Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
78.72% |
111 / 141 |
|
16.67% |
1 / 6 |
CRAP | |
0.00% |
0 / 1 |
| TitleCleanup | |
78.72% |
111 / 141 |
|
16.67% |
1 / 6 |
70.19 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
1 | |||
| execute | |
80.00% |
4 / 5 |
|
0.00% |
0 / 1 |
5.20 | |||
| processRow | |
89.47% |
17 / 19 |
|
0.00% |
0 / 1 |
8.07 | |||
| fileExists | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
2 | |||
| moveIllegalPage | |
75.44% |
43 / 57 |
|
0.00% |
0 / 1 |
21.28 | |||
| moveInconsistentPage | |
89.36% |
42 / 47 |
|
0.00% |
0 / 1 |
16.31 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Clean up broken, unparseable titles. |
| 4 | * |
| 5 | * Copyright © 2005 Brooke Vibber <bvibber@wikimedia.org> |
| 6 | * https://www.mediawiki.org/ |
| 7 | * |
| 8 | * This program is free software; you can redistribute it and/or modify |
| 9 | * it under the terms of the GNU General Public License as published by |
| 10 | * the Free Software Foundation; either version 2 of the License, or |
| 11 | * (at your option) any later version. |
| 12 | * |
| 13 | * This program is distributed in the hope that it will be useful, |
| 14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 16 | * GNU General Public License for more details. |
| 17 | * |
| 18 | * You should have received a copy of the GNU General Public License along |
| 19 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 20 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 21 | * http://www.gnu.org/copyleft/gpl.html |
| 22 | * |
| 23 | * @file |
| 24 | * @author Brooke Vibber <bvibber@wikimedia.org> |
| 25 | * @ingroup Maintenance |
| 26 | */ |
| 27 | |
| 28 | use MediaWiki\Title\Title; |
| 29 | use Wikimedia\Rdbms\IDBAccessObject; |
| 30 | |
| 31 | // @codeCoverageIgnoreStart |
| 32 | require_once __DIR__ . '/TableCleanup.php'; |
| 33 | // @codeCoverageIgnoreEnd |
| 34 | |
| 35 | /** |
| 36 | * Maintenance script to clean up broken, unparseable titles. |
| 37 | * |
| 38 | * @ingroup Maintenance |
| 39 | */ |
| 40 | class TitleCleanup extends TableCleanup { |
| 41 | |
| 42 | private string $prefix; |
| 43 | |
| 44 | public function __construct() { |
| 45 | parent::__construct(); |
| 46 | $this->addDescription( 'Script to clean up broken, unparseable titles' ); |
| 47 | $this->addOption( 'prefix', "Broken pages will be renamed to titles with " . |
| 48 | "<prefix> prepended before the article name. Defaults to 'Broken'", false, true ); |
| 49 | $this->setBatchSize( 1000 ); |
| 50 | } |
| 51 | |
| 52 | /** |
| 53 | * @inheritDoc |
| 54 | */ |
| 55 | public function execute() { |
| 56 | $this->prefix = $this->getOption( 'prefix', 'Broken' ) . "/"; |
| 57 | // Make sure the prefix itself is a valid title now |
| 58 | // rather than spewing errors for every page being cleaned up |
| 59 | // if it's not (We assume below that concatenating the prefix to a title leaves it in NS0) |
| 60 | // The trailing slash above ensures that concatenating the title to something |
| 61 | // can't turn it into a namespace or interwiki |
| 62 | $title = Title::newFromText( $this->prefix ); |
| 63 | if ( !$title || !$title->canExist() || $title->getInterwiki() || $title->getNamespace() !== 0 ) { |
| 64 | $this->fatalError( "Invalid prefix {$this->prefix}. Must be a valid mainspace title." ); |
| 65 | } |
| 66 | parent::execute(); |
| 67 | } |
| 68 | |
| 69 | /** |
| 70 | * @param stdClass $row |
| 71 | */ |
| 72 | protected function processRow( $row ) { |
| 73 | $display = Title::makeName( $row->page_namespace, $row->page_title ); |
| 74 | $verified = $this->getServiceContainer()->getContentLanguage()->normalize( $display ); |
| 75 | $title = Title::newFromText( $verified ); |
| 76 | |
| 77 | if ( $title !== null |
| 78 | && $title->canExist() |
| 79 | && $title->getNamespace() == $row->page_namespace |
| 80 | && $title->getDBkey() === $row->page_title |
| 81 | ) { |
| 82 | // all is fine |
| 83 | $this->progress( 0 ); |
| 84 | |
| 85 | return; |
| 86 | } |
| 87 | |
| 88 | if ( $row->page_namespace == NS_FILE && $this->fileExists( $row->page_title ) ) { |
| 89 | $this->output( "file $row->page_title needs cleanup, please run cleanupImages.php.\n" ); |
| 90 | $this->progress( 0 ); |
| 91 | } elseif ( $title === null ) { |
| 92 | $this->output( "page $row->page_id ($display) is illegal.\n" ); |
| 93 | $this->moveIllegalPage( $row ); |
| 94 | $this->progress( 1 ); |
| 95 | } else { |
| 96 | $this->output( "page $row->page_id ($display) doesn't match self.\n" ); |
| 97 | $this->moveInconsistentPage( $row, $title ); |
| 98 | $this->progress( 1 ); |
| 99 | } |
| 100 | } |
| 101 | |
| 102 | /** |
| 103 | * @param string $name |
| 104 | * @return bool |
| 105 | */ |
| 106 | protected function fileExists( $name ) { |
| 107 | // XXX: Doesn't actually check for file existence, just presence of image record. |
| 108 | // This is reasonable, since cleanupImages.php only iterates over the image table. |
| 109 | $dbr = $this->getReplicaDB(); |
| 110 | $row = $dbr->newSelectQueryBuilder() |
| 111 | ->select( '*' ) |
| 112 | ->from( 'image' ) |
| 113 | ->where( [ 'img_name' => $name ] ) |
| 114 | ->caller( __METHOD__ ) |
| 115 | ->fetchRow(); |
| 116 | |
| 117 | return $row !== false; |
| 118 | } |
| 119 | |
| 120 | /** |
| 121 | * @param stdClass $row |
| 122 | */ |
| 123 | protected function moveIllegalPage( $row ) { |
| 124 | $legalChars = Title::legalChars(); |
| 125 | $legalizedUnprefixed = preg_replace_callback( "/([^$legalChars])/", |
| 126 | [ $this, 'hexChar' ], |
| 127 | $row->page_title ); |
| 128 | if ( $legalizedUnprefixed == '.' ) { |
| 129 | $legalizedUnprefixed = '(dot)'; |
| 130 | } |
| 131 | if ( $legalizedUnprefixed == '_' ) { |
| 132 | $legalizedUnprefixed = '(space)'; |
| 133 | } |
| 134 | $ns = (int)$row->page_namespace; |
| 135 | |
| 136 | $title = null; |
| 137 | // Try to move "Talk:Project:Foo" -> "Project talk:Foo" |
| 138 | if ( $ns === 1 ) { |
| 139 | $subjectTitle = Title::newFromText( $legalizedUnprefixed ); |
| 140 | if ( $subjectTitle && !$subjectTitle->isTalkPage() ) { |
| 141 | $talkTitle = $subjectTitle->getTalkPageIfDefined(); |
| 142 | if ( $talkTitle !== null && !$talkTitle->exists() ) { |
| 143 | $ns = $talkTitle->getNamespace(); |
| 144 | $title = $talkTitle; |
| 145 | } |
| 146 | } |
| 147 | } |
| 148 | |
| 149 | if ( $title === null ) { |
| 150 | // Not a talk page or that didn't work |
| 151 | // move any other broken pages to the main namespace so they can be found together |
| 152 | if ( $ns !== 0 ) { |
| 153 | $namespaceInfo = $this->getServiceContainer()->getNamespaceInfo(); |
| 154 | $namespaceName = $namespaceInfo->getCanonicalName( $ns ); |
| 155 | if ( $namespaceName === false ) { |
| 156 | $namespaceName = "NS$ns"; // Fallback for unknown namespaces |
| 157 | } |
| 158 | $ns = 0; |
| 159 | $legalizedUnprefixed = "$namespaceName:$legalizedUnprefixed"; |
| 160 | } |
| 161 | $title = Title::newFromText( $this->prefix . $legalizedUnprefixed ); |
| 162 | } |
| 163 | |
| 164 | if ( $title === null ) { |
| 165 | // It's still not a valid title, try again with a much smaller |
| 166 | // allowed character set. This will mangle any titles with non-ASCII |
| 167 | // characters, but if we don't do this the result will be |
| 168 | // falling back to the Broken/id:foo failsafe below which is worse |
| 169 | $legalizedUnprefixed = preg_replace_callback( '!([^A-Za-z0-9_:\\-])!', |
| 170 | [ $this, 'hexChar' ], |
| 171 | $legalizedUnprefixed |
| 172 | ); |
| 173 | $title = Title::newFromText( $this->prefix . $legalizedUnprefixed ); |
| 174 | } |
| 175 | |
| 176 | if ( $title === null ) { |
| 177 | // Oh well, we tried |
| 178 | $clean = $this->prefix . 'id:' . $row->page_id; |
| 179 | $legalized = $this->prefix . $legalizedUnprefixed; |
| 180 | $this->output( "Couldn't legalize; form '$legalized' still invalid; using '$clean'\n" ); |
| 181 | $title = Title::newFromText( $clean ); |
| 182 | } elseif ( $title->exists( IDBAccessObject::READ_LATEST ) ) { |
| 183 | $clean = $this->prefix . 'id:' . $row->page_id; |
| 184 | $conflict = $title->getDBKey(); |
| 185 | $this->output( "Legalized for '$conflict' exists; using '$clean'\n" ); |
| 186 | $title = Title::newFromText( $clean ); |
| 187 | } |
| 188 | |
| 189 | if ( !$title || $title->exists( IDBAccessObject::READ_LATEST ) ) { |
| 190 | // This can happen in corner cases like if numbers are made not valid |
| 191 | // title characters using the (deprecated) $wgLegalTitleChars or |
| 192 | // a 'Broken/id:foo' title already exists |
| 193 | $this->error( "Destination page {$title->getText()} is invalid or already exists, skipping." ); |
| 194 | return; |
| 195 | } |
| 196 | |
| 197 | $dest = $title->getDBkey(); |
| 198 | if ( $this->dryrun ) { |
| 199 | $this->output( "DRY RUN: would rename $row->page_id ($row->page_namespace," . |
| 200 | "'$row->page_title') to ($ns,'$dest')\n" ); |
| 201 | } else { |
| 202 | $this->output( "renaming $row->page_id ($row->page_namespace," . |
| 203 | "'$row->page_title') to ($ns,'$dest')\n" ); |
| 204 | $this->getPrimaryDB() |
| 205 | ->newUpdateQueryBuilder() |
| 206 | ->update( 'page' ) |
| 207 | ->set( [ 'page_title' => $dest, 'page_namespace' => $ns ] ) |
| 208 | ->where( [ 'page_id' => $row->page_id ] ) |
| 209 | ->caller( __METHOD__ )->execute(); |
| 210 | } |
| 211 | } |
| 212 | |
| 213 | /** |
| 214 | * @param stdClass $row |
| 215 | * @param Title $title |
| 216 | */ |
| 217 | protected function moveInconsistentPage( $row, Title $title ) { |
| 218 | $titleImpossible = $title->getInterwiki() || !$title->canExist(); |
| 219 | if ( $title->exists( IDBAccessObject::READ_LATEST ) || $titleImpossible ) { |
| 220 | if ( $titleImpossible ) { |
| 221 | $prior = $title->getPrefixedDBkey(); |
| 222 | } else { |
| 223 | $prior = $title->getDBkey(); |
| 224 | } |
| 225 | |
| 226 | $ns = (int)$row->page_namespace; |
| 227 | # If a page is saved in the main namespace with a namespace prefix then try to move it into |
| 228 | # that namespace. If there's no conflict then it will succeed. Otherwise it will hit the condition |
| 229 | # } else if ($ns !== 0) { and be moved to Broken/Namespace:Title |
| 230 | # whereas without this check it would just go to Broken/Title |
| 231 | if ( $ns === 0 ) { |
| 232 | $ns = $title->getNamespace(); |
| 233 | } |
| 234 | |
| 235 | # Old cleanupTitles could move articles there. See T25147. |
| 236 | # or a page could be stored as (0, "Special:Foo") in which case the $titleImpossible |
| 237 | # condition would be true and we've already added a prefix so pretend we're in mainspace |
| 238 | # and don't add another |
| 239 | if ( $ns < 0 ) { |
| 240 | $ns = 0; |
| 241 | } |
| 242 | |
| 243 | # Namespace which no longer exists. Put the page in the main namespace |
| 244 | # since we don't have any idea of the old namespace name. See T70501. |
| 245 | # We build the new title ourself rather than relying on getDBKey() because |
| 246 | # that will return Special:BadTitle |
| 247 | $namespaceInfo = $this->getServiceContainer()->getNamespaceInfo(); |
| 248 | if ( !$namespaceInfo->exists( $ns ) ) { |
| 249 | $clean = "{$this->prefix}NS$ns:$row->page_title"; |
| 250 | $ns = 0; |
| 251 | } elseif ( !$titleImpossible && !$title->exists( IDBAccessObject::READ_LATEST ) ) { |
| 252 | // Looks like the current title, after cleaning it up, is valid and available |
| 253 | $clean = $prior; |
| 254 | } elseif ( $ns !== 0 ) { |
| 255 | // Put all broken pages in the main namespace so that they can be found via Special:PrefixIndex |
| 256 | $nsName = $namespaceInfo->getCanonicalName( $ns ); |
| 257 | $clean = "{$this->prefix}$nsName:{$prior}"; |
| 258 | $ns = 0; |
| 259 | } else { |
| 260 | $clean = $this->prefix . $prior; |
| 261 | } |
| 262 | $verified = Title::makeTitleSafe( $ns, $clean ); |
| 263 | if ( !$verified || $verified->exists( IDBAccessObject::READ_LATEST ) ) { |
| 264 | $lastResort = "{$this->prefix}id: {$row->page_id}"; |
| 265 | $this->output( "Couldn't legalize; form '$clean' exists; using '$lastResort'\n" ); |
| 266 | $verified = Title::makeTitleSafe( $ns, $lastResort ); |
| 267 | if ( !$verified || $verified->exists( IDBAccessObject::READ_LATEST ) ) { |
| 268 | // This can happen in corner cases like if numbers are made not valid |
| 269 | // title characters using the (deprecated) $wgLegalTitleChars or |
| 270 | // a 'Broken/id:foo' title already exists |
| 271 | $this->error( "Destination page $lastResort invalid or already exists." ); |
| 272 | return; |
| 273 | } |
| 274 | } |
| 275 | $title = $verified; |
| 276 | } |
| 277 | |
| 278 | $ns = $title->getNamespace(); |
| 279 | $dest = $title->getDBkey(); |
| 280 | |
| 281 | if ( $this->dryrun ) { |
| 282 | $this->output( "DRY RUN: would rename $row->page_id ($row->page_namespace," . |
| 283 | "'$row->page_title') to ($ns,'$dest')\n" ); |
| 284 | } else { |
| 285 | $this->output( "renaming $row->page_id ($row->page_namespace," . |
| 286 | "'$row->page_title') to ($ns,'$dest')\n" ); |
| 287 | $this->getPrimaryDB() |
| 288 | ->newUpdateQueryBuilder() |
| 289 | ->update( 'page' ) |
| 290 | ->set( [ |
| 291 | 'page_namespace' => $ns, |
| 292 | 'page_title' => $dest |
| 293 | ] ) |
| 294 | ->where( [ 'page_id' => $row->page_id ] ) |
| 295 | ->caller( __METHOD__ )->execute(); |
| 296 | $this->getServiceContainer()->getLinkCache()->clear(); |
| 297 | } |
| 298 | } |
| 299 | } |
| 300 | |
| 301 | // @codeCoverageIgnoreStart |
| 302 | $maintClass = TitleCleanup::class; |
| 303 | require_once RUN_MAINTENANCE_IF_MAIN; |
| 304 | // @codeCoverageIgnoreEnd |