generateSitemap can now optionally skip redirects

Slightly enhanced on commit.

Patchset2:
 - get ride of unneeded arguments passed to addOption()
 - rename $redirects to $skippedRedirects. Easier to search.
 - easier to read conditional test of page_is_redirect

Change-Id: I4ae528a584c5a9d73b615a6281d75b14d056e5e7

diff --git a/CREDITS b/CREDITS
index d0c4623..5c41d96 100644 (file)
--- a/CREDITS
+++ b/CREDITS
@@ -86,6 +86,7 @@ following names for their contribution to the product.
 * Amir E. Aharoni
 * Andrew Dunbar
 * Antonio Ospite
+* Asier Lostalé
 * Azliq7
 * Bagariavivek
 * Beau

diff --git a/RELEASE-NOTES-1.20 b/RELEASE-NOTES-1.20
index 5fb07b2..a0b3d4d 100644 (file)
--- a/RELEASE-NOTES-1.20
+++ b/RELEASE-NOTES-1.20
@@ -36,6 +36,7 @@ production.
 * (bug 35728) Git revisions are now linked on Special:Version.
 * "Show Changes" on default messages shows now diff against default message text
 * (bug 23006) create #speciale parser function.
+* generateSitemap can now optionally skip redirect pages.
 
 === Bug fixes in 1.20 ===
 * (bug 30245) Use the correct way to construct a log page title.

diff --git a/maintenance/generateSitemap.php b/maintenance/generateSitemap.php
index 8f19ef5..7e83d5f 100644 (file)
--- a/maintenance/generateSitemap.php
+++ b/maintenance/generateSitemap.php
@@ -71,6 +71,13 @@ class GenerateSitemap extends Maintenance {
         */
        var $compress;
 
+       /**
+        * Whether or not to include redirection pages
+        *
+        * @var bool
+        */
+       var $skipRedirects;
+
        /**
         * The number of entries to save in each sitemap file
         *
@@ -137,6 +144,7 @@ class GenerateSitemap extends Maintenance {
                $this->addOption( 'fspath', 'The file system path to save to, e.g. /tmp/sitemap; defaults to current directory', false, true );
                $this->addOption( 'urlpath', 'The URL path corresponding to --fspath, prepended to filenames in the index; defaults to an empty string', false, true );
                $this->addOption( 'compress', 'Compress the sitemap files, can take value yes|no, default yes', false, true );
+               $this->addOption( 'skip-redirects', 'Do not include redirecting articles in the sitemap' );
                $this->addOption( 'identifier', 'What site identifier to use for the wiki, defaults to $wgDBname', false, true );
        }
 
@@ -154,6 +162,7 @@ class GenerateSitemap extends Maintenance {
                }
                $this->identifier = $this->getOption( 'identifier', wfWikiID() );
                $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
+               $this->skipRedirects = $this->getOption( 'skip-redirects', false ) !== false ;
                $this->dbr = wfGetDB( DB_SLAVE );
                $this->generateNamespaces();
                $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
@@ -279,6 +288,7 @@ class GenerateSitemap extends Maintenance {
                                'page_namespace',
                                'page_title',
                                'page_touched',
+                               'page_is_redirect'
                        ),
                        array( 'page_namespace' => $namespace ),
                        __METHOD__
@@ -302,7 +312,13 @@ class GenerateSitemap extends Maintenance {
 
                        $fns = $wgContLang->getFormattedNsText( $namespace );
                        $this->output( "$namespace ($fns)\n" );
+                       $skippedRedirects = 0;  // Number of redirects skipped for that namespace
                        foreach ( $res as $row ) {
+                               if ($this->skipRedirects && $row->page_is_redirect ) {
+                                       $skippedRedirects++;
+                                       continue;
+                               }
+
                                if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
                                        if ( $this->file !== false ) {
                                                $this->write( $this->file, $this->closeFile() );
@@ -332,6 +348,11 @@ class GenerateSitemap extends Maintenance {
                                        }
                                }
                        }
+
+                       if ($this->skipRedirects && $skippedRedirects > 0) {
+                               $this->output( "  skipped $skippedRedirects redirect(s)\n" );
+                       }
+
                        if ( $this->file ) {
                                $this->write( $this->file, $this->closeFile() );
                                $this->close( $this->file );