tests/phpunit/maintenance/DumpTestCase.php

   1 <?php
   2
   3 /**
   4  * Base TestCase for dumps
   5  */
   6 abstract class DumpTestCase extends MediaWikiLangTestCase {
   7
   8         /**
   9          * exception to be rethrown once in sound PHPUnit surrounding
  10          *
  11          * As the current MediaWikiTestCase::run is not robust enough to recover
  12          * from thrown exceptions directly, we cannot throw frow within
  13          * self::addDBData, although it would be appropriate. Hence, we catch the
  14          * exception and store it until we are in setUp and may finally rethrow
  15          * the exception without crashing the test suite.
  16          *
  17          * @var Exception|null
  18          */
  19         protected $exceptionFromAddDBData = null;
  20
  21         /**
  22          * Holds the xmlreader used for analyzing an xml dump
  23          *
  24          * @var XMLReader|null
  25          */
  26         protected $xml = null;
  27
  28         /**
  29          * Adds a revision to a page, while returning the resuting revision's id
  30          *
  31          * @param Page $page Page to add the revision to
  32          * @param string $text Revisions text
  33          * @param string $summary Revisions summare
  34          * @return array
  35          * @throws MWException
  36          */
  37         protected function addRevision( Page $page, $text, $summary ) {
  38                 $status = $page->doEditContent(
  39                         ContentHandler::makeContent( $text, $page->getTitle() ),
  40                         $summary
  41                 );
  42
  43                 if ( $status->isGood() ) {
  44                         $value = $status->getValue();
  45                         $revision = $value['revision'];
  46                         $revision_id = $revision->getId();
  47                         $text_id = $revision->getTextId();
  48
  49                         if ( ( $revision_id > 0 ) && ( $text_id > 0 ) ) {
  50                                 return array( $revision_id, $text_id );
  51                         }
  52                 }
  53
  54                 throw new MWException( "Could not determine revision id (" . $status->getWikiText() . ")" );
  55         }
  56
  57         /**
  58          * gunzips the given file and stores the result in the original file name
  59          *
  60          * @param string $fname Filename to read the gzipped data from and stored
  61          *   the gunzipped data into
  62          */
  63         protected function gunzip( $fname ) {
  64                 $gzipped_contents = file_get_contents( $fname );
  65                 if ( $gzipped_contents === false ) {
  66                         $this->fail( "Could not get contents of $fname" );
  67                 }
  68
  69                 $contents = gzdecode( $gzipped_contents );
  70
  71                 $this->assertEquals(
  72                         strlen( $contents ),
  73                         file_put_contents( $fname, $contents ),
  74                         '# bytes written'
  75                 );
  76         }
  77
  78         /**
  79          * Default set up function.
  80          *
  81          * Clears $wgUser, and reports errors from addDBData to PHPUnit
  82          */
  83         protected function setUp() {
  84                 parent::setUp();
  85
  86                 // Check if any Exception is stored for rethrowing from addDBData
  87                 // @see self::exceptionFromAddDBData
  88                 if ( $this->exceptionFromAddDBData !== null ) {
  89                         throw $this->exceptionFromAddDBData;
  90                 }
  91
  92                 $this->setMwGlobals( 'wgUser', new User() );
  93         }
  94
  95         /**
  96          * Checks for test output consisting only of lines containing ETA announcements
  97          */
  98         function expectETAOutput() {
  99                 // Newer PHPUnits require assertion about the output using PHPUnit's own
 100                 // expectOutput[...] functions. However, the PHPUnit shipped prediactes
 101                 // do not allow to check /each/ line of the output using /readable/ REs.
 102                 // So we ...
 103                 //
 104                 // 1. ... add a dummy output checking to make PHPUnit not complain
 105                 //    about unchecked test output
 106                 $this->expectOutputRegex( '//' );
 107
 108                 // 2. Do the real output checking on our own.
 109                 $lines = explode( "\n", $this->getActualOutput() );
 110                 $this->assertGreaterThan( 1, count( $lines ), "Minimal lines of produced output" );
 111                 $this->assertEquals( '', array_pop( $lines ), "Output ends in LF" );
 112                 $timestamp_re = "[0-9]{4}-[01][0-9]-[0-3][0-9] [0-2][0-9]:[0-5][0-9]:[0-6][0-9]";
 113                 foreach ( $lines as $line ) {
 114                         $this->assertRegExp(
 115                                 "/$timestamp_re: .* \(ID [0-9]+\) [0-9]* pages .*, [0-9]* revs .*, ETA/",
 116                                 $line
 117                         );
 118                 }
 119         }
 120
 121         /**
 122          * Step the current XML reader until node end of given name is found.
 123          *
 124          * @param string $name Name of the closing element to look for
 125          *   (e.g.: "mediawiki" when looking for </mediawiki>)
 126          *
 127          * @return bool True if the end node could be found. false otherwise.
 128          */
 129         protected function skipToNodeEnd( $name ) {
 130                 while ( $this->xml->read() ) {
 131                         if ( $this->xml->nodeType == XMLReader::END_ELEMENT &&
 132                                 $this->xml->name == $name
 133                         ) {
 134                                 return true;
 135                         }
 136                 }
 137
 138                 return false;
 139         }
 140
 141         /**
 142          * Step the current XML reader to the first element start after the node
 143          * end of a given name.
 144          *
 145          * @param string $name Name of the closing element to look for
 146          *   (e.g.: "mediawiki" when looking for </mediawiki>)
 147          *
 148          * @return bool True if new element after the closing of $name could be
 149          *   found. false otherwise.
 150          */
 151         protected function skipPastNodeEnd( $name ) {
 152                 $this->assertTrue( $this->skipToNodeEnd( $name ),
 153                         "Skipping to end of $name" );
 154                 while ( $this->xml->read() ) {
 155                         if ( $this->xml->nodeType == XMLReader::ELEMENT ) {
 156                                 return true;
 157                         }
 158                 }
 159
 160                 return false;
 161         }
 162
 163         /**
 164          * Opens an XML file to analyze and optionally skips past siteinfo.
 165          *
 166          * @param string $fname Name of file to analyze
 167          * @param bool $skip_siteinfo (optional) If true, step the xml reader
 168          *   to the first element after </siteinfo>
 169          */
 170         protected function assertDumpStart( $fname, $skip_siteinfo = true ) {
 171                 $this->xml = new XMLReader();
 172                 $this->assertTrue( $this->xml->open( $fname ),
 173                         "Opening temporary file $fname via XMLReader failed" );
 174                 if ( $skip_siteinfo ) {
 175                         $this->assertTrue( $this->skipPastNodeEnd( "siteinfo" ),
 176                                 "Skipping past end of siteinfo" );
 177                 }
 178         }
 179
 180         /**
 181          * Asserts that the xml reader is at the final closing tag of an xml file and
 182          * closes the reader.
 183          *
 184          * @param string $name (optional) the name of the final tag
 185          *   (e.g.: "mediawiki" for </mediawiki>)
 186          */
 187         protected function assertDumpEnd( $name = "mediawiki" ) {
 188                 $this->assertNodeEnd( $name, false );
 189                 if ( $this->xml->read() ) {
 190                         $this->skipWhitespace();
 191                 }
 192                 $this->assertEquals( $this->xml->nodeType, XMLReader::NONE,
 193                         "No proper entity left to parse" );
 194                 $this->xml->close();
 195         }
 196
 197         /**
 198          * Steps the xml reader over white space
 199          */
 200         protected function skipWhitespace() {
 201                 $cont = true;
 202                 while ( $cont && ( ( $this->xml->nodeType == XMLReader::WHITESPACE )
 203                         || ( $this->xml->nodeType == XMLReader::SIGNIFICANT_WHITESPACE ) ) ) {
 204                         $cont = $this->xml->read();
 205                 }
 206         }
 207
 208         /**
 209          * Asserts that the xml reader is at an element of given name, and optionally
 210          * skips past it.
 211          *
 212          * @param string $name The name of the element to check for
 213          *   (e.g.: "mediawiki" for <mediawiki>)
 214          * @param bool $skip (optional) if true, skip past the found element
 215          */
 216         protected function assertNodeStart( $name, $skip = true ) {
 217                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 218                 $this->assertEquals( XMLReader::ELEMENT, $this->xml->nodeType, "Node type" );
 219                 if ( $skip ) {
 220                         $this->assertTrue( $this->xml->read(), "Skipping past start tag" );
 221                 }
 222         }
 223
 224         /**
 225          * Asserts that the xml reader is at an closing element of given name, and optionally
 226          * skips past it.
 227          *
 228          * @param string $name The name of the closing element to check for
 229          *   (e.g.: "mediawiki" for </mediawiki>)
 230          * @param bool $skip (optional) if true, skip past the found element
 231          */
 232         protected function assertNodeEnd( $name, $skip = true ) {
 233                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 234                 $this->assertEquals( XMLReader::END_ELEMENT, $this->xml->nodeType, "Node type" );
 235                 if ( $skip ) {
 236                         $this->assertTrue( $this->xml->read(), "Skipping past end tag" );
 237                 }
 238         }
 239
 240         /**
 241          * Asserts that the xml reader is at an element of given tag that contains a given text,
 242          * and skips over the element.
 243          *
 244          * @param string $name The name of the element to check for
 245          *   (e.g.: "mediawiki" for <mediawiki>...</mediawiki>)
 246          * @param string|bool $text If string, check if it equals the elements text.
 247          *   If false, ignore the element's text
 248          * @param bool $skip_ws (optional) if true, skip past white spaces that trail the
 249          *   closing element.
 250          */
 251         protected function assertTextNode( $name, $text, $skip_ws = true ) {
 252                 $this->assertNodeStart( $name );
 253
 254                 if ( $text !== false ) {
 255                         $this->assertEquals( $text, $this->xml->value, "Text of node " . $name );
 256                 }
 257                 $this->assertTrue( $this->xml->read(), "Skipping past processed text of " . $name );
 258                 $this->assertNodeEnd( $name );
 259
 260                 if ( $skip_ws ) {
 261                         $this->skipWhitespace();
 262                 }
 263         }
 264
 265         /**
 266          * Asserts that the xml reader is at the start of a page element and skips over the first
 267          * tags, after checking them.
 268          *
 269          * Besides the opening page element, this function also checks for and skips over the
 270          * title, ns, and id tags. Hence after this function, the xml reader is at the first
 271          * revision of the current page.
 272          *
 273          * @param int $id Id of the page to assert
 274          * @param int $ns Number of namespage to assert
 275          * @param string $name Title of the current page
 276          */
 277         protected function assertPageStart( $id, $ns, $name ) {
 278
 279                 $this->assertNodeStart( "page" );
 280                 $this->skipWhitespace();
 281
 282                 $this->assertTextNode( "title", $name );
 283                 $this->assertTextNode( "ns", $ns );
 284                 $this->assertTextNode( "id", $id );
 285         }
 286
 287         /**
 288          * Asserts that the xml reader is at the page's closing element and skips to the next
 289          * element.
 290          */
 291         protected function assertPageEnd() {
 292                 $this->assertNodeEnd( "page" );
 293                 $this->skipWhitespace();
 294         }
 295
 296         /**
 297          * Asserts that the xml reader is at a revision and checks its representation before
 298          * skipping over it.
 299          *
 300          * @param int $id Id of the revision
 301          * @param string $summary Summary of the revision
 302          * @param int $text_id Id of the revision's text
 303          * @param int $text_bytes Number of bytes in the revision's text
 304          * @param string $text_sha1 The base36 SHA-1 of the revision's text
 305          * @param string|bool $text (optional) The revision's string, or false to check for a
 306          *            revision stub
 307          * @param int|bool $parentid (optional) id of the parent revision
 308          * @param string $model The expected content model id (default: CONTENT_MODEL_WIKITEXT)
 309          * @param string $format The expected format model id (default: CONTENT_FORMAT_WIKITEXT)
 310          */
 311         protected function assertRevision( $id, $summary, $text_id, $text_bytes,
 312                 $text_sha1, $text = false, $parentid = false,
 313                 $model = CONTENT_MODEL_WIKITEXT, $format = CONTENT_FORMAT_WIKITEXT
 314         ) {
 315                 $this->assertNodeStart( "revision" );
 316                 $this->skipWhitespace();
 317
 318                 $this->assertTextNode( "id", $id );
 319                 if ( $parentid !== false ) {
 320                         $this->assertTextNode( "parentid", $parentid );
 321                 }
 322                 $this->assertTextNode( "timestamp", false );
 323
 324                 $this->assertNodeStart( "contributor" );
 325                 $this->skipWhitespace();
 326                 $this->assertTextNode( "ip", false );
 327                 $this->assertNodeEnd( "contributor" );
 328                 $this->skipWhitespace();
 329
 330                 $this->assertTextNode( "comment", $summary );
 331                 $this->skipWhitespace();
 332
 333                 $this->assertTextNode( "model", $model );
 334                 $this->skipWhitespace();
 335
 336                 $this->assertTextNode( "format", $format );
 337                 $this->skipWhitespace();
 338
 339                 if ( $this->xml->name == "text" ) {
 340                         // note: <text> tag may occur here or at the very end.
 341                         $text_found = true;
 342                         $this->assertText( $id, $text_id, $text_bytes, $text );
 343                 } else {
 344                         $text_found = false;
 345                 }
 346
 347                 $this->assertTextNode( "sha1", $text_sha1 );
 348
 349                 if ( !$text_found ) {
 350                         $this->assertText( $id, $text_id, $text_bytes, $text );
 351                 }
 352
 353                 $this->assertNodeEnd( "revision" );
 354                 $this->skipWhitespace();
 355         }
 356
 357         protected function assertText( $id, $text_id, $text_bytes, $text ) {
 358                 $this->assertNodeStart( "text", false );
 359                 if ( $text_bytes !== false ) {
 360                         $this->assertEquals( $this->xml->getAttribute( "bytes" ), $text_bytes,
 361                                 "Attribute 'bytes' of revision " . $id );
 362                 }
 363
 364                 if ( $text === false ) {
 365                         // Testing for a stub
 366                         $this->assertEquals( $this->xml->getAttribute( "id" ), $text_id,
 367                                 "Text id of revision " . $id );
 368                         $this->assertFalse( $this->xml->hasValue, "Revision has text" );
 369                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 370                         if ( ( $this->xml->nodeType == XMLReader::END_ELEMENT )
 371                                 && ( $this->xml->name == "text" )
 372                         ) {
 373
 374                                 $this->xml->read();
 375                         }
 376                         $this->skipWhitespace();
 377                 } else {
 378                         // Testing for a real dump
 379                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 380                         $this->assertEquals( $text, $this->xml->value, "Text of revision " . $id );
 381                         $this->assertTrue( $this->xml->read(), "Skipping past text" );
 382                         $this->assertNodeEnd( "text" );
 383                         $this->skipWhitespace();
 384                 }
 385         }
 386 }