Merge "Live Preview: Cope with the edit summary being an OOjs UI widget"
[lhc/web/wiklou.git] / tests / phpunit / includes / SanitizerTest.php
1 <?php
2
3 /**
4 * @todo Tests covering decodeCharReferences can be refactored into a single
5 * method and dataprovider.
6 *
7 * @group Sanitizer
8 */
9 class SanitizerTest extends MediaWikiTestCase {
10
11 protected function tearDown() {
12 MWTidy::destroySingleton();
13 parent::tearDown();
14 }
15
16 /**
17 * @covers Sanitizer::decodeCharReferences
18 */
19 public function testDecodeNamedEntities() {
20 $this->assertEquals(
21 "\xc3\xa9cole",
22 Sanitizer::decodeCharReferences( '&eacute;cole' ),
23 'decode named entities'
24 );
25 }
26
27 /**
28 * @covers Sanitizer::decodeCharReferences
29 */
30 public function testDecodeNumericEntities() {
31 $this->assertEquals(
32 "\xc4\x88io bonas dans l'\xc3\xa9cole!",
33 Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&#233;cole!" ),
34 'decode numeric entities'
35 );
36 }
37
38 /**
39 * @covers Sanitizer::decodeCharReferences
40 */
41 public function testDecodeMixedEntities() {
42 $this->assertEquals(
43 "\xc4\x88io bonas dans l'\xc3\xa9cole!",
44 Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&eacute;cole!" ),
45 'decode mixed numeric/named entities'
46 );
47 }
48
49 /**
50 * @covers Sanitizer::decodeCharReferences
51 */
52 public function testDecodeMixedComplexEntities() {
53 $this->assertEquals(
54 "\xc4\x88io bonas dans l'\xc3\xa9cole! (mais pas &#x108;io dans l'&eacute;cole)",
55 Sanitizer::decodeCharReferences(
56 "&#x108;io bonas dans l'&eacute;cole! (mais pas &amp;#x108;io dans l'&#38;eacute;cole)"
57 ),
58 'decode mixed complex entities'
59 );
60 }
61
62 /**
63 * @covers Sanitizer::decodeCharReferences
64 */
65 public function testInvalidAmpersand() {
66 $this->assertEquals(
67 'a & b',
68 Sanitizer::decodeCharReferences( 'a & b' ),
69 'Invalid ampersand'
70 );
71 }
72
73 /**
74 * @covers Sanitizer::decodeCharReferences
75 */
76 public function testInvalidEntities() {
77 $this->assertEquals(
78 '&foo;',
79 Sanitizer::decodeCharReferences( '&foo;' ),
80 'Invalid named entity'
81 );
82 }
83
84 /**
85 * @covers Sanitizer::decodeCharReferences
86 */
87 public function testInvalidNumberedEntities() {
88 $this->assertEquals(
89 UtfNormal\Constants::UTF8_REPLACEMENT,
90 Sanitizer::decodeCharReferences( "&#88888888888888;" ),
91 'Invalid numbered entity'
92 );
93 }
94
95 /**
96 * @covers Sanitizer::removeHTMLtags
97 * @dataProvider provideHtml5Tags
98 *
99 * @param string $tag Name of an HTML5 element (ie: 'video')
100 * @param bool $escaped Whether sanitizer let the tag in or escape it (ie: '&lt;video&gt;')
101 */
102 public function testRemovehtmltagsOnHtml5Tags( $tag, $escaped ) {
103 MWTidy::setInstance( false );
104
105 if ( $escaped ) {
106 $this->assertEquals( "&lt;$tag&gt;",
107 Sanitizer::removeHTMLtags( "<$tag>" )
108 );
109 } else {
110 $this->assertEquals( "<$tag></$tag>\n",
111 Sanitizer::removeHTMLtags( "<$tag>" )
112 );
113 }
114 }
115
116 /**
117 * Provide HTML5 tags
118 */
119 public static function provideHtml5Tags() {
120 $ESCAPED = true; # We want tag to be escaped
121 $VERBATIM = false; # We want to keep the tag
122 return [
123 [ 'data', $VERBATIM ],
124 [ 'mark', $VERBATIM ],
125 [ 'time', $VERBATIM ],
126 [ 'video', $ESCAPED ],
127 ];
128 }
129
130 function dataRemoveHTMLtags() {
131 return [
132 // former testSelfClosingTag
133 [
134 '<div>Hello world</div />',
135 '<div>Hello world</div>',
136 'Self-closing closing div'
137 ],
138 // Make sure special nested HTML5 semantics are not broken
139 // https://html.spec.whatwg.org/multipage/semantics.html#the-kbd-element
140 [
141 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
142 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
143 'Nested <kbd>.'
144 ],
145 // https://html.spec.whatwg.org/multipage/semantics.html#the-sub-and-sup-elements
146 [
147 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
148 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
149 'Nested <var>.'
150 ],
151 // https://html.spec.whatwg.org/multipage/semantics.html#the-dfn-element
152 [
153 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
154 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
155 '<abbr> inside <dfn>',
156 ],
157 ];
158 }
159
160 /**
161 * @dataProvider dataRemoveHTMLtags
162 * @covers Sanitizer::removeHTMLtags
163 */
164 public function testRemoveHTMLtags( $input, $output, $msg = null ) {
165 MWTidy::setInstance( false );
166 $this->assertEquals( $output, Sanitizer::removeHTMLtags( $input ), $msg );
167 }
168
169 /**
170 * @dataProvider provideTagAttributesToDecode
171 * @covers Sanitizer::decodeTagAttributes
172 */
173 public function testDecodeTagAttributes( $expected, $attributes, $message = '' ) {
174 $this->assertEquals( $expected,
175 Sanitizer::decodeTagAttributes( $attributes ),
176 $message
177 );
178 }
179
180 public static function provideTagAttributesToDecode() {
181 return [
182 [ [ 'foo' => 'bar' ], 'foo=bar', 'Unquoted attribute' ],
183 [ [ 'עברית' => 'bar' ], 'עברית=bar', 'Non-Latin attribute' ],
184 [ [ '६' => 'bar' ], '६=bar', 'Devanagari number' ],
185 [ [ '搭𨋢' => 'bar' ], '搭𨋢=bar', 'Non-BMP character' ],
186 [ [], 'ńgh=bar', 'Combining accent is not allowed' ],
187 [ [ 'foo' => 'bar' ], ' foo = bar ', 'Spaced attribute' ],
188 [ [ 'foo' => 'bar' ], 'foo="bar"', 'Double-quoted attribute' ],
189 [ [ 'foo' => 'bar' ], 'foo=\'bar\'', 'Single-quoted attribute' ],
190 [
191 [ 'foo' => 'bar', 'baz' => 'foo' ],
192 'foo=\'bar\' baz="foo"',
193 'Several attributes'
194 ],
195 [
196 [ 'foo' => 'bar', 'baz' => 'foo' ],
197 'foo=\'bar\' baz="foo"',
198 'Several attributes'
199 ],
200 [
201 [ 'foo' => 'bar', 'baz' => 'foo' ],
202 'foo=\'bar\' baz="foo"',
203 'Several attributes'
204 ],
205 [ [ ':foo' => 'bar' ], ':foo=\'bar\'', 'Leading :' ],
206 [ [ '_foo' => 'bar' ], '_foo=\'bar\'', 'Leading _' ],
207 [ [ 'foo' => 'bar' ], 'Foo=\'bar\'', 'Leading capital' ],
208 [ [ 'foo' => 'BAR' ], 'FOO=BAR', 'Attribute keys are normalized to lowercase' ],
209
210 # Invalid beginning
211 [ [], '-foo=bar', 'Leading - is forbidden' ],
212 [ [], '.foo=bar', 'Leading . is forbidden' ],
213 [ [ 'foo-bar' => 'bar' ], 'foo-bar=bar', 'A - is allowed inside the attribute' ],
214 [ [ 'foo-' => 'bar' ], 'foo-=bar', 'A - is allowed inside the attribute' ],
215 [ [ 'foo.bar' => 'baz' ], 'foo.bar=baz', 'A . is allowed inside the attribute' ],
216 [ [ 'foo.' => 'baz' ], 'foo.=baz', 'A . is allowed as last character' ],
217 [ [ 'foo6' => 'baz' ], 'foo6=baz', 'Numbers are allowed' ],
218
219 # This bit is more relaxed than XML rules, but some extensions use
220 # it, like ProofreadPage (see T29539)
221 [ [ '1foo' => 'baz' ], '1foo=baz', 'Leading numbers are allowed' ],
222 [ [], 'foo$=baz', 'Symbols are not allowed' ],
223 [ [], 'foo@=baz', 'Symbols are not allowed' ],
224 [ [], 'foo~=baz', 'Symbols are not allowed' ],
225 [
226 [ 'foo' => '1[#^`*%w/(' ],
227 'foo=1[#^`*%w/(',
228 'All kind of characters are allowed as values'
229 ],
230 [
231 [ 'foo' => '1[#^`*%\'w/(' ],
232 'foo="1[#^`*%\'w/("',
233 'Double quotes are allowed if quoted by single quotes'
234 ],
235 [
236 [ 'foo' => '1[#^`*%"w/(' ],
237 'foo=\'1[#^`*%"w/(\'',
238 'Single quotes are allowed if quoted by double quotes'
239 ],
240 [ [ 'foo' => '&"' ], 'foo=&amp;&quot;', 'Special chars can be provided as entities' ],
241 [ [ 'foo' => '&foobar;' ], 'foo=&foobar;', 'Entity-like items are accepted' ],
242 ];
243 }
244
245 /**
246 * @dataProvider provideDeprecatedAttributes
247 * @covers Sanitizer::fixTagAttributes
248 */
249 public function testDeprecatedAttributesUnaltered( $inputAttr, $inputEl, $message = '' ) {
250 $this->assertEquals( " $inputAttr",
251 Sanitizer::fixTagAttributes( $inputAttr, $inputEl ),
252 $message
253 );
254 }
255
256 public static function provideDeprecatedAttributes() {
257 /** [ <attribute>, <element>, [message] ] */
258 return [
259 [ 'clear="left"', 'br' ],
260 [ 'clear="all"', 'br' ],
261 [ 'width="100"', 'td' ],
262 [ 'nowrap="true"', 'td' ],
263 [ 'nowrap=""', 'td' ],
264 [ 'align="right"', 'td' ],
265 [ 'align="center"', 'table' ],
266 [ 'align="left"', 'tr' ],
267 [ 'align="center"', 'div' ],
268 [ 'align="left"', 'h1' ],
269 [ 'align="left"', 'p' ],
270 ];
271 }
272
273 /**
274 * @dataProvider provideCssCommentsFixtures
275 * @covers Sanitizer::checkCss
276 */
277 public function testCssCommentsChecking( $expected, $css, $message = '' ) {
278 $this->assertEquals( $expected,
279 Sanitizer::checkCss( $css ),
280 $message
281 );
282 }
283
284 public static function provideCssCommentsFixtures() {
285 /** [ <expected>, <css>, [message] ] */
286 return [
287 // Valid comments spanning entire input
288 [ '/**/', '/**/' ],
289 [ '/* comment */', '/* comment */' ],
290 // Weird stuff
291 [ ' ', '/****/' ],
292 [ ' ', '/* /* */' ],
293 [ 'display: block;', "display:/* foo */block;" ],
294 [ 'display: block;', "display:\\2f\\2a foo \\2a\\2f block;",
295 'Backslash-escaped comments must be stripped (T30450)' ],
296 [ '', '/* unfinished comment structure',
297 'Remove anything after a comment-start token' ],
298 [ '', "\\2f\\2a unifinished comment'",
299 'Remove anything after a backslash-escaped comment-start token' ],
300 [
301 '/* insecure input */',
302 'filter: progid:DXImageTransform.Microsoft.AlphaImageLoader'
303 . '(src=\'asdf.png\',sizingMethod=\'scale\');'
304 ],
305 [
306 '/* insecure input */',
307 '-ms-filter: "progid:DXImageTransform.Microsoft.AlphaImageLoader'
308 . '(src=\'asdf.png\',sizingMethod=\'scale\')";'
309 ],
310 [ '/* insecure input */', 'width: expression(1+1);' ],
311 [ '/* insecure input */', 'background-image: image(asdf.png);' ],
312 [ '/* insecure input */', 'background-image: -webkit-image(asdf.png);' ],
313 [ '/* insecure input */', 'background-image: -moz-image(asdf.png);' ],
314 [ '/* insecure input */', 'background-image: image-set("asdf.png" 1x, "asdf.png" 2x);' ],
315 [
316 '/* insecure input */',
317 'background-image: -webkit-image-set("asdf.png" 1x, "asdf.png" 2x);'
318 ],
319 [
320 '/* insecure input */',
321 'background-image: -moz-image-set("asdf.png" 1x, "asdf.png" 2x);'
322 ],
323 [ '/* insecure input */', 'foo: attr( title, url );' ],
324 [ '/* insecure input */', 'foo: attr( title url );' ],
325 ];
326 }
327
328 /**
329 * @dataProvider provideEscapeHtmlAllowEntities
330 * @covers Sanitizer::escapeHtmlAllowEntities
331 */
332 public function testEscapeHtmlAllowEntities( $expected, $html ) {
333 $this->assertEquals(
334 $expected,
335 Sanitizer::escapeHtmlAllowEntities( $html )
336 );
337 }
338
339 public static function provideEscapeHtmlAllowEntities() {
340 return [
341 [ 'foo', 'foo' ],
342 [ 'a¡b', 'a&#161;b' ],
343 [ 'foo&#039;bar', "foo'bar" ],
344 [ '&lt;script&gt;foo&lt;/script&gt;', '<script>foo</script>' ],
345 ];
346 }
347
348 /**
349 * Test Sanitizer::escapeId
350 *
351 * @dataProvider provideEscapeId
352 * @covers Sanitizer::escapeId
353 */
354 public function testEscapeId( $input, $output ) {
355 $this->assertEquals(
356 $output,
357 Sanitizer::escapeId( $input, [ 'noninitial', 'legacy' ] )
358 );
359 }
360
361 public static function provideEscapeId() {
362 return [
363 [ '+', '.2B' ],
364 [ '&', '.26' ],
365 [ '=', '.3D' ],
366 [ ':', ':' ],
367 [ ';', '.3B' ],
368 [ '@', '.40' ],
369 [ '$', '.24' ],
370 [ '-_.', '-_.' ],
371 [ '!', '.21' ],
372 [ '*', '.2A' ],
373 [ '/', '.2F' ],
374 [ '[]', '.5B.5D' ],
375 [ '<>', '.3C.3E' ],
376 [ '\'', '.27' ],
377 [ '§', '.C2.A7' ],
378 [ 'Test:A & B/Here', 'Test:A_.26_B.2FHere' ],
379 [ 'A&B&amp;C&amp;amp;D&amp;amp;amp;E', 'A.26B.26C.26amp.3BD.26amp.3Bamp.3BE' ],
380 ];
381 }
382
383 /**
384 * Test escapeIdReferenceList for consistency with escapeIdForAttribute
385 *
386 * @dataProvider provideEscapeIdReferenceList
387 * @covers Sanitizer::escapeIdReferenceList
388 */
389 public function testEscapeIdReferenceList( $referenceList, $id1, $id2 ) {
390 $this->assertEquals(
391 Sanitizer::escapeIdReferenceList( $referenceList, 'noninitial' ),
392 Sanitizer::escapeIdForAttribute( $id1 )
393 . ' '
394 . Sanitizer::escapeIdForAttribute( $id2 )
395 );
396 }
397
398 public static function provideEscapeIdReferenceList() {
399 /** [ <reference list>, <individual id 1>, <individual id 2> ] */
400 return [
401 [ 'foo bar', 'foo', 'bar' ],
402 [ '#1 #2', '#1', '#2' ],
403 [ '+1 +2', '+1', '+2' ],
404 ];
405 }
406
407 /**
408 * @dataProvider provideIsReservedDataAttribute
409 */
410 public function testIsReservedDataAttribute( $attr, $expected ) {
411 $this->assertSame( $expected, Sanitizer::isReservedDataAttribute( $attr ) );
412 }
413
414 public static function provideIsReservedDataAttribute() {
415 return [
416 [ 'foo', false ],
417 [ 'data', false ],
418 [ 'data-foo', false ],
419 [ 'data-mw', true ],
420 [ 'data-ooui', true ],
421 [ 'data-parsoid', true ],
422 [ 'data-mw-foo', true ],
423 [ 'data-ooui-foo', true ],
424 [ 'data-mwfoo', true ], // could be false but this is how it's implemented currently
425 ];
426 }
427
428 /**
429 * @dataProvider provideEscapeIdForStuff
430 *
431 * @covers Sanitizer::escapeIdForAttribute()
432 * @covers Sanitizer::escapeIdForLink()
433 * @covers Sanitizer::escapeIdForExternalInterwiki()
434 * @covers Sanitizer::escapeIdInternal()
435 * @covers Sanitizer::urlEscapeId()
436 *
437 * @param string $stuff
438 * @param string[] $config
439 * @param string $id
440 * @param string|false $expected
441 * @param int|null $mode
442 */
443 public function testEscapeIdForStuff( $stuff, array $config, $id, $expected, $mode = null ) {
444 $func = "Sanitizer::escapeIdFor{$stuff}";
445 $iwFlavor = array_pop( $config );
446 $this->setMwGlobals( [
447 'wgFragmentMode' => $config,
448 'wgExternalInterwikiFragmentMode' => $iwFlavor,
449 ] );
450 $escaped = call_user_func( $func, $id, $mode );
451 self::assertEquals( $expected, $escaped );
452 }
453
454 public function provideEscapeIdForStuff() {
455 // Test inputs and outputs
456 $text = 'foo тест_#%!\'()[]:<>';
457 $legacyEncoded = 'foo_.D1.82.D0.B5.D1.81.D1.82_.23.25.21.27.28.29.5B.5D:.3C.3E';
458 $html5Encoded = 'foo_тест_#%!\'()[]:<>';
459 $html5Escaped = 'foo_%D1%82%D0%B5%D1%81%D1%82_%23%25%21%27%28%29%5B%5D:%3C%3E';
460 $html5Experimental = 'foo_тест_!_()[]:<>';
461
462 // Settings: last element is $wgExternalInterwikiFragmentMode, the rest is $wgFragmentMode
463 $legacy = [ 'legacy', 'legacy' ];
464 $legacyNew = [ 'legacy', 'html5', 'legacy' ];
465 $newLegacy = [ 'html5', 'legacy', 'legacy' ];
466 $new = [ 'html5', 'legacy' ];
467 $allNew = [ 'html5', 'html5' ];
468 $experimentalLegacy = [ 'html5-legacy', 'legacy', 'legacy' ];
469 $newExperimental = [ 'html5', 'html5-legacy', 'legacy' ];
470
471 return [
472 // Pure legacy: how MW worked before 2017
473 [ 'Attribute', $legacy, $text, $legacyEncoded, Sanitizer::ID_PRIMARY ],
474 [ 'Attribute', $legacy, $text, false, Sanitizer::ID_FALLBACK ],
475 [ 'Link', $legacy, $text, $legacyEncoded ],
476 [ 'ExternalInterwiki', $legacy, $text, $legacyEncoded ],
477
478 // Transition to a new world: legacy links with HTML5 fallback
479 [ 'Attribute', $legacyNew, $text, $legacyEncoded, Sanitizer::ID_PRIMARY ],
480 [ 'Attribute', $legacyNew, $text, $html5Encoded, Sanitizer::ID_FALLBACK ],
481 [ 'Link', $legacyNew, $text, $legacyEncoded ],
482 [ 'ExternalInterwiki', $legacyNew, $text, $legacyEncoded ],
483
484 // New world: HTML5 links, legacy fallbacks
485 [ 'Attribute', $newLegacy, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
486 [ 'Attribute', $newLegacy, $text, $legacyEncoded, Sanitizer::ID_FALLBACK ],
487 [ 'Link', $newLegacy, $text, $html5Escaped ],
488 [ 'ExternalInterwiki', $newLegacy, $text, $legacyEncoded ],
489
490 // Distant future: no legacy fallbacks, but still linking to leagacy wikis
491 [ 'Attribute', $new, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
492 [ 'Attribute', $new, $text, false, Sanitizer::ID_FALLBACK ],
493 [ 'Link', $new, $text, $html5Escaped ],
494 [ 'ExternalInterwiki', $new, $text, $legacyEncoded ],
495
496 // Just before the heat death of universe: external interwikis are also HTML5 \m/
497 [ 'Attribute', $allNew, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
498 [ 'Attribute', $allNew, $text, false, Sanitizer::ID_FALLBACK ],
499 [ 'Link', $allNew, $text, $html5Escaped ],
500 [ 'ExternalInterwiki', $allNew, $text, $html5Escaped ],
501
502 // Someone flipped $wgExperimentalHtmlIds on
503 [ 'Attribute', $experimentalLegacy, $text, $html5Experimental, Sanitizer::ID_PRIMARY ],
504 [ 'Attribute', $experimentalLegacy, $text, $legacyEncoded, Sanitizer::ID_FALLBACK ],
505 [ 'Link', $experimentalLegacy, $text, $html5Experimental ],
506 [ 'ExternalInterwiki', $experimentalLegacy, $text, $legacyEncoded ],
507
508 // Migration from $wgExperimentalHtmlIds to modern HTML5
509 [ 'Attribute', $newExperimental, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
510 [ 'Attribute', $newExperimental, $text, $html5Experimental, Sanitizer::ID_FALLBACK ],
511 [ 'Link', $newExperimental, $text, $html5Escaped ],
512 [ 'ExternalInterwiki', $newExperimental, $text, $legacyEncoded ],
513 ];
514 }
515
516 /**
517 * @expectedException InvalidArgumentException
518 * @covers Sanitizer::escapeIdInternal()
519 */
520 public function testInvalidFragmentThrows() {
521 $this->setMwGlobals( 'wgFragmentMode', [ 'boom!' ] );
522 Sanitizer::escapeIdForAttribute( 'This should throw' );
523 }
524
525 /**
526 * @expectedException UnexpectedValueException
527 * @covers Sanitizer::escapeIdForAttribute()
528 */
529 public function testNoPrimaryFragmentModeThrows() {
530 $this->setMwGlobals( 'wgFragmentMode', [ 666 => 'html5' ] );
531 Sanitizer::escapeIdForAttribute( 'This should throw' );
532 }
533
534 /**
535 * @expectedException UnexpectedValueException
536 * @covers Sanitizer::escapeIdForLink()
537 */
538 public function testNoPrimaryFragmentModeThrows2() {
539 $this->setMwGlobals( 'wgFragmentMode', [ 666 => 'html5' ] );
540 Sanitizer::escapeIdForLink( 'This should throw' );
541 }
542 }