
|
|
337
|
|
|
$epilogue = trim( $changelog->getEpilogue() ); |
338
|
|
|
if ( '' !== $epilogue ) { |
339
|
|
|
$ret .= "$epilogue\n"; |
340
|
|
|
} |
341
|
|
|
|
342
|
|
|
$ret = trim( $ret ) . "\n"; |
343
|
|
|
|
344
|
|
|
if ( $links ) { |
|
|
|
|
345
|
|
|
if ( ! $this->endsInLink( $ret ) ) { |
346
|
|
|
$ret .= "\n"; |
347
|
|
|
} |
348
|
|
|
foreach ( $links as $k => $v ) { |
349
|
|
|
$ret .= "[$k]: $v\n"; |
350
|
|
|
} |
351
|
|
|
} |
352
|
|
|
|
353
|
|
|
return $ret; |
354
|
|
|
} |
355
|
|
|
|
356
|
|
|
} |
357
|
|
|
|
This check looks for
@param
annotations where the type inferred by our type inference engine differs from the declared type.It makes a suggestion as to what type it considers more descriptive.
Most often this is a case of a parameter that can be null in addition to its declared types.