31 require_once __DIR__ .
'/dumpIterator.php';
43 private $saveFailed =
false;
45 private $stripParametersEnabled;
47 private $showParsedOutput;
56 parent::__construct();
57 $this->
addDescription(
'Run a file or dump with several parsers' );
58 $this->
addOption(
'parser1',
'The first parser to compare.',
true,
true );
59 $this->
addOption(
'parser2',
'The second parser to compare.',
true,
true );
62 'Folder in which articles which differ will be stored.',
66 $this->
addOption(
'show-diff',
'Show a diff of the two renderings.',
false,
false );
69 'Binary to use for diffing (can also be provided by DIFF env var).',
75 'Remove parameters of html tags to increase readability.',
81 'Show the parsed html if both Parsers give the same output.',
88 if ( $this->
hasOption(
'save-failed' ) ) {
89 $this->saveFailed = $this->
getOption(
'save-failed' );
92 $this->stripParametersEnabled = $this->
hasOption(
'strip-parameters' );
93 $this->showParsedOutput = $this->
hasOption(
'show-parsed-output' );
95 $this->showDiff = $this->
hasOption(
'show-diff' );
96 if ( $this->showDiff ) {
97 $bin = $this->
getOption(
'diff-bin', getenv(
'DIFF' ) );
111 $this->
error(
"{$this->failed} failed revisions out of {$this->count}" );
112 if ( $this->count > 0 ) {
113 $this->
output(
" (" . ( $this->failed / $this->count ) .
"%)\n" );
117 private function stripParameters( $text ) {
118 if ( !$this->stripParametersEnabled ) {
122 return preg_replace(
'/(<a) [^>]+>/',
'$1>', $text );
132 $parser1Name = $this->
getOption(
'parser1' );
133 $parser2Name = $this->
getOption(
'parser2' );
135 self::checkParserLocally( $parser1Name );
136 self::checkParserLocally( $parser2Name );
138 $parser1 =
new $parser1Name();
139 $parser2 =
new $parser2Name();
144 $this->
error(
"Page {$title->getPrefixedText()} does not contain wikitext "
145 .
"but {$content->getModel()}\n" );
151 '@phan-var WikitextContent $content';
152 $text = strval(
$content->getText() );
154 $output1 = $parser1->parse( $text,
$title, $this->options );
155 $output2 = $parser2->parse( $text,
$title, $this->options );
157 if ( $output1->getText() != $output2->getText() ) {
159 $this->
error(
"Parsing for {$title->getPrefixedText()} differs\n" );
161 if ( $this->saveFailed ) {
163 $this->saveFailed .
'/' . rawurlencode(
$title->getPrefixedText() ) .
".txt",
167 if ( $this->showDiff ) {
169 explode(
"\n", $this->stripParameters( $output1->getText() ) ),
170 explode(
"\n", $this->stripParameters( $output2->getText() ) )
173 $unifiedDiff = $formatter->format( $diffs );
175 $this->
output( $unifiedDiff );
180 if ( $this->showParsedOutput ) {
181 $this->
output( $this->stripParameters( $output1->getText() ) );
186 private static function checkParserLocally( $parserName ) {
188 if ( !class_exists( $parserName ) && file_exists(
"$parserName.php" ) ) {
196 require_once RUN_MAINTENANCE_IF_MAIN;
const CONTENT_MODEL_WIKITEXT
Maintenance script to take page text out of an XML dump file and render basic HTML out to files.
conclusions()
Stub function for giving data about what was computed.
processRevision(WikiRevision $rev)
Callback function for each revision, parse with both parsers and compare.
__construct()
Default constructor.
checkOptions()
Stub function for processing additional options.
Class representing a 'diff' between two sequences of strings.
Base class for iterating over a dump.
error( $err, $die=0)
Throw an error to the user.
output( $out, $channel=null)
Throw some output to the user.
hasOption( $name)
Checks to see if a particular option was set.
addDescription( $text)
Set the description text.
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
getOption( $name, $default=null)
Get an option, or return the default.
static newFromUser( $user)
Get a ParserOptions object from a given user.
The User object encapsulates all of the user-specific settings (user_id, name, rights,...
Represents a revision, log entry or upload during the import process.
getContent( $role=SlotRecord::MAIN)
$wgDiff
Config variable stub for the Diff setting, for use by phpdoc and IDEs.