/**
* Convert wikitext to HTML
* Do not call this function recursively.
*
* @param string $text text we want to parse
* @param $title Title object
* @param $options ParserOptions
* @param $linestart boolean
* @param $clearState boolean
* @param int $revid number to pass in {{REVISIONID}}
* @return ParserOutput a ParserOutput
*/
public function parse($text, Title $title, ParserOptions $options, $linestart = true, $clearState = true, $revid = null)
{
/**
* First pass--just handle <nowiki> sections, pass the rest off
* to internalParse() which does all the real work.
*/
global $wgUseTidy, $wgAlwaysUseTidy, $wgShowHostnames;
$fname = __METHOD__ . '-' . wfGetCaller();
wfProfileIn(__METHOD__);
wfProfileIn($fname);
$this->startParse($title, $options, self::OT_HTML, $clearState);
$this->mInputSize = strlen($text);
if ($this->mOptions->getEnableLimitReport()) {
$this->mOutput->resetParseStartTime();
}
# Remove the strip marker tag prefix from the input, if present.
if ($clearState) {
$text = str_replace($this->mUniqPrefix, '', $text);
}
$oldRevisionId = $this->mRevisionId;
$oldRevisionObject = $this->mRevisionObject;
$oldRevisionTimestamp = $this->mRevisionTimestamp;
$oldRevisionUser = $this->mRevisionUser;
$oldRevisionSize = $this->mRevisionSize;
if ($revid !== null) {
$this->mRevisionId = $revid;
$this->mRevisionObject = null;
$this->mRevisionTimestamp = null;
$this->mRevisionUser = null;
$this->mRevisionSize = null;
}
wfRunHooks('ParserBeforeStrip', array(&$this, &$text, &$this->mStripState));
# No more strip!
wfRunHooks('ParserAfterStrip', array(&$this, &$text, &$this->mStripState));
$text = $this->internalParse($text);
wfRunHooks('ParserAfterParse', array(&$this, &$text, &$this->mStripState));
$text = $this->mStripState->unstripGeneral($text);
# Clean up special characters, only run once, next-to-last before doBlockLevels
$fixtags = array('/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ', '/(\\302\\253) /' => '\\1 ', '/ (!\\s*important)/' => ' \\1');
$text = preg_replace(array_keys($fixtags), array_values($fixtags), $text);
$text = $this->doBlockLevels($text, $linestart);
$this->replaceLinkHolders($text);
/**
* The input doesn't get language converted if
* a) It's disabled
* b) Content isn't converted
* c) It's a conversion table
* d) it is an interface message (which is in the user language)
*/
if (!($options->getDisableContentConversion() || isset($this->mDoubleUnderscores['nocontentconvert']))) {
if (!$this->mOptions->getInterfaceMessage()) {
# The position of the convert() call should not be changed. it
# assumes that the links are all replaced and the only thing left
# is the <nowiki> mark.
$text = $this->getConverterLanguage()->convert($text);
}
}
/**
* A converted title will be provided in the output object if title and
* content conversion are enabled, the article text does not contain
* a conversion-suppressing double-underscore tag, and no
* {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
* automatic link conversion.
*/
if (!($options->getDisableTitleConversion() || isset($this->mDoubleUnderscores['nocontentconvert']) || isset($this->mDoubleUnderscores['notitleconvert']) || $this->mOutput->getDisplayTitle() !== false)) {
$convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
if ($convruletitle) {
$this->mOutput->setTitleText($convruletitle);
} else {
$titleText = $this->getConverterLanguage()->convertTitle($title);
$this->mOutput->setTitleText($titleText);
}
}
$text = $this->mStripState->unstripNoWiki($text);
wfRunHooks('ParserBeforeTidy', array(&$this, &$text));
$text = $this->replaceTransparentTags($text);
$text = $this->mStripState->unstripGeneral($text);
$text = Sanitizer::normalizeCharReferences($text);
if ($wgUseTidy && $this->mOptions->getTidy() || $wgAlwaysUseTidy) {
$text = MWTidy::tidy($text);
} else {
# attempt to sanitize at least some nesting problems
# (bug #2702 and quite a few others)
$tidyregs = array('/(<([bi])>)(<([bi])>)?([^<]*)(<\\/?a[^<]*>)([^<]*)(<\\/\\4>)?(<\\/\\2>)/' => '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9', '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\\/a>(.*)<\\/a>/' => '\\1\\2</a>\\3</a>\\1\\4</a>', '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\\/div>)([^<]*)(<\\/\\2>)/' => '\\1\\3<div\\5>\\6</div>\\8\\9', '/<([bi])><\\/\\1>/' => '');
$text = preg_replace(array_keys($tidyregs), array_values($tidyregs), $text);
}
if ($this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit()) {
$this->limitationWarn('expensive-parserfunction', $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit());
}
wfRunHooks('ParserAfterTidy', array(&$this, &$text));
# Information on include size limits, for the benefit of users who try to skirt them
if ($this->mOptions->getEnableLimitReport()) {
$max = $this->mOptions->getMaxIncludeSize();
$cpuTime = $this->mOutput->getTimeSinceStart('cpu');
if ($cpuTime !== null) {
$this->mOutput->setLimitReportData('limitreport-cputime', sprintf("%.3f", $cpuTime));
}
$wallTime = $this->mOutput->getTimeSinceStart('wall');
$this->mOutput->setLimitReportData('limitreport-walltime', sprintf("%.3f", $wallTime));
$this->mOutput->setLimitReportData('limitreport-ppvisitednodes', array($this->mPPNodeCount, $this->mOptions->getMaxPPNodeCount()));
$this->mOutput->setLimitReportData('limitreport-ppgeneratednodes', array($this->mGeneratedPPNodeCount, $this->mOptions->getMaxGeneratedPPNodeCount()));
$this->mOutput->setLimitReportData('limitreport-postexpandincludesize', array($this->mIncludeSizes['post-expand'], $max));
$this->mOutput->setLimitReportData('limitreport-templateargumentsize', array($this->mIncludeSizes['arg'], $max));
$this->mOutput->setLimitReportData('limitreport-expansiondepth', array($this->mHighestExpansionDepth, $this->mOptions->getMaxPPExpandDepth()));
$this->mOutput->setLimitReportData('limitreport-expensivefunctioncount', array($this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit()));
wfRunHooks('ParserLimitReportPrepare', array($this, $this->mOutput));
$limitReport = "NewPP limit report\n";
if ($wgShowHostnames) {
$limitReport .= 'Parsed by ' . wfHostname() . "\n";
}
foreach ($this->mOutput->getLimitReportData() as $key => $value) {
if (wfRunHooks('ParserLimitReportFormat', array($key, &$value, &$limitReport, false, false))) {
$keyMsg = wfMessage($key)->inLanguage('en')->useDatabase(false);
$valueMsg = wfMessage(array("{$key}-value-text", "{$key}-value"))->inLanguage('en')->useDatabase(false);
if (!$valueMsg->exists()) {
$valueMsg = new RawMessage('$1');
}
if (!$keyMsg->isDisabled() && !$valueMsg->isDisabled()) {
$valueMsg->params($value);
$limitReport .= "{$keyMsg->text()}: {$valueMsg->text()}\n";
}
}
}
// Since we're not really outputting HTML, decode the entities and
// then re-encode the things that need hiding inside HTML comments.
$limitReport = htmlspecialchars_decode($limitReport);
wfRunHooks('ParserLimitReport', array($this, &$limitReport));
// Sanitize for comment. Note '‐' in the replacement is U+2010,
// which looks much like the problematic '-'.
$limitReport = str_replace(array('-', '&'), array('‐', '&'), $limitReport);
$text .= "\n<!-- \n{$limitReport}-->\n";
if ($this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10) {
wfDebugLog('generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' . $this->mTitle->getPrefixedDBkey());
}
}
$this->mOutput->setText($text);
$this->mRevisionId = $oldRevisionId;
$this->mRevisionObject = $oldRevisionObject;
$this->mRevisionTimestamp = $oldRevisionTimestamp;
$this->mRevisionUser = $oldRevisionUser;
$this->mRevisionSize = $oldRevisionSize;
$this->mInputSize = false;
wfProfileOut($fname);
wfProfileOut(__METHOD__);
return $this->mOutput;
}
/**
* Convert wikitext to HTML
* Do not call this function recursively.
*
* @param $text String: text we want to parse
* @param $title Title object
* @param $options ParserOptions
* @param $linestart boolean
* @param $clearState boolean
* @param $revid Int: number to pass in {{REVISIONID}}
* @return ParserOutput a ParserOutput
*/
public function parse($text, Title $title, ParserOptions $options, $linestart = true, $clearState = true, $revid = null)
{
/**
* First pass--just handle <nowiki> sections, pass the rest off
* to internalParse() which does all the real work.
*/
global $wgUseTidy, $wgAlwaysUseTidy;
$fname = __METHOD__ . '-' . wfGetCaller();
wfProfileIn(__METHOD__);
wfProfileIn($fname);
$this->startParse($title, $options, self::OT_HTML, $clearState);
# Remove the strip marker tag prefix from the input, if present.
if ($clearState) {
$text = str_replace($this->mUniqPrefix, '', $text);
}
$oldRevisionId = $this->mRevisionId;
$oldRevisionObject = $this->mRevisionObject;
$oldRevisionTimestamp = $this->mRevisionTimestamp;
$oldRevisionUser = $this->mRevisionUser;
if ($revid !== null) {
$this->mRevisionId = $revid;
$this->mRevisionObject = null;
$this->mRevisionTimestamp = null;
$this->mRevisionUser = null;
}
wfRunHooks('ParserBeforeStrip', array(&$this, &$text, &$this->mStripState));
# No more strip!
wfRunHooks('ParserAfterStrip', array(&$this, &$text, &$this->mStripState));
$text = $this->internalParse($text);
wfRunHooks('ParserAfterParse', array(&$this, &$text, &$this->mStripState));
$text = $this->mStripState->unstripGeneral($text);
# Clean up special characters, only run once, next-to-last before doBlockLevels
$fixtags = array('/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ', '/(\\302\\253) /' => '\\1 ', '/ (!\\s*important)/' => ' \\1');
$text = preg_replace(array_keys($fixtags), array_values($fixtags), $text);
$text = $this->doBlockLevels($text, $linestart);
$this->replaceLinkHolders($text);
/**
* The input doesn't get language converted if
* a) It's disabled
* b) Content isn't converted
* c) It's a conversion table
* d) it is an interface message (which is in the user language)
*/
if (!($options->getDisableContentConversion() || isset($this->mDoubleUnderscores['nocontentconvert']))) {
# Run convert unconditionally in 1.18-compatible mode
global $wgBug34832TransitionalRollback;
if ($wgBug34832TransitionalRollback || !$this->mOptions->getInterfaceMessage()) {
# The position of the convert() call should not be changed. it
# assumes that the links are all replaced and the only thing left
# is the <nowiki> mark.
$text = $this->getConverterLanguage()->convert($text);
}
}
/**
* A converted title will be provided in the output object if title and
* content conversion are enabled, the article text does not contain
* a conversion-suppressing double-underscore tag, and no
* {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
* automatic link conversion.
*/
if (!($options->getDisableTitleConversion() || isset($this->mDoubleUnderscores['nocontentconvert']) || isset($this->mDoubleUnderscores['notitleconvert']) || $this->mOutput->getDisplayTitle() !== false)) {
$convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
if ($convruletitle) {
$this->mOutput->setTitleText($convruletitle);
} else {
$titleText = $this->getConverterLanguage()->convertTitle($title);
$this->mOutput->setTitleText($titleText);
}
}
$text = $this->mStripState->unstripNoWiki($text);
wfRunHooks('ParserBeforeTidy', array(&$this, &$text));
$text = $this->replaceTransparentTags($text);
$text = $this->mStripState->unstripGeneral($text);
$text = Sanitizer::normalizeCharReferences($text);
if ($wgUseTidy && $this->mOptions->getTidy() || $wgAlwaysUseTidy) {
$text = MWTidy::tidy($text);
} else {
# attempt to sanitize at least some nesting problems
# (bug #2702 and quite a few others)
$tidyregs = array('/(<([bi])>)(<([bi])>)?([^<]*)(<\\/?a[^<]*>)([^<]*)(<\\/\\4>)?(<\\/\\2>)/' => '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9', '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\\/a>(.*)<\\/a>/' => '\\1\\2</a>\\3</a>\\1\\4</a>', '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\\/div>)([^<]*)(<\\/\\2>)/' => '\\1\\3<div\\5>\\6</div>\\8\\9', '/<([bi])><\\/\\1>/' => '');
$text = preg_replace(array_keys($tidyregs), array_values($tidyregs), $text);
}
if ($this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit()) {
$this->limitationWarn('expensive-parserfunction', $this->mExpensiveFunctionCount, $this->mOptions->getExpensiveParserFunctionLimit());
}
wfRunHooks('ParserAfterTidy', array(&$this, &$text));
# Information on include size limits, for the benefit of users who try to skirt them
if ($this->mOptions->getEnableLimitReport()) {
$max = $this->mOptions->getMaxIncludeSize();
$PFreport = "Expensive parser function count: {$this->mExpensiveFunctionCount}/{$this->mOptions->getExpensiveParserFunctionLimit()}\n";
$limitReport = "NewPP limit report\n" . "Preprocessor visited node count: {$this->mPPNodeCount}/{$this->mOptions->getMaxPPNodeCount()}\n" . "Preprocessor generated node count: " . "{$this->mGeneratedPPNodeCount}/{$this->mOptions->getMaxGeneratedPPNodeCount()}\n" . "Post-expand include size: {$this->mIncludeSizes['post-expand']}/{$max} bytes\n" . "Template argument size: {$this->mIncludeSizes['arg']}/{$max} bytes\n" . "Highest expansion depth: {$this->mHighestExpansionDepth}/{$this->mOptions->getMaxPPExpandDepth()}\n" . $PFreport;
wfRunHooks('ParserLimitReport', array($this, &$limitReport));
$text .= "\n<!-- \n{$limitReport}-->\n";
if ($this->mGeneratedPPNodeCount > $this->mOptions->getMaxGeneratedPPNodeCount() / 10) {
wfDebugLog('generated-pp-node-count', $this->mGeneratedPPNodeCount . ' ' . $this->mTitle->getPrefixedDBkey());
}
}
$this->mOutput->setText($text);
$this->mRevisionId = $oldRevisionId;
$this->mRevisionObject = $oldRevisionObject;
$this->mRevisionTimestamp = $oldRevisionTimestamp;
$this->mRevisionUser = $oldRevisionUser;
wfProfileOut($fname);
wfProfileOut(__METHOD__);
return $this->mOutput;
}
/**
* Helper function for parse() that transforms half-parsed HTML into fully
* parsed HTML.
*
* @param string $text
* @param bool $isMain
* @param bool $linestart
* @return string
*/
private function internalParseHalfParsed($text, $isMain = true, $linestart = true)
{
$text = $this->mStripState->unstripGeneral($text);
if ($isMain) {
Hooks::run('ParserAfterUnstrip', array(&$this, &$text));
}
# Clean up special characters, only run once, next-to-last before doBlockLevels
$fixtags = array('/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1 ', '/(\\302\\253) /' => '\\1 ', '/ (!\\s*important)/' => ' \\1');
$text = preg_replace(array_keys($fixtags), array_values($fixtags), $text);
$text = $this->doBlockLevels($text, $linestart);
$this->replaceLinkHolders($text);
/**
* The input doesn't get language converted if
* a) It's disabled
* b) Content isn't converted
* c) It's a conversion table
* d) it is an interface message (which is in the user language)
*/
if (!($this->mOptions->getDisableContentConversion() || isset($this->mDoubleUnderscores['nocontentconvert']))) {
if (!$this->mOptions->getInterfaceMessage()) {
# The position of the convert() call should not be changed. it
# assumes that the links are all replaced and the only thing left
# is the <nowiki> mark.
$text = $this->getConverterLanguage()->convert($text);
}
}
$text = $this->mStripState->unstripNoWiki($text);
if ($isMain) {
Hooks::run('ParserBeforeTidy', array(&$this, &$text));
}
$text = $this->replaceTransparentTags($text);
$text = $this->mStripState->unstripGeneral($text);
$text = Sanitizer::normalizeCharReferences($text);
if (MWTidy::isEnabled() && $this->mOptions->getTidy()) {
$text = MWTidy::tidy($text);
} else {
# attempt to sanitize at least some nesting problems
# (bug #2702 and quite a few others)
$tidyregs = array('/(<([bi])>)(<([bi])>)?([^<]*)(<\\/?a[^<]*>)([^<]*)(<\\/\\4>)?(<\\/\\2>)/' => '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9', '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\\/a>(.*)<\\/a>/' => '\\1\\2</a>\\3</a>\\1\\4</a>', '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\\/div>)([^<]*)(<\\/\\2>)/' => '\\1\\3<div\\5>\\6</div>\\8\\9', '/<([bi])><\\/\\1>/' => '');
$text = preg_replace(array_keys($tidyregs), array_values($tidyregs), $text);
}
if ($isMain) {
Hooks::run('ParserAfterTidy', array(&$this, &$text));
}
return $text;
}
/**
* Parse the form definition and return it
*/
public static function getFormDefinition(Parser $parser, $form_def = null, $form_id = null)
{
if ($form_id !== null) {
$cachedDef = self::getFormDefinitionFromCache($form_id, $parser);
if ($cachedDef !== null) {
return $cachedDef;
}
}
if ($form_id !== null) {
$form_title = Title::newFromID($form_id);
$form_def = SFUtils::getPageText($form_title);
} elseif ($form_def == null) {
// No id, no text -> nothing to do
return '';
}
// Remove <noinclude> sections and <includeonly> tags from form definition
$form_def = StringUtils::delimiterReplace('<noinclude>', '</noinclude>', '', $form_def);
$form_def = strtr($form_def, array('<includeonly>' => '', '</includeonly>' => ''));
// We need to replace all SF tags in the form definition by strip items. But we can not just use
// the Parser strip state because the Parser would during parsing replace all strip items and then
// mangle them into HTML code. So we have to use our own. Which means we also can not just use
// Parser::insertStripItem() (see below).
$prefix = "UNIQ" . Parser::getRandomString();
$stripState = new StripState($prefix);
// This regexp will find any SF triple braced tags (including correct handling of contained braces), i.e.
// {{{field|foo|default={{Bar}}}}} is not a problem. When used with preg_match and friends, $matches[0] will
// contain the whole SF tag, $matches[1] will contain the tag without the enclosing triple braces.
$regexp = '#\\{\\{\\{((?>[^\\{\\}]+)|(\\{((?>[^\\{\\}]+)|(?-2))*\\}))*\\}\\}\\}#';
// replace all SF tags by strip markers
$form_def = preg_replace_callback($regexp, function (array $matches) use($stripState, $prefix) {
static $markerIndex = 0;
$rnd = "{$prefix}-item-{$markerIndex}-" . Parser::MARKER_SUFFIX;
$markerIndex++;
$stripState->addGeneral($rnd, $matches[0]);
return $rnd;
}, $form_def);
$title = is_object($parser->getTitle()) ? $parser->getTitle() : new Title();
// parse wiki-text
$output = $parser->parse($form_def, $title, $parser->getOptions());
$form_def = $stripState->unstripGeneral($output->getText());
if ($output->getCacheTime() == -1) {
$form_article = Article::newFromID($form_id);
self::purgeCache($form_article);
wfDebug("Caching disabled for form definition {$form_id}\n");
} elseif ($form_id !== null) {
self::cacheFormDefinition($form_id, $form_def, $parser);
}
return $form_def;
}
