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