|
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
|
|
|
* String used as the date for an unreleased version. |
|
37
|
|
|
* |
|
38
|
|
|
* @var string |
|
39
|
|
|
*/ |
|
40
|
|
|
private $unreleased = 'unreleased'; |
|
41
|
|
|
|
|
42
|
|
|
/** |
|
43
|
|
|
* If true, try to parse authors from entries. |
|
44
|
|
|
* |
|
45
|
|
|
* @var bool |
|
46
|
|
|
*/ |
|
47
|
|
|
private $parseAuthors = false; |
|
48
|
|
|
|
|
49
|
|
|
/** |
|
50
|
|
|
* Constructor. |
|
51
|
|
|
* |
|
52
|
|
|
* @param array $config Configuration. |
|
53
|
|
|
* - bullet: (string) Bullet for changes. Default '-'. |
|
54
|
|
|
* - dateFormat: (string) Date format to use in output. Default 'Y-m-d'. |
|
55
|
|
|
* - parseAuthors: (bool) Try to parse authors out of change entries. Default false. |
|
56
|
|
|
* - unreleased: (string) String used as the date for an unreleased version. |
|
57
|
|
|
*/ |
|
58
|
|
|
public function __construct( array $config = array() ) { |
|
59
|
|
|
if ( ! empty( $config['bullet'] ) ) { |
|
60
|
|
|
$this->bullet = $config['bullet']; |
|
61
|
|
|
} |
|
62
|
|
|
if ( ! empty( $config['dateFormat'] ) ) { |
|
63
|
|
|
$this->dateFormat = $config['dateFormat']; |
|
64
|
|
|
} |
|
65
|
|
|
$this->parseAuthors = ! empty( $config['parseAuthors'] ); |
|
66
|
|
|
if ( ! empty( $config['unreleased'] ) ) { |
|
67
|
|
|
$this->unreleased = $config['unreleased']; |
|
68
|
|
|
} |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
|
|
/** |
|
72
|
|
|
* Test if there's a link at the end of a string. |
|
73
|
|
|
* |
|
74
|
|
|
* @param string $s String. |
|
75
|
|
|
* @return array|null Match data. |
|
76
|
|
|
*/ |
|
77
|
|
|
private function endsInLink( $s ) { |
|
78
|
|
|
if ( preg_match( '/^\[([^]]+)\]: *(\S+(?: +(?:"(?:[^"]|\\\\.)*"|\'(?:[^\']|\\\\.)*\'|\((?:[^()]|\\\\.)*\)))?)\s*\Z/m', $s, $m ) ) { |
|
79
|
|
|
return array( |
|
80
|
|
|
'match' => $m[0], |
|
81
|
|
|
'id' => $m[1], |
|
82
|
|
|
'link' => $m[2], |
|
83
|
|
|
); |
|
84
|
|
|
} |
|
85
|
|
|
return null; |
|
86
|
|
|
} |
|
87
|
|
|
|
|
88
|
|
|
/** |
|
89
|
|
|
* Split a string in two at the first occurrence of a substring. |
|
90
|
|
|
* |
|
91
|
|
|
* @param string $haystack String to split. |
|
92
|
|
|
* @param string[] ...$needles Strings to split on. Earliest match in $haystack wins. |
|
93
|
|
|
* @return string[] Two elements: The part before $needles and the part after, both trimmed. |
|
94
|
|
|
*/ |
|
|
|
|
|
|
95
|
|
|
private function split( $haystack, ...$needles ) { |
|
96
|
|
|
$i = false; |
|
97
|
|
|
foreach ( $needles as $needle ) { |
|
98
|
|
|
$j = strpos( $haystack, $needle ); |
|
99
|
|
|
if ( false !== $j ) { |
|
100
|
|
|
$i = false === $i ? $j : min( $i, $j ); |
|
101
|
|
|
} |
|
102
|
|
|
} |
|
103
|
|
|
if ( false === $i ) { |
|
104
|
|
|
return array( trim( $haystack, "\n" ), '' ); |
|
105
|
|
|
} |
|
106
|
|
|
return array( |
|
107
|
|
|
trim( substr( $haystack, 0, $i ), "\n" ), |
|
108
|
|
|
trim( substr( $haystack, $i ), "\n" ), |
|
109
|
|
|
); |
|
110
|
|
|
} |
|
111
|
|
|
|
|
112
|
|
|
/** |
|
113
|
|
|
* Parse changelog data into a Changelog object. |
|
114
|
|
|
* |
|
115
|
|
|
* This does not handle all markdown! In particular, it makes the following assumptions: |
|
116
|
|
|
* |
|
117
|
|
|
* - All level-2 ATX headings with no indentation are changelog entry headings. |
|
118
|
|
|
* - Changelog entry headings consist of either a bare version number or a version number as |
|
119
|
|
|
* link text with no destination or title, followed by a spaced ASCII hyphen, followed by a timestamp. |
|
120
|
|
|
* - All level-3 ATX headings with no indentation are changelog entry subheadings. |
|
121
|
|
|
* - All change entries are formatted as lists starting with the configured bullet followed by a space, |
|
122
|
|
|
* and do not make use of lazy continuation. Indentation of continued |
|
123
|
|
|
* lines is equal to the length of the bullet plus the space. |
|
124
|
|
|
* - All link definitions come at the end of the document, with no intervening blank lines or |
|
125
|
|
|
* other content, and are not indented and do not contain newlines. Link definitions for |
|
126
|
|
|
* changelog entries have no titles. |
|
127
|
|
|
* |
|
128
|
|
|
* @param string $changelog Changelog contents. |
|
129
|
|
|
* @return Changelog |
|
130
|
|
|
* @throws InvalidArgumentException If the changelog data cannot be parsed. |
|
131
|
|
|
*/ |
|
132
|
|
|
public function parse( $changelog ) { |
|
133
|
|
|
$ret = new Changelog(); |
|
134
|
|
|
|
|
135
|
|
|
$bullet = $this->bullet . ' '; |
|
136
|
|
|
$len = strlen( $bullet ); |
|
137
|
|
|
$indent = str_repeat( ' ', $len ); |
|
138
|
|
|
|
|
139
|
|
|
// Fix newlines and expand tabs. |
|
140
|
|
|
$changelog = strtr( $changelog, array( "\r\n" => "\n" ) ); |
|
141
|
|
|
$changelog = strtr( $changelog, array( "\r" => "\n" ) ); |
|
142
|
|
|
while ( strpos( $changelog, "\t" ) !== false ) { |
|
143
|
|
|
$changelog = preg_replace_callback( |
|
144
|
|
|
'/^([^\t\n]*)\t/m', |
|
145
|
|
|
function ( $m ) { |
|
146
|
|
|
return $m[1] . str_repeat( ' ', 4 - ( mb_strlen( $m[1] ) % 4 ) ); |
|
147
|
|
|
}, |
|
148
|
|
|
$changelog |
|
149
|
|
|
); |
|
150
|
|
|
} |
|
151
|
|
|
|
|
152
|
|
|
// Extract link definitions. |
|
153
|
|
|
$links = array(); |
|
154
|
|
|
$usedlinks = array(); |
|
155
|
|
|
while ( ( $m = $this->endsInLink( $changelog ) ) ) { // phpcs:ignore WordPress.CodeAnalysis.AssignmentInCondition.FoundInWhileCondition |
|
156
|
|
|
$links[ $m['id'] ] = $m['link']; |
|
157
|
|
|
$usedlinks[ $m['id'] ] = false; |
|
158
|
|
|
$changelog = (string) substr( $changelog, 0, -strlen( $m['match'] ) ); |
|
159
|
|
|
} |
|
160
|
|
|
$links = array_reverse( $links ); |
|
161
|
|
|
|
|
162
|
|
|
// Everything up to the first level-2 ATX heading is the prologue. |
|
163
|
|
|
list( $prologue, $changelog ) = $this->split( "\n$changelog", "\n## " ); |
|
164
|
|
|
$ret->setPrologue( $prologue ); |
|
165
|
|
|
|
|
166
|
|
|
// Entries make up the rest of the document. |
|
167
|
|
|
$entries = array(); |
|
168
|
|
|
while ( '' !== $changelog ) { |
|
169
|
|
|
// Extract the first entry from the changelog file, then extract the heading from it. |
|
170
|
|
|
list( $content, $changelog ) = $this->split( $changelog, "\n## " ); |
|
171
|
|
|
list( $heading, $content ) = $this->split( $content, "\n" ); |
|
172
|
|
|
|
|
173
|
|
|
// Parse the heading and create a ChangelogEntry for it. |
|
174
|
|
|
if ( ! preg_match( '/^## +(\[?[^] ]+\]?) - (.+?) *$/', $heading, $m ) ) { |
|
175
|
|
|
throw new InvalidArgumentException( "Invalid heading: $heading" ); |
|
176
|
|
|
} |
|
177
|
|
|
$link = null; |
|
178
|
|
|
$version = $m[1]; |
|
179
|
|
|
$timestamp = $m[2]; |
|
180
|
|
|
if ( '[' === $version[0] && ']' === substr( $version, -1 ) ) { |
|
181
|
|
|
$version = substr( $version, 1, -1 ); |
|
182
|
|
|
if ( ! isset( $links[ $version ] ) ) { |
|
183
|
|
|
throw new InvalidArgumentException( "Heading seems to have a linked version, but link was not found: $heading" ); |
|
184
|
|
|
} |
|
185
|
|
|
$link = $links[ $version ]; |
|
186
|
|
|
$usedlinks[ $version ] = true; |
|
187
|
|
|
} |
|
188
|
|
|
if ( $timestamp === $this->unreleased ) { |
|
189
|
|
|
$timestamp = null; |
|
190
|
|
|
$entryTimestamp = new DateTime( 'now', new DateTimeZone( 'UTC' ) ); |
|
191
|
|
|
} else { |
|
192
|
|
|
try { |
|
193
|
|
|
$timestamp = new DateTime( $timestamp, new DateTimeZone( 'UTC' ) ); |
|
194
|
|
|
} catch ( \Exception $ex ) { |
|
195
|
|
|
throw new InvalidArgumentException( "Heading has an invalid timestamp: $heading", 0, $ex ); |
|
196
|
|
|
} |
|
197
|
|
|
if ( strtotime( $m[2], 0 ) !== strtotime( $m[2], 1000000000 ) ) { |
|
198
|
|
|
throw new InvalidArgumentException( "Heading has a relative timestamp: $heading" ); |
|
199
|
|
|
} |
|
200
|
|
|
$entryTimestamp = $timestamp; |
|
201
|
|
|
} |
|
202
|
|
|
$entry = $this->newChangelogEntry( |
|
203
|
|
|
$version, |
|
204
|
|
|
array( |
|
205
|
|
|
'link' => $link, |
|
206
|
|
|
'timestamp' => $timestamp, |
|
207
|
|
|
) |
|
208
|
|
|
); |
|
209
|
|
|
$entries[] = $entry; |
|
210
|
|
|
|
|
211
|
|
|
// Extract the prologue, if any. |
|
212
|
|
|
list( $prologue, $content ) = $this->split( "\n$content", "\n### ", "\n$bullet" ); |
|
213
|
|
|
$entry->setPrologue( $prologue ); |
|
214
|
|
|
|
|
215
|
|
|
if ( '' === $content ) { |
|
216
|
|
|
// Huh, no changes. |
|
217
|
|
|
continue; |
|
218
|
|
|
} |
|
219
|
|
|
|
|
220
|
|
|
// Inject an empty heading if necessary so the change parsing can be more straightforward. |
|
221
|
|
|
if ( '#' !== $content[0] ) { |
|
222
|
|
|
$content = "### \n$content"; |
|
223
|
|
|
} |
|
224
|
|
|
|
|
225
|
|
|
// Now parse all the subheadings and changes. |
|
226
|
|
|
while ( '' !== $content ) { |
|
227
|
|
|
list( $section, $content ) = $this->split( $content, "\n### " ); |
|
228
|
|
|
list( $subheading, $section ) = $this->split( $section, "\n" ); |
|
229
|
|
|
$subheading = trim( substr( $subheading, 4 ) ); |
|
230
|
|
|
$changes = array(); |
|
231
|
|
|
$cur = ''; |
|
232
|
|
|
$section = explode( "\n", $section ); |
|
233
|
|
|
while ( $section ) { |
|
|
|
|
|
|
234
|
|
|
$line = array_shift( $section ); |
|
235
|
|
|
$prefix = substr( $line, 0, $len ); |
|
236
|
|
|
if ( $prefix === $bullet ) { |
|
237
|
|
|
$cur = trim( $cur ); |
|
238
|
|
|
if ( '' !== $cur ) { |
|
239
|
|
|
$changes[] = $cur; |
|
240
|
|
|
} |
|
241
|
|
|
$cur = substr( $line, $len ) . "\n"; |
|
242
|
|
|
} elseif ( $prefix === $indent ) { |
|
243
|
|
|
$cur .= substr( $line, $len ) . "\n"; |
|
244
|
|
|
} elseif ( '' === $line ) { |
|
245
|
|
|
$cur .= "\n"; |
|
246
|
|
|
} else { |
|
247
|
|
|
// If there are no more subsections and the rest of the lines don't contain |
|
248
|
|
|
// bullets, assume it's an epilogue. Otherwise, assume it's an error. |
|
249
|
|
|
$section = $line . "\n" . implode( "\n", $section ); |
|
250
|
|
|
if ( '' === $content && strpos( $section, "\n$bullet" ) === false ) { |
|
251
|
|
|
$entry->setEpilogue( $section ); |
|
252
|
|
|
break; |
|
253
|
|
|
} else { |
|
254
|
|
|
throw new InvalidArgumentException( "Malformatted changes list near \"$line\"" ); |
|
255
|
|
|
} |
|
256
|
|
|
} |
|
257
|
|
|
} |
|
258
|
|
|
$cur = trim( $cur ); |
|
259
|
|
|
if ( '' !== $cur ) { |
|
260
|
|
|
$changes[] = $cur; |
|
261
|
|
|
} |
|
262
|
|
|
foreach ( $changes as $change ) { |
|
263
|
|
|
$author = ''; |
|
264
|
|
|
if ( $this->parseAuthors && preg_match( '/ \(([^()\n]+)\)$/', $change, $m ) ) { |
|
265
|
|
|
$author = $m[1]; |
|
266
|
|
|
$change = substr( $change, 0, -strlen( $m[0] ) ); |
|
267
|
|
|
} |
|
268
|
|
|
$entry->appendChange( |
|
269
|
|
|
$this->newChangeEntry( |
|
270
|
|
|
array( |
|
271
|
|
|
'subheading' => $subheading, |
|
272
|
|
|
'author' => $author, |
|
273
|
|
|
'content' => $change, |
|
274
|
|
|
'timestamp' => $entryTimestamp, |
|
275
|
|
|
) |
|
276
|
|
|
) |
|
277
|
|
|
); |
|
278
|
|
|
} |
|
279
|
|
|
} |
|
280
|
|
|
} |
|
281
|
|
|
$ret->setEntries( $entries ); |
|
282
|
|
|
|
|
283
|
|
|
// Append any unused links to the epilogue. |
|
284
|
|
|
$epilogue = $ret->getEpilogue(); |
|
285
|
|
|
foreach ( $links as $id => $content ) { |
|
286
|
|
|
if ( empty( $usedlinks[ $id ] ) ) { |
|
287
|
|
|
$epilogue .= "\n[$id]: $content"; |
|
288
|
|
|
} |
|
289
|
|
|
} |
|
290
|
|
|
$ret->setEpilogue( $epilogue ); |
|
291
|
|
|
|
|
292
|
|
|
return $ret; |
|
293
|
|
|
} |
|
294
|
|
|
|
|
295
|
|
|
/** |
|
296
|
|
|
* Write a Changelog object to a string. |
|
297
|
|
|
* |
|
298
|
|
|
* @param Changelog $changelog Changelog object. |
|
299
|
|
|
* @return string |
|
300
|
|
|
*/ |
|
301
|
|
|
public function format( Changelog $changelog ) { |
|
302
|
|
|
$ret = ''; |
|
303
|
|
|
|
|
304
|
|
|
$bullet = $this->bullet . ' '; |
|
305
|
|
|
$indent = str_repeat( ' ', strlen( $bullet ) ); |
|
306
|
|
|
|
|
307
|
|
|
$prologue = trim( $changelog->getPrologue() ); |
|
308
|
|
|
if ( '' !== $prologue ) { |
|
309
|
|
|
$ret .= "$prologue\n\n"; |
|
310
|
|
|
} |
|
311
|
|
|
|
|
312
|
|
|
$links = array(); |
|
313
|
|
|
foreach ( $changelog->getEntries() as $entry ) { |
|
314
|
|
|
$ret .= '## '; |
|
315
|
|
|
if ( $entry->getLink() !== null ) { |
|
316
|
|
|
$links[ $entry->getVersion() ] = $entry->getLink(); |
|
317
|
|
|
$ret .= "[{$entry->getVersion()}]"; |
|
318
|
|
|
} else { |
|
319
|
|
|
$ret .= $entry->getVersion(); |
|
320
|
|
|
} |
|
321
|
|
|
$timestamp = $entry->getTimestamp(); |
|
322
|
|
|
$ret .= ' - ' . ( null === $timestamp ? $this->unreleased : $timestamp->format( $this->dateFormat ) ) . "\n"; |
|
323
|
|
|
|
|
324
|
|
|
$prologue = trim( $entry->getPrologue() ); |
|
325
|
|
|
if ( '' !== $prologue ) { |
|
326
|
|
|
$ret .= "\n$prologue\n\n"; |
|
327
|
|
|
} |
|
328
|
|
|
|
|
329
|
|
|
foreach ( $entry->getChangesBySubheading() as $heading => $changes ) { |
|
330
|
|
|
if ( '' !== $heading ) { |
|
|
|
|
|
|
331
|
|
|
$heading = "### $heading\n"; |
|
332
|
|
|
} else { |
|
333
|
|
|
$heading = substr( $ret, -2 ) === "\n\n" ? '' : "\n"; |
|
334
|
|
|
} |
|
335
|
|
|
foreach ( $changes as $change ) { |
|
|
|
|
|
|
336
|
|
|
$text = trim( $change->getContent() ); |
|
337
|
|
|
if ( '' !== $text ) { |
|
338
|
|
|
if ( $change->getAuthor() !== '' ) { |
|
339
|
|
|
$text .= " ({$change->getAuthor()})"; |
|
340
|
|
|
} |
|
341
|
|
|
$ret .= $heading . $bullet . str_replace( "\n", "\n$indent", $text ) . "\n"; |
|
342
|
|
|
$heading = ''; |
|
343
|
|
|
} |
|
344
|
|
|
} |
|
345
|
|
|
$ret .= "\n"; |
|
346
|
|
|
} |
|
347
|
|
|
|
|
348
|
|
|
$epilogue = trim( $entry->getEpilogue() ); |
|
349
|
|
|
if ( '' !== $epilogue ) { |
|
350
|
|
|
$ret = trim( $ret ) . "\n\n$epilogue\n"; |
|
351
|
|
|
} |
|
352
|
|
|
$ret = trim( $ret ) . "\n\n"; |
|
353
|
|
|
} |
|
354
|
|
|
|
|
355
|
|
|
$epilogue = trim( $changelog->getEpilogue() ); |
|
356
|
|
|
if ( '' !== $epilogue ) { |
|
357
|
|
|
$ret .= "$epilogue\n"; |
|
358
|
|
|
} |
|
359
|
|
|
|
|
360
|
|
|
$ret = trim( $ret ) . "\n"; |
|
361
|
|
|
|
|
362
|
|
|
if ( $links ) { |
|
|
|
|
|
|
363
|
|
|
if ( ! $this->endsInLink( $ret ) ) { |
|
364
|
|
|
$ret .= "\n"; |
|
365
|
|
|
} |
|
366
|
|
|
foreach ( $links as $k => $v ) { |
|
367
|
|
|
$ret .= "[$k]: $v\n"; |
|
368
|
|
|
} |
|
369
|
|
|
} |
|
370
|
|
|
|
|
371
|
|
|
return $ret; |
|
372
|
|
|
} |
|
373
|
|
|
|
|
374
|
|
|
} |
|
375
|
|
|
|
This check looks for
@paramannotations 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.