tests/phpunit/includes/tidy/BalancerTest.php

   1 <?php
   2
   3 class BalancerTest extends MediaWikiTestCase {
   4         private $balancer;
   5
   6         /**
   7          * Anything that needs to happen before your tests should go here.
   8          */
   9         protected function setUp() {
  10                 // Be sure to do call the parent setup and teardown functions.
  11                 // This makes sure that all the various cleanup and restorations
  12                 // happen as they should (including the restoration for setMwGlobals).
  13                 parent::setUp();
  14                 $this->balancer = new MediaWiki\Tidy\Balancer( [
  15                         'strict' => false, /* not strict */
  16                         'allowedHtmlElements' => null, /* no sanitization */
  17                         'tidyCompat' => false, /* standard parser */
  18                         'allowComments' => true, /* comment parsing */
  19                 ] );
  20         }
  21
  22         /**
  23          * Anything cleanup you need to do should go here.
  24          */
  25         protected function tearDown() {
  26                 parent::tearDown();
  27         }
  28
  29         /**
  30          * @covers Balancer::balance
  31          * @dataProvider provideBalancerTests
  32          */
  33         public function testBalancer( $description, $input, $expected ) {
  34                 $output = $this->balancer->balance( $input );
  35
  36                 // Ignore self-closing tags
  37                 $output = preg_replace( '/\s*\/>/', '>', $output );
  38
  39                 $this->assertEquals( $expected, $output, $description );
  40         }
  41
  42         public static function provideBalancerTests() {
  43                 // Get the tests from html5lib-tests.json
  44                 $json = json_decode( file_get_contents(
  45                         __DIR__ . '/html5lib-tests.json'
  46                 ), true );
  47                 // Munge this slightly into the format phpunit expects
  48                 // for providers, and filter out HTML constructs which
  49                 // the balancer doesn't support.
  50                 $tests = [];
  51                 $okre = "~ \A
  52                         (?i:<!DOCTYPE\ html>)?
  53                         <html><head></head><body>
  54                         .*
  55                         </body></html>
  56                 \z ~xs";
  57                 foreach ( $json as $filename => $cases ) {
  58                         foreach ( $cases as $case ) {
  59                                 $html = $case['document']['html'];
  60                                 if ( !preg_match( $okre, $html ) ) {
  61                                         // Skip tests which involve stuff in the <head> or
  62                                         // weird doctypes.
  63                                         continue;
  64                                 }
  65                                 // We used to do this:
  66                                 //   $html = substr( $html, strlen( $start ), -strlen( $end ) );
  67                                 // But now we use a different field in the test case,
  68                                 // which reports how domino would parse this case in a
  69                                 // no-quirks <body> context.  (The original test case may
  70                                 // have had a different context, or relied on quirks mode.)
  71                                 $html = $case['document']['noQuirksBodyHtml'];
  72                                 // Normalize case of SVG attributes.
  73                                 $html = str_replace( 'foreignObject', 'foreignobject', $html );
  74                                 // Normalize case of MathML attributes.
  75                                 $html = str_replace( 'definitionURL', 'definitionurl', $html );
  76
  77                                 if (
  78                                         isset( $case['document']['props']['comment'] ) &&
  79                                         preg_match( ',<!--[^>]*<,', $html )
  80                                 ) {
  81                                         // Skip tests which include HTML comments containing
  82                                         // the < character, which we don't support.
  83                                         continue;
  84                                 }
  85                                 if ( strpos( $case['data'], '<![CDATA[' ) !== false ) {
  86                                         // Skip tests involving <![CDATA[ ]]> quoting.
  87                                         continue;
  88                                 }
  89                                 if (
  90                                         stripos( $case['data'], '<!DOCTYPE' ) !== false &&
  91                                         stripos( $case['data'], '<!DOCTYPE html>' ) === false
  92                                 ) {
  93                                         // Skip tests involving unusual doctypes.
  94                                         continue;
  95                                 }
  96                                 $literalre = "~ <rdar: | <isindex | < /? (
  97                                         html | head | body | frame | frameset | plaintext
  98                                 ) > ~xi";
  99                                 if ( preg_match( $literalre, $case['data'] ) ) {
 100                                         // Skip tests involving some literal tags, which are
 101                                         // unsupported but don't show up in the expected output.
 102                                         continue;
 103                                 }
 104                                 if (
 105                                         isset( $case['document']['props']['tags']['iframe'] ) ||
 106                                         isset( $case['document']['props']['tags']['noembed'] ) ||
 107                                         isset( $case['document']['props']['tags']['noscript'] ) ||
 108                                         isset( $case['document']['props']['tags']['script'] ) ||
 109                                         isset( $case['document']['props']['tags']['svg script'] ) ||
 110                                         isset( $case['document']['props']['tags']['svg title'] ) ||
 111                                         isset( $case['document']['props']['tags']['title'] ) ||
 112                                         isset( $case['document']['props']['tags']['xmp'] )
 113                                 ) {
 114                                         // Skip tests with unsupported tags which *do* show
 115                                         // up in the expected output.
 116                                         continue;
 117                                 }
 118                                 if (
 119                                         $filename === 'entities01.dat' ||
 120                                         $filename === 'entities02.dat' ||
 121                                         preg_match( '/&([a-z]+|#x[0-9A-F]+);/i', $case['data'] ) ||
 122                                         preg_match( '/^(&|&#|&#X|&#x|&#45|&x-test|&AMP)$/', $case['data'] )
 123                                 ) {
 124                                         // Skip tests involving entity encoding.
 125                                         continue;
 126                                 }
 127                                 if (
 128                                         isset( $case['document']['props']['tagWithLt'] ) ||
 129                                         isset( $case['document']['props']['attrWithFunnyChar'] ) ||
 130                                         preg_match( ':^(</b test|<di|<foo bar=qux/>)$:', $case['data'] ) ||
 131                                         preg_match( ':</p<p>:', $case['data'] ) ||
 132                                         preg_match( ':<b &=&amp>|<p/x/y/z>:', $case['data'] )
 133                                 ) {
 134                                         // Skip tests with funny tag or attribute names,
 135                                         // which are really tests of the HTML tokenizer, not
 136                                         // the tree builder.
 137                                         continue;
 138                                 }
 139                                 if (
 140                                         preg_match( ':encoding=" text/html "|type=" hidden":', $case['data'] )
 141                                 ) {
 142                                         // The Sanitizer normalizes whitespace in attribute
 143                                         // values, which makes this test case invalid.
 144                                         continue;
 145                                 }
 146                                 if ( $filename === 'plain-text-unsafe.dat' ) {
 147                                         // Skip tests with ASCII null, etc.
 148                                         continue;
 149                                 }
 150                                 $data = preg_replace(
 151                                         '~<!DOCTYPE html>~i', '', $case['data']
 152                                 );
 153                                 $tests[] = [
 154                                         $filename, # use better description?
 155                                         $data,
 156                                         $html
 157                                 ];
 158                         }
 159                 }
 160                 return $tests;
 161         }
 162 }