Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
66.67% |
56 / 84 |
|
57.14% |
4 / 7 |
CRAP | |
0.00% |
0 / 1 |
| ParsoidOutputAccess | |
66.67% |
56 / 84 |
|
57.14% |
4 / 7 |
57.04 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
1 | |||
| supportsContentModel | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
4 | |||
| getParserOutput | |
58.33% |
7 / 12 |
|
0.00% |
0 / 1 |
5.16 | |||
| getCachedParserOutput | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| parseUncacheable | |
61.11% |
11 / 18 |
|
0.00% |
0 / 1 |
9.88 | |||
| resolveRevision | |
51.52% |
17 / 33 |
|
0.00% |
0 / 1 |
18.23 | |||
| adjustParserOptions | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Copyright (C) 2011-2022 Wikimedia Foundation and others. |
| 4 | * |
| 5 | * This program is free software; you can redistribute it and/or modify |
| 6 | * it under the terms of the GNU General Public License as published by |
| 7 | * the Free Software Foundation; either version 2 of the License, or |
| 8 | * (at your option) any later version. |
| 9 | * |
| 10 | * This program is distributed in the hope that it will be useful, |
| 11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 | * GNU General Public License for more details. |
| 14 | * |
| 15 | * You should have received a copy of the GNU General Public License along |
| 16 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 17 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. |
| 18 | */ |
| 19 | |
| 20 | namespace MediaWiki\Parser\Parsoid; |
| 21 | |
| 22 | use MediaWiki\Content\IContentHandlerFactory; |
| 23 | use MediaWiki\Page\PageIdentity; |
| 24 | use MediaWiki\Page\PageLookup; |
| 25 | use MediaWiki\Page\PageRecord; |
| 26 | use MediaWiki\Page\ParserOutputAccess; |
| 27 | use MediaWiki\Parser\ParserOutput; |
| 28 | use MediaWiki\Revision\RevisionAccessException; |
| 29 | use MediaWiki\Revision\RevisionLookup; |
| 30 | use MediaWiki\Revision\RevisionRecord; |
| 31 | use MediaWiki\Revision\SlotRecord; |
| 32 | use MediaWiki\Status\Status; |
| 33 | use MWUnknownContentModelException; |
| 34 | use ParserOptions; |
| 35 | use Wikimedia\Parsoid\Config\SiteConfig; |
| 36 | use Wikimedia\Parsoid\Core\ClientError; |
| 37 | use Wikimedia\Parsoid\Core\ResourceLimitExceededException; |
| 38 | |
| 39 | /** |
| 40 | * MediaWiki service for getting rendered page content. |
| 41 | * |
| 42 | * This is very similar to ParserOutputAccess and only exists as a |
| 43 | * separate class as an interim solution and should be removed soon. |
| 44 | * |
| 45 | * It is different from ParserOutputAccess in two aspects: |
| 46 | * - it forces Parsoid to be used when possible |
| 47 | * - it supports on-the-fly parsing through parseUncacheable() |
| 48 | * |
| 49 | * @since 1.39 |
| 50 | * @unstable |
| 51 | */ |
| 52 | class ParsoidOutputAccess { |
| 53 | private ParsoidParserFactory $parsoidParserFactory; |
| 54 | private PageLookup $pageLookup; |
| 55 | private RevisionLookup $revisionLookup; |
| 56 | private ParserOutputAccess $parserOutputAccess; |
| 57 | private SiteConfig $siteConfig; |
| 58 | private IContentHandlerFactory $contentHandlerFactory; |
| 59 | |
| 60 | /** |
| 61 | * @param ParsoidParserFactory $parsoidParserFactory |
| 62 | * @param ParserOutputAccess $parserOutputAccess |
| 63 | * @param PageLookup $pageLookup |
| 64 | * @param RevisionLookup $revisionLookup |
| 65 | * @param SiteConfig $siteConfig |
| 66 | * @param IContentHandlerFactory $contentHandlerFactory |
| 67 | */ |
| 68 | public function __construct( |
| 69 | ParsoidParserFactory $parsoidParserFactory, |
| 70 | ParserOutputAccess $parserOutputAccess, |
| 71 | PageLookup $pageLookup, |
| 72 | RevisionLookup $revisionLookup, |
| 73 | SiteConfig $siteConfig, |
| 74 | IContentHandlerFactory $contentHandlerFactory |
| 75 | ) { |
| 76 | $this->parsoidParserFactory = $parsoidParserFactory; |
| 77 | $this->parserOutputAccess = $parserOutputAccess; |
| 78 | $this->pageLookup = $pageLookup; |
| 79 | $this->revisionLookup = $revisionLookup; |
| 80 | $this->siteConfig = $siteConfig; |
| 81 | $this->contentHandlerFactory = $contentHandlerFactory; |
| 82 | } |
| 83 | |
| 84 | /** |
| 85 | * @param string $model |
| 86 | * |
| 87 | * @return bool |
| 88 | */ |
| 89 | public function supportsContentModel( string $model ): bool { |
| 90 | if ( $model === CONTENT_MODEL_WIKITEXT ) { |
| 91 | return true; |
| 92 | } |
| 93 | |
| 94 | // Check if the content model serializes to wikitext. |
| 95 | // NOTE: We could use isSupportedFormat( CONTENT_FORMAT_WIKITEXT ) if PageContent::getContent() |
| 96 | // would specify the format when calling serialize(). |
| 97 | try { |
| 98 | $handler = $this->contentHandlerFactory->getContentHandler( $model ); |
| 99 | if ( $handler->getDefaultFormat() === CONTENT_FORMAT_WIKITEXT ) { |
| 100 | return true; |
| 101 | } |
| 102 | } catch ( MWUnknownContentModelException $ex ) { |
| 103 | // If the content model is not known, it can't be supported. |
| 104 | return false; |
| 105 | } |
| 106 | |
| 107 | return $this->siteConfig->getContentModelHandler( $model ) !== null; |
| 108 | } |
| 109 | |
| 110 | /** |
| 111 | * @param PageIdentity $page |
| 112 | * @param ParserOptions $parserOpts |
| 113 | * @param RevisionRecord|int|null $revision |
| 114 | * @param int $options See the OPT_XXX constants |
| 115 | * @param bool $lenientRevHandling |
| 116 | * |
| 117 | * @return Status<ParserOutput> |
| 118 | */ |
| 119 | public function getParserOutput( |
| 120 | PageIdentity $page, |
| 121 | ParserOptions $parserOpts, |
| 122 | $revision = null, |
| 123 | int $options = 0, |
| 124 | bool $lenientRevHandling = false |
| 125 | ): Status { |
| 126 | [ $page, $revision, $uncacheable ] = $this->resolveRevision( $page, $revision, $lenientRevHandling ); |
| 127 | |
| 128 | try { |
| 129 | if ( $uncacheable ) { |
| 130 | $options |= ParserOutputAccess::OPT_NO_UPDATE_CACHE; |
| 131 | } |
| 132 | |
| 133 | $this->adjustParserOptions( $revision, $parserOpts ); |
| 134 | $status = $this->parserOutputAccess->getParserOutput( |
| 135 | $page, $parserOpts, $revision, $options |
| 136 | ); |
| 137 | } catch ( ClientError $e ) { |
| 138 | $status = Status::newFatal( 'parsoid-client-error', $e->getMessage() ); |
| 139 | } catch ( ResourceLimitExceededException $e ) { |
| 140 | $status = Status::newFatal( 'parsoid-resource-limit-exceeded', $e->getMessage() ); |
| 141 | } |
| 142 | return $status; |
| 143 | } |
| 144 | |
| 145 | /** |
| 146 | * @param PageIdentity $page |
| 147 | * @param ParserOptions $parserOpts |
| 148 | * @param RevisionRecord|int|null $revision |
| 149 | * @param bool $lenientRevHandling |
| 150 | * |
| 151 | * @return ?ParserOutput |
| 152 | */ |
| 153 | public function getCachedParserOutput( |
| 154 | PageIdentity $page, |
| 155 | ParserOptions $parserOpts, |
| 156 | $revision = null, |
| 157 | bool $lenientRevHandling = false |
| 158 | ): ?ParserOutput { |
| 159 | [ $page, $revision, $ignored ] = $this->resolveRevision( $page, $revision, $lenientRevHandling ); |
| 160 | |
| 161 | $this->adjustParserOptions( $revision, $parserOpts ); |
| 162 | return $this->parserOutputAccess->getCachedParserOutput( $page, $parserOpts, $revision ); |
| 163 | } |
| 164 | |
| 165 | /** |
| 166 | * This is to be called only for parsing posted wikitext that is actually |
| 167 | * not part of any real revision. |
| 168 | * |
| 169 | * @param PageIdentity $page |
| 170 | * @param ParserOptions $parserOpts |
| 171 | * @param RevisionRecord|int|null $revision |
| 172 | * @param bool $lenientRevHandling |
| 173 | * |
| 174 | * @return Status |
| 175 | */ |
| 176 | public function parseUncacheable( |
| 177 | PageIdentity $page, |
| 178 | ParserOptions $parserOpts, |
| 179 | $revision, |
| 180 | bool $lenientRevHandling = false |
| 181 | ): Status { |
| 182 | // NOTE: If we have a RevisionRecord already, just use it, there is no need to resolve $page to |
| 183 | // a PageRecord (and it may not be possible if the page doesn't exist). |
| 184 | if ( !$revision instanceof RevisionRecord ) { |
| 185 | [ $page, $revision, $ignored ] = $this->resolveRevision( $page, $revision, $lenientRevHandling ); |
| 186 | } |
| 187 | |
| 188 | // Enforce caller expectation |
| 189 | $revId = $revision->getId(); |
| 190 | if ( $revId !== 0 && $revId !== null ) { |
| 191 | return Status::newFatal( 'parsoid-revision-access', |
| 192 | "parseUncacheable should not be called for a real revision" ); |
| 193 | } |
| 194 | |
| 195 | try { |
| 196 | // Since we aren't caching this output, there is no need to |
| 197 | // call setUseParsoid() here. |
| 198 | $parser = $this->parsoidParserFactory->create(); |
| 199 | $parserOutput = $this->parsoidParserFactory->create()->parseFakeRevision( |
| 200 | $revision, $page, $parserOpts ); |
| 201 | $parserOutput->updateCacheExpiry( 0 ); // Ensure this isn't accidentally cached |
| 202 | $status = Status::newGood( $parserOutput ); |
| 203 | } catch ( RevisionAccessException $e ) { |
| 204 | return Status::newFatal( 'parsoid-revision-access', $e->getMessage() ); |
| 205 | } catch ( ClientError $e ) { |
| 206 | $status = Status::newFatal( 'parsoid-client-error', $e->getMessage() ); |
| 207 | } catch ( ResourceLimitExceededException $e ) { |
| 208 | $status = Status::newFatal( 'parsoid-resource-limit-exceeded', $e->getMessage() ); |
| 209 | } |
| 210 | return $status; |
| 211 | } |
| 212 | |
| 213 | /** |
| 214 | * @param PageIdentity $page |
| 215 | * @param RevisionRecord|int|null $revision |
| 216 | * @param bool $lenientRevHandling |
| 217 | * |
| 218 | * @return array [ PageRecord $page, RevisionRecord $revision ] |
| 219 | */ |
| 220 | private function resolveRevision( PageIdentity $page, $revision, bool $lenientRevHandling = false ): array { |
| 221 | $uncacheable = false; |
| 222 | if ( !$page instanceof PageRecord ) { |
| 223 | $name = "$page"; |
| 224 | $page = $this->pageLookup->getPageByReference( $page ); |
| 225 | if ( !$page ) { |
| 226 | throw new RevisionAccessException( |
| 227 | 'Page {name} not found', |
| 228 | [ 'name' => $name ] |
| 229 | ); |
| 230 | } |
| 231 | } |
| 232 | |
| 233 | if ( $revision === null ) { |
| 234 | $revision = $page->getLatest(); |
| 235 | } |
| 236 | |
| 237 | if ( is_int( $revision ) ) { |
| 238 | $revId = $revision; |
| 239 | $revision = $this->revisionLookup->getRevisionById( $revId ); |
| 240 | |
| 241 | if ( !$revision ) { |
| 242 | throw new RevisionAccessException( |
| 243 | 'Revision {revId} not found', |
| 244 | [ 'revId' => $revId ] |
| 245 | ); |
| 246 | } |
| 247 | } |
| 248 | |
| 249 | if ( $page->getId() !== $revision->getPageId() ) { |
| 250 | if ( $lenientRevHandling ) { |
| 251 | $page = $this->pageLookup->getPageById( $revision->getPageId() ); |
| 252 | if ( !$page ) { |
| 253 | // This should ideally never trigger! |
| 254 | throw new \RuntimeException( |
| 255 | "Unexpected NULL page for pageid " . $revision->getPageId() . |
| 256 | " from revision " . $revision->getId() |
| 257 | ); |
| 258 | } |
| 259 | // Don't cache this! |
| 260 | $uncacheable = true; |
| 261 | } else { |
| 262 | throw new RevisionAccessException( |
| 263 | 'Revision {revId} does not belong to page {name}', |
| 264 | [ 'name' => $page->getDBkey(), 'revId' => $revision->getId() ] |
| 265 | ); |
| 266 | } |
| 267 | } |
| 268 | |
| 269 | return [ $page, $revision, $uncacheable ]; |
| 270 | } |
| 271 | |
| 272 | private function adjustParserOptions( RevisionRecord $revision, ParserOptions $parserOpts ): void { |
| 273 | $mainSlot = $revision->getSlot( SlotRecord::MAIN ); |
| 274 | $contentModel = $mainSlot->getModel(); |
| 275 | if ( $this->supportsContentModel( $contentModel ) ) { |
| 276 | // Since we know Parsoid supports this content model, explicitly |
| 277 | // call ParserOptions::setUseParsoid. This ensures that when |
| 278 | // we query the parser-cache, the right cache key is called. |
| 279 | // This is an optional transition step to using ParserOutputAccess. |
| 280 | $parserOpts->setUseParsoid(); |
| 281 | } |
| 282 | } |
| 283 | } |