From f3a4fd56c020aa717dfd6eb69bdfe6a10ea35393 Mon Sep 17 00:00:00 2001
From: Lee Daniel Crocker <lcrocker@users.mediawiki.org>
Date: Fri, 2 May 2003 22:55:37 +0000
Subject: [PATCH] Completed several maintenance scripts for index rebuilding.

---
 maintenance/rebuildIndex.php                  |  46 -------
 maintenance/rebuildall.php                    |  39 ++++++
 .../{rebuildLinks.inc => rebuildlinks.inc}    |   2 +-
 .../{rebuildLinks.php => rebuildlinks.php}    |  19 +--
 maintenance/rebuildrecentchanges.inc          | 115 ++++++++++++++++++
 maintenance/rebuildrecentchanges.php          |  30 +++++
 maintenance/rebuildtextindex.inc              |  43 +++++++
 maintenance/rebuildtextindex.php              |  31 +++++
 maintenance/tables.sql                        |  12 +-
 9 files changed, 277 insertions(+), 60 deletions(-)
 delete mode 100644 maintenance/rebuildIndex.php
 create mode 100644 maintenance/rebuildall.php
 rename maintenance/{rebuildLinks.inc => rebuildlinks.inc} (99%)
 rename maintenance/{rebuildLinks.php => rebuildlinks.php} (53%)
 create mode 100644 maintenance/rebuildrecentchanges.inc
 create mode 100644 maintenance/rebuildrecentchanges.php
 create mode 100644 maintenance/rebuildtextindex.inc
 create mode 100644 maintenance/rebuildtextindex.php

diff --git a/maintenance/rebuildIndex.php b/maintenance/rebuildIndex.php
deleted file mode 100644
index 1bd085cff4..0000000000
--- a/maintenance/rebuildIndex.php
+++ /dev/null
@@ -1,46 +0,0 @@
-<?
-
-# Rebuild the fulltext search indexes. This may take a while
-# depending on the database size and server configuration.
-
-global $IP;
-include_once( "../LocalSettings.php" );
-include_once( "$IP/Setup.php" );
-include_once( "$IP/SearchUpdate.php" );
-set_time_limit(0);
-
-$wgDBuser			= "wikiadmin";
-$wgDBpassword		= $wgDBadminpassword;
-
-# May run faster if you drop the index; but will break attempts to search
-# while it's running if you're online.
-#echo "Dropping index...\n";
-##$sql = "ALTER TABLE searchindex DROP INDEX si_title, DROP INDEX si_text";
-#$res = wfQuery($sql);
-
-$sql = "SELECT COUNT(*) AS count FROM cur";
-$res = wfQuery($sql);
-$s = wfFetchObject($res);
-echo "Rebuilding index fields for {$s->count} pages...\n";
-$n = 0;
-
-$sql = "SELECT cur_id, cur_namespace, cur_title, cur_text FROM cur";
-$res = wfQuery($sql);
-while( $s = wfFetchObject($res)) {
-	$u = new SearchUpdate( $s->cur_id, $s->cur_title, $s->cur_text );
-	$u->doUpdate();
-	if ( ( (++$n) % 500) == 0) {
-		echo "$n\n";
-	}
-}
-wfFreeResult( $res );
-
-#echo "Rebuild the index...\n";
-##$sql = "ALTER TABLE searchindex ADD FULLTEXT si_title (si_title),
-##  ADD FULLTEXT si_text (si_text)";
-#$res = wfQuery($sql);
-
-print "Done.\n";
-exit();
-
-?>
diff --git a/maintenance/rebuildall.php b/maintenance/rebuildall.php
new file mode 100644
index 0000000000..9e64d193c3
--- /dev/null
+++ b/maintenance/rebuildall.php
@@ -0,0 +1,39 @@
+<?
+
+# Rebuild link tracking tables from scratch.  This takes several
+# hours, depending on the database size and server configuration.
+
+if ( ! is_readable( "../LocalSettings.php" ) ) {
+	print "A copy of your installation's LocalSettings.php\n" .
+	  "must exist in the source directory.\n";
+	exit();
+}
+
+$DP = "../includes";
+include_once( "../LocalSettings.php" );
+include_once( "../AdminSettings.php" );
+
+include_once( "{$IP}/Setup.php" );
+include_once( "./rebuildlinks.inc" );
+include_once( "./rebuildtextindex.inc" );
+include_once( "./rebuildrecentchanges.inc" );
+$wgTitle = Title::newFromText( "Rebuild links script" );
+set_time_limit(0);
+
+$wgDBuser			= $wgDBadminuser;
+$wgDBpassword		= $wgDBadminpassword;
+
+rebuildLinkTablesPass1();
+rebuildLinkTablesPass2();
+
+dropTextIndex();
+rebuildTextIndex();
+createTextIndex();
+
+rebuildRecentChangesTablePass1();
+rebuildRecentChangesTablePass2();
+
+print "Done.\n";
+exit();
+
+?>
diff --git a/maintenance/rebuildLinks.inc b/maintenance/rebuildlinks.inc
similarity index 99%
rename from maintenance/rebuildLinks.inc
rename to maintenance/rebuildlinks.inc
index e98d6d7635..7dd3ddee50 100644
--- a/maintenance/rebuildLinks.inc
+++ b/maintenance/rebuildlinks.inc
@@ -2,7 +2,7 @@
 
 # Functions for rebuilding the link tracking tables; must
 # be included within a script that also includes the Setup.
-# See convertdb.php, for example.
+# See rebuildlinks.php, for example.
 #
 
 function rebuildLinkTablesPass1()
diff --git a/maintenance/rebuildLinks.php b/maintenance/rebuildlinks.php
similarity index 53%
rename from maintenance/rebuildLinks.php
rename to maintenance/rebuildlinks.php
index 2890507dfe..de756c1911 100644
--- a/maintenance/rebuildLinks.php
+++ b/maintenance/rebuildlinks.php
@@ -3,22 +3,27 @@
 # Rebuild link tracking tables from scratch.  This takes several
 # hours, depending on the database size and server configuration.
 
-global $IP;
+if ( ! is_readable( "../LocalSettings.php" ) ) {
+	print "A copy of your installation's LocalSettings.php\n" .
+	  "must exist in the source directory.\n";
+	exit();
+}
+
+$DP = "../includes";
 include_once( "../LocalSettings.php" );
-include_once( "$IP/Setup.php" );
-include_once( "./rebuildLinks.inc" );
-include_once( "./rebuildRecentChanges.inc" );
+include_once( "../AdminSettings.php" );
+
+include_once( "{$IP}/Setup.php" );
+include_once( "./rebuildlinks.inc" );
 $wgTitle = Title::newFromText( "Rebuild links script" );
 set_time_limit(0);
 
-$wgDBuser			= "wikiadmin";
+$wgDBuser			= $wgDBadminuser;
 $wgDBpassword		= $wgDBadminpassword;
 
 rebuildLinkTablesPass1();
 rebuildLinkTablesPass2();
 
-rebuildRecentChangesTable();
-
 print "Done.\n";
 exit();
 
diff --git a/maintenance/rebuildrecentchanges.inc b/maintenance/rebuildrecentchanges.inc
new file mode 100644
index 0000000000..2d9c00933a
--- /dev/null
+++ b/maintenance/rebuildrecentchanges.inc
@@ -0,0 +1,115 @@
+<?
+
+# Rebuild recent changes table.
+#
+
+function rebuildRecentChangesTablePass1()
+{
+	$sql = "DROP TABLE IF EXISTS recentchanges";
+	wfQuery( $sql );
+
+	$sql = "CREATE TABLE recentchanges (
+  rc_timestamp varchar(14) binary NOT NULL default '',
+  rc_cur_time varchar(14) binary NOT NULL default '',
+  rc_user int(10) unsigned NOT NULL default '0',
+  rc_user_text varchar(255) binary NOT NULL default '',
+  rc_namespace tinyint(3) unsigned NOT NULL default '0',
+  rc_title varchar(255) binary NOT NULL default '',
+  rc_comment varchar(255) binary NOT NULL default '',
+  rc_minor tinyint(3) unsigned NOT NULL default '0',
+  rc_bot tinyint(3) unsigned NOT NULL default '0',
+  rc_new tinyint(3) unsigned NOT NULL default '0',
+  rc_cur_id int(10) unsigned NOT NULL default '0',
+  rc_this_oldid int(10) unsigned NOT NULL default '0',
+  rc_last_oldid int(10) unsigned NOT NULL default '0',
+  INDEX rc_cur_id (rc_cur_id),
+  INDEX rc_cur_time (rc_cur_time),
+  INDEX rc_timestamp (rc_timestamp),
+  INDEX rc_namespace (rc_namespace),
+  INDEX rc_title (rc_title)
+) TYPE=MyISAM PACK_KEYS=1;";
+	wfQuery( $sql );
+
+	print( "Loading from CUR table...\n" );
+
+	$sql = "INSERT INTO recentchanges (rc_timestamp,rc_cur_time,rc_user," .
+	  "rc_user_text,rc_namespace,rc_title,rc_comment,rc_minor,rc_bot,rc_new," .
+	  "rc_cur_id,rc_this_oldid,rc_last_oldid) SELECT cur_timestamp," .
+	  "cur_timestamp,cur_user,cur_user_text,cur_namespace,cur_title," .
+	  "cur_comment,cur_minor_edit,0,cur_is_new,cur_id,0,0 FROM cur " .
+	  "ORDER BY cur_timestamp DESC LIMIT 5000";
+	wfQuery( $sql );
+
+	print( "Loading from OLD table...\n" );
+
+	$sql = "INSERT INTO recentchanges (rc_timestamp,rc_cur_time,rc_user," .
+      "rc_user_text,rc_namespace,rc_title,rc_comment,rc_minor,rc_bot,rc_new," .
+      "rc_cur_id,rc_this_oldid,rc_last_oldid) SELECT old_timestamp,''," .
+	  "old_user,old_user_text,old_namespace,old_title,old_comment," .
+	  "old_minor_edit,0,0,0,old_id,0 FROM old ORDER BY old_timestamp " .
+	  "DESC LIMIT 5000";
+	wfQuery( $sql );
+
+	$sql = "SELECT rc_timestamp FROM recentchanges " .
+	  "ORDER BY rc_timestamp DESC LIMIT 5000,1";
+	$res = wfQuery( $sql );
+	$obj = wfFetchObject( $res );
+	$ts = $obj->rc_timestamp;
+
+	$sql = "DELETE FROM recentchanges WHERE rc_timestamp < '{$ts}'";
+	wfQuery( $sql );
+}
+
+function rebuildRecentChangesTablePass2()
+{
+	$ns = $id = $count = 0;
+	$title = $ct =  "";
+
+	print( "Updating links...\n" );
+
+	$sql = "SELECT rc_namespace,rc_title,rc_timestamp FROM recentchanges " .
+	  "ORDER BY rc_namespace,rc_title,rc_timestamp DESC";
+	$res = wfQuery( $sql );
+
+	while ( $obj = wfFetchObject( $res ) ) {
+		if ( ! ( $ns == $obj->rc_namespace &&
+			   0 == strcmp( $title, wfStrencode( $obj->rc_title ) ) ) ) {
+
+			$ns = $obj->rc_namespace;
+			$title = wfStrencode( $obj->rc_title );
+
+			$sql = "SELECT cur_id,cur_timestamp FROM cur WHERE " .
+			  "cur_namespace={$ns} AND cur_title='{$title}'";
+			$res2 = wfQuery( $sql );
+			$obj2 = wfFetchObject( $res2 );
+
+			$id = $obj2->cur_id;
+			$ct = $obj2->cur_timestamp;
+		}
+		$sql = "SELECT old_id FROM old WHERE old_namespace={$ns} " .
+		  "AND old_title='{$title}' AND old_timestamp < '" .
+		  "{$obj->rc_timestamp}' ORDER BY old_timestamp DESC LIMIT 1";
+		$res2 = wfQuery( $sql );
+
+		if ( 0 != wfNumRows( $res2 ) ) {
+			$obj2 = wfFetchObject( $res2 );
+
+			$sql = "UPDATE recentchanges SET rc_cur_id={$id},rc_cur_time=" .
+			  "'{$ct}',rc_last_oldid={$obj2->old_id} WHERE " .
+			  "rc_namespace={$ns} AND rc_title='{$title}' AND " .
+			  "rc_timestamp='{$obj->rc_timestamp}'";
+			wfQuery( $sql );
+		} else {
+			$sql = "UPDATE recentchanges SET rc_cur_id={$id},rc_cur_time=" .
+			  "'{$ct}' WHERE rc_namespace={$ns} AND rc_title='{$title}' " .
+			  "AND rc_timestamp='{$obj->rc_timestamp}'";
+			wfQuery( $sql );
+		}
+
+		if ( 0 == ( ++$count % 500 ) ) {
+			printf( "%d records processed.\n", $count );
+		}
+	}
+}
+
+?>
diff --git a/maintenance/rebuildrecentchanges.php b/maintenance/rebuildrecentchanges.php
new file mode 100644
index 0000000000..894353f179
--- /dev/null
+++ b/maintenance/rebuildrecentchanges.php
@@ -0,0 +1,30 @@
+<?
+
+# Rebuild link tracking tables from scratch.  This takes several
+# hours, depending on the database size and server configuration.
+
+if ( ! is_readable( "../LocalSettings.php" ) ) {
+	print "A copy of your installation's LocalSettings.php\n" .
+	  "must exist in the source directory.\n";
+	exit();
+}
+
+$DP = "../includes";
+include_once( "../LocalSettings.php" );
+include_once( "../AdminSettings.php" );
+
+include_once( "{$IP}/Setup.php" );
+include_once( "./rebuildrecentchanges.inc" );
+$wgTitle = Title::newFromText( "Rebuild recent changes script" );
+set_time_limit(0);
+
+$wgDBuser			= $wgDBadminuser;
+$wgDBpassword		= $wgDBadminpassword;
+
+rebuildRecentChangesTablePass1();
+rebuildRecentChangesTablePass2();
+
+print "Done.\n";
+exit();
+
+?>
diff --git a/maintenance/rebuildtextindex.inc b/maintenance/rebuildtextindex.inc
new file mode 100644
index 0000000000..8d681c8fdf
--- /dev/null
+++ b/maintenance/rebuildtextindex.inc
@@ -0,0 +1,43 @@
+<?
+
+# Rebuild the fulltext search indexes. This may take a while
+# depending on the database size and server configuration.
+
+# Rebuilding is faster if you drop the index and recreate it,
+# but that will prevent searches from working while it runs.
+
+function dropTextIndex()
+{
+	echo "Dropping index...\n";
+	$sql = "ALTER TABLE searchindex DROP INDEX si_title, DROP INDEX si_text";
+	$res = wfQuery($sql);
+}
+
+function createTextIndex()
+{
+	echo "Rebuild the index...\n";
+	$sql = "ALTER TABLE searchindex ADD FULLTEXT si_title (si_title), " .
+	  "ADD FULLTEXT si_text (si_text)";
+	$res = wfQuery($sql);
+}
+
+function rebuildTextIndex()
+{
+	$sql = "SELECT COUNT(*) AS count FROM cur";
+	$res = wfQuery($sql);
+	$s = wfFetchObject($res);
+	echo "Rebuilding index fields for {$s->count} pages...\n";
+	$n = 0;
+
+	$sql = "SELECT cur_id, cur_namespace, cur_title, cur_text FROM cur";
+	$res = wfQuery($sql);
+
+	while( $s = wfFetchObject($res) ) {
+		$u = new SearchUpdate( $s->cur_id, $s->cur_title, $s->cur_text );
+		$u->doUpdate();
+		if ( ( (++$n) % 500) == 0) { echo "$n\n"; }
+	}
+	wfFreeResult( $res );
+}
+
+?>
diff --git a/maintenance/rebuildtextindex.php b/maintenance/rebuildtextindex.php
new file mode 100644
index 0000000000..bbc5327d3b
--- /dev/null
+++ b/maintenance/rebuildtextindex.php
@@ -0,0 +1,31 @@
+<?
+
+# Rebuild link tracking tables from scratch.  This takes several
+# hours, depending on the database size and server configuration.
+
+if ( ! is_readable( "../LocalSettings.php" ) ) {
+	print "A copy of your installation's LocalSettings.php\n" .
+	  "must exist in the source directory.\n";
+	exit();
+}
+
+$DP = "../includes";
+include_once( "../LocalSettings.php" );
+include_once( "../AdminSettings.php" );
+
+include_once( "{$IP}/Setup.php" );
+include_once( "./rebuildtextindex.inc" );
+$wgTitle = Title::newFromText( "Rebuild text index script" );
+set_time_limit(0);
+
+$wgDBuser			= $wgDBadminuser;
+$wgDBpassword		= $wgDBadminpassword;
+
+dropTextIndex();
+rebuildTextIndex();
+createTextIndex();
+
+print "Done.\n";
+exit();
+
+?>
diff --git a/maintenance/tables.sql b/maintenance/tables.sql
index 09556af74d..9ca9f5e334 100644
--- a/maintenance/tables.sql
+++ b/maintenance/tables.sql
@@ -165,12 +165,12 @@ CREATE TABLE watchlist (
 
 DROP TABLE IF EXISTS math;
 CREATE TABLE math (
-    math_inputhash varchar(16) NOT NULL,
-    math_outputhash varchar(16) NOT NULL,
-    math_html_conservativeness tinyint(1) NOT NULL,
-    math_html text,
-    math_mathml text,
-    UNIQUE KEY math_inputhash (math_inputhash)
+  math_inputhash varchar(16) NOT NULL,
+  math_outputhash varchar(16) NOT NULL,
+  math_html_conservativeness tinyint(1) NOT NULL,
+  math_html text,
+  math_mathml text,
+  UNIQUE KEY math_inputhash (math_inputhash)
 ) TYPE=MyISAM;
 
 -- Table searchindex must be MyISAM for fulltext support
-- 
2.20.1