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                 // We resort to use gzinflate instead of gzdecode, as gzdecode
  64                 // need not be available
  65                 $contents = gzinflate( substr( $gzipped_contents, 10, -8 ) );
  66                 $this->assertEquals( strlen( $contents ),
  67                         file_put_contents( $fname, $contents ), "# bytes written" );
  68         }
  69
  70         /**
  71          * Default set up function.
  72          *
  73          * Clears $wgUser, and reports errors from addDBData to PHPUnit
  74          */
  75         protected function setUp() {
  76                 parent::setUp();
  77
  78                 // Check if any Exception is stored for rethrowing from addDBData
  79                 // @see self::exceptionFromAddDBData
  80                 if ( $this->exceptionFromAddDBData !== null ) {
  81                         throw $this->exceptionFromAddDBData;
  82                 }
  83
  84                 $this->setMwGlobals( 'wgUser', new User() );
  85         }
  86
  87         /**
  88          * Checks for test output consisting only of lines containing ETA announcements
  89          */
  90         function expectETAOutput() {
  91                 // Newer PHPUnits require assertion about the output using PHPUnit's own
  92                 // expectOutput[...] functions. However, the PHPUnit shipped prediactes
  93                 // do not allow to check /each/ line of the output using /readable/ REs.
  94                 // So we ...
  95                 //
  96                 // 1. ... add a dummy output checking to make PHPUnit not complain
  97                 //    about unchecked test output
  98                 $this->expectOutputRegex( '//' );
  99
 100                 // 2. Do the real output checking on our own.
 101                 $lines = explode( "\n", $this->getActualOutput() );
 102                 $this->assertGreaterThan( 1, count( $lines ), "Minimal lines of produced output" );
 103                 $this->assertEquals( '', array_pop( $lines ), "Output ends in LF" );
 104                 $timestamp_re = "[0-9]{4}-[01][0-9]-[0-3][0-9] [0-2][0-9]:[0-5][0-9]:[0-6][0-9]";
 105                 foreach ( $lines as $line ) {
 106                         $this->assertRegExp( "/$timestamp_re: .* \(ID [0-9]+\) [0-9]* pages .*, [0-9]* revs .*, ETA/", $line );
 107                 }
 108         }
 109
 110
 111         /**
 112          * Step the current XML reader until node end of given name is found.
 113          *
 114          * @param $name string: name of the closing element to look for
 115          *           (e.g.: "mediawiki" when looking for </mediawiki>)
 116          *
 117          * @return bool: true iff the end node could be found. false otherwise.
 118          */
 119         protected function skipToNodeEnd( $name ) {
 120                 while ( $this->xml->read() ) {
 121                         if ( $this->xml->nodeType == XMLReader::END_ELEMENT &&
 122                                 $this->xml->name == $name
 123                         ) {
 124                                 return true;
 125                         }
 126                 }
 127                 return false;
 128         }
 129
 130         /**
 131          * Step the current XML reader to the first element start after the node
 132          * end of a given name.
 133          *
 134          * @param $name string: name of the closing element to look for
 135          *           (e.g.: "mediawiki" when looking for </mediawiki>)
 136          *
 137          * @return bool: true iff new element after the closing of $name could be
 138          *           found. false otherwise.
 139          */
 140         protected function skipPastNodeEnd( $name ) {
 141                 $this->assertTrue( $this->skipToNodeEnd( $name ),
 142                         "Skipping to end of $name" );
 143                 while ( $this->xml->read() ) {
 144                         if ( $this->xml->nodeType == XMLReader::ELEMENT ) {
 145                                 return true;
 146                         }
 147                 }
 148                 return false;
 149         }
 150
 151         /**
 152          * Opens an XML file to analyze and optionally skips past siteinfo.
 153          *
 154          * @param $fname string: name of file to analyze
 155          * @param $skip_siteinfo bool: (optional) If true, step the xml reader
 156          *           to the first element after </siteinfo>
 157          */
 158         protected function assertDumpStart( $fname, $skip_siteinfo = true ) {
 159                 $this->xml = new XMLReader();
 160                 $this->assertTrue( $this->xml->open( $fname ),
 161                         "Opening temporary file $fname via XMLReader failed" );
 162                 if ( $skip_siteinfo ) {
 163                         $this->assertTrue( $this->skipPastNodeEnd( "siteinfo" ),
 164                                 "Skipping past end of siteinfo" );
 165                 }
 166         }
 167
 168         /**
 169          * Asserts that the xml reader is at the final closing tag of an xml file and
 170          * closes the reader.
 171          *
 172          * @param $tag string: (optional) the name of the final tag
 173          *           (e.g.: "mediawiki" for </mediawiki>)
 174          */
 175         protected function assertDumpEnd( $name = "mediawiki" ) {
 176                 $this->assertNodeEnd( $name, false );
 177                 if ( $this->xml->read() ) {
 178                         $this->skipWhitespace();
 179                 }
 180                 $this->assertEquals( $this->xml->nodeType, XMLReader::NONE,
 181                         "No proper entity left to parse" );
 182                 $this->xml->close();
 183         }
 184
 185         /**
 186          * Steps the xml reader over white space
 187          */
 188         protected function skipWhitespace() {
 189                 $cont = true;
 190                 while ( $cont && ( ( $this->xml->nodeType == XMLReader::WHITESPACE )
 191                         || ( $this->xml->nodeType == XMLReader::SIGNIFICANT_WHITESPACE ) ) ) {
 192                         $cont = $this->xml->read();
 193                 }
 194         }
 195
 196         /**
 197          * Asserts that the xml reader is at an element of given name, and optionally
 198          * skips past it.
 199          *
 200          * @param $name string: the name of the element to check for
 201          *           (e.g.: "mediawiki" for <mediawiki>)
 202          * @param $skip bool: (optional) if true, skip past the found element
 203          */
 204         protected function assertNodeStart( $name, $skip = true ) {
 205                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 206                 $this->assertEquals( XMLReader::ELEMENT, $this->xml->nodeType, "Node type" );
 207                 if ( $skip ) {
 208                         $this->assertTrue( $this->xml->read(), "Skipping past start tag" );
 209                 }
 210         }
 211
 212         /**
 213          * Asserts that the xml reader is at an closing element of given name, and optionally
 214          * skips past it.
 215          *
 216          * @param $name string: the name of the closing element to check for
 217          *           (e.g.: "mediawiki" for </mediawiki>)
 218          * @param $skip bool: (optional) if true, skip past the found element
 219          */
 220         protected function assertNodeEnd( $name, $skip = true ) {
 221                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 222                 $this->assertEquals( XMLReader::END_ELEMENT, $this->xml->nodeType, "Node type" );
 223                 if ( $skip ) {
 224                         $this->assertTrue( $this->xml->read(), "Skipping past end tag" );
 225                 }
 226         }
 227
 228
 229         /**
 230          * Asserts that the xml reader is at an element of given tag that contains a given text,
 231          * and skips over the element.
 232          *
 233          * @param $name string: the name of the element to check for
 234          *           (e.g.: "mediawiki" for <mediawiki>...</mediawiki>)
 235          * @param $text string|false: If string, check if it equals the elements text.
 236          *           If false, ignore the element's text
 237          * @param $skip_ws bool: (optional) if true, skip past white spaces that trail the
 238          *           closing element.
 239          */
 240         protected function assertTextNode( $name, $text, $skip_ws = true ) {
 241                 $this->assertNodeStart( $name );
 242
 243                 if ( $text !== false ) {
 244                         $this->assertEquals( $text, $this->xml->value, "Text of node " . $name );
 245                 }
 246                 $this->assertTrue( $this->xml->read(), "Skipping past processed text of " . $name );
 247                 $this->assertNodeEnd( $name );
 248
 249                 if ( $skip_ws ) {
 250                         $this->skipWhitespace();
 251                 }
 252         }
 253
 254         /**
 255          * Asserts that the xml reader is at the start of a page element and skips over the first
 256          * tags, after checking them.
 257          *
 258          * Besides the opening page element, this function also checks for and skips over the
 259          * title, ns, and id tags. Hence after this function, the xml reader is at the first
 260          * revision of the current page.
 261          *
 262          * @param $id int: id of the page to assert
 263          * @param $ns int: number of namespage to assert
 264          * @param $name string: title of the current page
 265          */
 266         protected function assertPageStart( $id, $ns, $name ) {
 267
 268                 $this->assertNodeStart( "page" );
 269                 $this->skipWhitespace();
 270
 271                 $this->assertTextNode( "title", $name );
 272                 $this->assertTextNode( "ns", $ns );
 273                 $this->assertTextNode( "id", $id );
 274
 275         }
 276
 277         /**
 278          * Asserts that the xml reader is at the page's closing element and skips to the next
 279          * element.
 280          */
 281         protected function assertPageEnd() {
 282                 $this->assertNodeEnd( "page" );
 283                 $this->skipWhitespace();
 284         }
 285
 286         /**
 287          * Asserts that the xml reader is at a revision and checks its representation before
 288          * skipping over it.
 289          *
 290          * @param $id int: id of the revision
 291          * @param $summary string: summary of the revision
 292          * @param $text_id int: id of the revision's text
 293          * @param $text_bytes int: # of bytes in the revision's text
 294          * @param $text_sha1 string: the base36 SHA-1 of the revision's text
 295          * @param $text string|false: (optional) The revision's string, or false to check for a
 296          *            revision stub
 297          * @param $model String: the expected content model id (default: CONTENT_MODEL_WIKITEXT)
 298          * @param $format String: the expected format model id (default: CONTENT_FORMAT_WIKITEXT)
 299          * @param $parentid int|false: (optional) id of the parent revision
 300          */
 301         protected function assertRevision( $id, $summary, $text_id, $text_bytes, $text_sha1, $text = false, $parentid = false,
 302                                                                            $model = CONTENT_MODEL_WIKITEXT, $format = CONTENT_FORMAT_WIKITEXT ) {
 303
 304                 $this->assertNodeStart( "revision" );
 305                 $this->skipWhitespace();
 306
 307                 $this->assertTextNode( "id", $id );
 308                 if ( $parentid !== false ) {
 309                         $this->assertTextNode( "parentid", $parentid );
 310                 }
 311                 $this->assertTextNode( "timestamp", false );
 312
 313                 $this->assertNodeStart( "contributor" );
 314                 $this->skipWhitespace();
 315                 $this->assertTextNode( "ip", false );
 316                 $this->assertNodeEnd( "contributor" );
 317                 $this->skipWhitespace();
 318
 319                 $this->assertTextNode( "comment", $summary );
 320                 $this->skipWhitespace();
 321
 322                 if ( $this->xml->name == "text" ) {
 323                         // note: <text> tag may occur here or at the very end.
 324                         $text_found = true;
 325                         $this->assertText( $id, $text_id, $text_bytes, $text );
 326                 } else {
 327                         $text_found = false;
 328                 }
 329
 330                 $this->assertTextNode( "sha1", $text_sha1 );
 331
 332                 $this->assertTextNode( "model", $model );
 333                 $this->skipWhitespace();
 334
 335                 $this->assertTextNode( "format", $format );
 336                 $this->skipWhitespace();
 337
 338                 if ( !$text_found ) {
 339                         $this->assertText( $id, $text_id, $text_bytes, $text );
 340                 }
 341
 342                 $this->assertNodeEnd( "revision" );
 343                 $this->skipWhitespace();
 344         }
 345
 346         protected function assertText( $id, $text_id, $text_bytes, $text ) {
 347                 $this->assertNodeStart( "text", false );
 348                 if ( $text_bytes !== false ) {
 349                         $this->assertEquals( $this->xml->getAttribute( "bytes" ), $text_bytes,
 350                                 "Attribute 'bytes' of revision " . $id );
 351                 }
 352
 353                 if ( $text === false ) {
 354                         // Testing for a stub
 355                         $this->assertEquals( $this->xml->getAttribute( "id" ), $text_id,
 356                                 "Text id of revision " . $id );
 357                         $this->assertFalse( $this->xml->hasValue, "Revision has text" );
 358                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 359                         if ( ( $this->xml->nodeType == XMLReader::END_ELEMENT )
 360                                 && ( $this->xml->name == "text" )
 361                         ) {
 362
 363                                 $this->xml->read();
 364                         }
 365                         $this->skipWhitespace();
 366                 } else {
 367                         // Testing for a real dump
 368                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 369                         $this->assertEquals( $text, $this->xml->value, "Text of revision " . $id );
 370                         $this->assertTrue( $this->xml->read(), "Skipping past text" );
 371                         $this->assertNodeEnd( "text" );
 372                         $this->skipWhitespace();
 373                 }
 374         }
 375 }