$this->charset = DB_CHARSET;
}
- if ( ( $this->use_mysqli && ! ( $this->dbh instanceof mysqli ) )
- || ( empty( $this->dbh ) || ! ( $this->dbh instanceof mysqli ) ) ) {
+ if ( ( $this->use_mysqli && ! ( $this->dbh instanceof mysqli ) ) || empty( $this->dbh ) ) {
return;
}
* @return int|false The number of rows affected, or false on error.
*/
function _insert_replace_helper( $table, $data, $format = null, $type = 'INSERT' ) {
+ $this->insert_id = 0;
+
if ( ! in_array( strtoupper( $type ), array( 'REPLACE', 'INSERT' ) ) ) {
return false;
}
$sql = "$type INTO `$table` ($fields) VALUES ($formats)";
- $this->insert_id = 0;
$this->check_current_query = false;
return $this->query( $this->prepare( $sql, $values ) );
}
*/
protected function process_fields( $table, $data, $format ) {
$data = $this->process_field_formats( $data, $format );
+ if ( false === $data ) {
+ return false;
+ }
+
$data = $this->process_field_charsets( $data, $table );
if ( false === $data ) {
return false;
}
+ $data = $this->process_field_lengths( $data, $table );
+ if ( false === $data ) {
+ return false;
+ }
+
$converted_data = $this->strip_invalid_text( $data );
if ( $data !== $converted_data ) {
// We can skip this field if we know it isn't a string.
// This checks %d/%f versus ! %s because it's sprintf() could take more.
$value['charset'] = false;
- } elseif ( $this->check_ascii( $value['value'] ) ) {
- // If it's ASCII, then we don't need the charset. We can skip this field.
- $value['charset'] = false;
} else {
$value['charset'] = $this->get_col_charset( $table, $field );
if ( is_wp_error( $value['charset'] ) ) {
return false;
}
+ }
- // This isn't ASCII. Don't have strip_invalid_text() re-check.
- $value['ascii'] = false;
+ $data[ $field ] = $value;
+ }
+
+ return $data;
+ }
+
+ /**
+ * For string fields, record the maximum string length that field can safely save.
+ *
+ * @since 4.2.1
+ * @access protected
+ *
+ * @param array $data As it comes from the wpdb::process_field_charsets() method.
+ * @param string $table Table name.
+ * @return array|False The same array as $data with additional 'length' keys, or false if
+ * any of the values were too long for their corresponding field.
+ */
+ protected function process_field_lengths( $data, $table ) {
+ foreach ( $data as $field => $value ) {
+ if ( '%d' === $value['format'] || '%f' === $value['format'] ) {
+ // We can skip this field if we know it isn't a string.
+ // This checks %d/%f versus ! %s because it's sprintf() could take more.
+ $value['length'] = false;
+ } else {
+ $value['length'] = $this->get_col_length( $table, $field );
+ if ( is_wp_error( $value['length'] ) ) {
+ return false;
+ }
}
$data[ $field ] = $value;
public function get_var( $query = null, $x = 0, $y = 0 ) {
$this->func_call = "\$db->get_var(\"$query\", $x, $y)";
- if ( $this->check_safe_collation( $query ) ) {
+ if ( $this->check_current_query && $this->check_safe_collation( $query ) ) {
$this->check_current_query = false;
}
public function get_row( $query = null, $output = OBJECT, $y = 0 ) {
$this->func_call = "\$db->get_row(\"$query\",$output,$y)";
- if ( $this->check_safe_collation( $query ) ) {
+ if ( $this->check_current_query && $this->check_safe_collation( $query ) ) {
$this->check_current_query = false;
}
* @return array Database query result. Array indexed from 0 by SQL result row number.
*/
public function get_col( $query = null , $x = 0 ) {
- if ( $this->check_safe_collation( $query ) ) {
+ if ( $this->check_current_query && $this->check_safe_collation( $query ) ) {
$this->check_current_query = false;
}
public function get_results( $query = null, $output = OBJECT ) {
$this->func_call = "\$db->get_results(\"$query\", $output)";
- if ( $this->check_safe_collation( $query ) ) {
+ if ( $this->check_current_query && $this->check_safe_collation( $query ) ) {
$this->check_current_query = false;
}
}
$charsets = $columns = array();
- $results = $this->get_results( "SHOW FULL COLUMNS FROM `$table`" );
+
+ $table_parts = explode( '.', $table );
+ $table = '`' . implode( '`.`', $table_parts ) . '`';
+ $results = $this->get_results( "SHOW FULL COLUMNS FROM $table" );
if ( ! $results ) {
return new WP_Error( 'wpdb_get_table_charset_failure' );
}
return $charset;
}
+ /**
+ * Retrieve the maximum string length allowed in a given column.
+ * The length may either be specified as a byte length or a character length.
+ *
+ * @since 4.2.1
+ * @access public
+ *
+ * @param string $table Table name.
+ * @param string $column Column name.
+ * @return mixed array( 'length' => (int), 'type' => 'byte' | 'char' )
+ * false if the column has no length (for example, numeric column)
+ * WP_Error object if there was an error.
+ */
+ public function get_col_length( $table, $column ) {
+ $tablekey = strtolower( $table );
+ $columnkey = strtolower( $column );
+
+ // Skip this entirely if this isn't a MySQL database.
+ if ( false === $this->is_mysql ) {
+ return false;
+ }
+
+ if ( empty( $this->col_meta[ $tablekey ] ) ) {
+ // This primes column information for us.
+ $table_charset = $this->get_table_charset( $table );
+ if ( is_wp_error( $table_charset ) ) {
+ return $table_charset;
+ }
+ }
+
+ if ( empty( $this->col_meta[ $tablekey ][ $columnkey ] ) ) {
+ return false;
+ }
+
+ $typeinfo = explode( '(', $this->col_meta[ $tablekey ][ $columnkey ]->Type );
+
+ $type = strtolower( $typeinfo[0] );
+ if ( ! empty( $typeinfo[1] ) ) {
+ $length = trim( $typeinfo[1], ')' );
+ } else {
+ $length = false;
+ }
+
+ switch( $type ) {
+ case 'char':
+ case 'varchar':
+ return array(
+ 'type' => 'char',
+ 'length' => (int) $length,
+ );
+ break;
+ case 'binary':
+ case 'varbinary':
+ return array(
+ 'type' => 'byte',
+ 'length' => (int) $length,
+ );
+ break;
+ case 'tinyblob':
+ case 'tinytext':
+ return array(
+ 'type' => 'byte',
+ 'length' => 255, // 2^8 - 1
+ );
+ break;
+ case 'blob':
+ case 'text':
+ return array(
+ 'type' => 'byte',
+ 'length' => 65535, // 2^16 - 1
+ );
+ break;
+ case 'mediumblob':
+ case 'mediumtext':
+ return array(
+ 'type' => 'byte',
+ 'length' => 16777215, // 2^24 - 1
+ );
+ break;
+ case 'longblob':
+ case 'longtext':
+ return array(
+ 'type' => 'byte',
+ 'length' => 4294967295, // 2^32 - 1
+ );
+ break;
+ default:
+ return false;
+ }
+
+ return false;
+ }
+
/**
* Check if a string is ASCII.
*
// We don't need to check the collation for queries that don't read data.
$query = ltrim( $query, "\r\n\t (" );
- if ( preg_match( '/^(?:SHOW|DESCRIBE|DESC|EXPLAIN)\s/i', $query ) ) {
+ if ( preg_match( '/^(?:SHOW|DESCRIBE|DESC|EXPLAIN|CREATE)\s/i', $query ) ) {
return true;
}
* remove invalid characters, a WP_Error object is returned.
*/
protected function strip_invalid_text( $data ) {
- // Some multibyte character sets that we can check in PHP.
- $mb_charsets = array(
- 'ascii' => 'ASCII',
- 'big5' => 'BIG-5',
- 'eucjpms' => 'eucJP-win',
- 'gb2312' => 'EUC-CN',
- 'ujis' => 'EUC-JP',
- 'utf32' => 'UTF-32',
- );
-
- $supported_charsets = array();
- if ( function_exists( 'mb_list_encodings' ) ) {
- $supported_charsets = mb_list_encodings();
- }
-
$db_check_string = false;
foreach ( $data as &$value ) {
$charset = $value['charset'];
- // Column isn't a string, or is latin1, which will will happily store anything.
- if ( false === $charset || 'latin1' === $charset ) {
+ if ( is_array( $value['length'] ) ) {
+ $length = $value['length']['length'];
+ $truncate_by_byte_length = 'byte' === $value['length']['type'];
+ } else {
+ $length = false;
+ // Since we have no length, we'll never truncate.
+ // Initialize the variable to false. true would take us
+ // through an unnecessary (for this case) codepath below.
+ $truncate_by_byte_length = false;
+ }
+
+ // There's no charset to work with.
+ if ( false === $charset ) {
continue;
}
+ // Column isn't a string.
if ( ! is_string( $value['value'] ) ) {
continue;
}
- // ASCII is always OK.
- if ( ! isset( $value['ascii'] ) && $this->check_ascii( $value['value'] ) ) {
- continue;
+ $needs_validation = true;
+ if (
+ // latin1 can store any byte sequence
+ 'latin1' === $charset
+ ||
+ // ASCII is always OK.
+ ( ! isset( $value['ascii'] ) && $this->check_ascii( $value['value'] ) )
+ ) {
+ $truncate_by_byte_length = true;
+ $needs_validation = false;
}
- // Convert the text locally.
- if ( $supported_charsets ) {
- if ( isset( $mb_charsets[ $charset ] ) && in_array( $mb_charsets[ $charset ], $supported_charsets ) ) {
- $value['value'] = mb_convert_encoding( $value['value'], $mb_charsets[ $charset ], $mb_charsets[ $charset ] );
+ if ( $truncate_by_byte_length ) {
+ mbstring_binary_safe_encoding();
+ if ( false !== $length && strlen( $value['value'] ) > $length ) {
+ $value['value'] = substr( $value['value'], 0, $length );
+ }
+ reset_mbstring_encoding();
+
+ if ( ! $needs_validation ) {
continue;
}
}
// utf8 can be handled by regex, which is a bunch faster than a DB lookup.
- if ( 'utf8' === $charset || 'utf8mb3' === $charset || 'utf8mb4' === $charset ) {
+ if ( ( 'utf8' === $charset || 'utf8mb3' === $charset || 'utf8mb4' === $charset ) && function_exists( 'mb_strlen' ) ) {
$regex = '/
(
(?: [\x00-\x7F] # single-byte sequences 0xxxxxxx
| \xED[\x80-\x9F][\x80-\xBF]
| [\xEE-\xEF][\x80-\xBF]{2}';
- if ( 'utf8mb4' === $charset) {
+ if ( 'utf8mb4' === $charset ) {
$regex .= '
| \xF0[\x90-\xBF][\x80-\xBF]{2} # four-byte sequences 11110xxx 10xxxxxx * 3
| [\xF1-\xF3][\x80-\xBF]{3}
';
}
- $regex .= '){1,50} # ...one or more times
+ $regex .= '){1,40} # ...one or more times
)
| . # anything else
/x';
$value['value'] = preg_replace( $regex, '$1', $value['value'] );
+
+
+ if ( false !== $length && mb_strlen( $value['value'], 'UTF-8' ) > $length ) {
+ $value['value'] = mb_substr( $value['value'], 0, $length, 'UTF-8' );
+ }
continue;
}
$queries = array();
foreach ( $data as $col => $value ) {
if ( ! empty( $value['db'] ) ) {
- if ( ! isset( $queries[ $value['charset'] ] ) ) {
- $queries[ $value['charset'] ] = array();
+ // We're going to need to truncate by characters or bytes, depending on the length value we have.
+ if ( 'byte' === $value['length']['type'] ) {
+ // Using binary causes LEFT() to truncate by bytes.
+ $charset = 'binary';
+ } else {
+ $charset = $value['charset'];
+ }
+
+ if ( is_array( $value['length'] ) ) {
+ $queries[ $col ] = $this->prepare( "CONVERT( LEFT( CONVERT( %s USING $charset ), %.0f ) USING {$this->charset} )", $value['value'], $value['length']['length'] );
+ } else if ( 'binary' !== $charset ) {
+ // If we don't have a length, there's no need to convert binary - it will always return the same result.
+ $queries[ $col ] = $this->prepare( "CONVERT( CONVERT( %s USING $charset ) USING {$this->charset} )", $value['value'] );
}
- // Split the CONVERT() calls by charset, so we can make sure the connection is right
- $queries[ $value['charset'] ][ $col ] = $this->prepare( "CONVERT( %s USING {$value['charset']} )", $value['value'] );
unset( $data[ $col ]['db'] );
}
}
- $connection_charset = $this->charset;
- foreach ( $queries as $charset => $query ) {
+ $sql = array();
+ foreach ( $queries as $column => $query ) {
if ( ! $query ) {
continue;
}
- // Change the charset to match the string(s) we're converting
- if ( $charset !== $connection_charset ) {
- $connection_charset = $charset;
- $this->set_charset( $this->dbh, $charset );
- }
-
- $this->check_current_query = false;
-
- $row = $this->get_row( "SELECT " . implode( ', ', $query ), ARRAY_N );
- if ( ! $row ) {
- $this->set_charset( $this->dbh, $connection_charset );
- return new WP_Error( 'wpdb_strip_invalid_text_failure' );
- }
+ $sql[] = $query . " AS x_$column";
+ }
- $cols = array_keys( $query );
- $col_count = count( $cols );
- for ( $ii = 0; $ii < $col_count; $ii++ ) {
- $data[ $cols[ $ii ] ]['value'] = $row[ $ii ];
- }
+ $this->check_current_query = false;
+ $row = $this->get_row( "SELECT " . implode( ', ', $sql ), ARRAY_A );
+ if ( ! $row ) {
+ return new WP_Error( 'wpdb_strip_invalid_text_failure' );
}
- // Don't forget to change the charset back!
- if ( $connection_charset !== $this->charset ) {
- $this->set_charset( $this->dbh );
+ foreach ( array_keys( $data ) as $column ) {
+ if ( isset( $row["x_$column"] ) ) {
+ $data[ $column ]['value'] = $row["x_$column"];
+ }
}
}
* @return string|WP_Error The converted query, or a WP_Error object if the conversion fails.
*/
protected function strip_invalid_text_from_query( $query ) {
+ // We don't need to check the collation for queries that don't read data.
+ $trimmed_query = ltrim( $query, "\r\n\t (" );
+ if ( preg_match( '/^(?:SHOW|DESCRIBE|DESC|EXPLAIN|CREATE)\s/i', $trimmed_query ) ) {
+ return $query;
+ }
+
$table = $this->get_table_from_query( $query );
if ( $table ) {
$charset = $this->get_table_charset( $table );
'value' => $query,
'charset' => $charset,
'ascii' => false,
+ 'length' => false,
);
$data = $this->strip_invalid_text( array( $data ) );
* @return string|WP_Error The converted string, or a WP_Error object if the conversion fails.
*/
public function strip_invalid_text_for_column( $table, $column, $value ) {
- if ( ! is_string( $value ) || $this->check_ascii( $value ) ) {
+ if ( ! is_string( $value ) ) {
return $value;
}
$column => array(
'value' => $value,
'charset' => $charset,
- 'ascii' => false,
+ 'length' => $this->get_col_length( $table, $column ),
)
);
// Allow (select...) union [...] style queries. Use the first query's table name.
$query = ltrim( $query, "\r\n\t (" );
- /*
- * Strip everything between parentheses except nested selects and use only 1,000
- * chars of the query.
- */
- $query = preg_replace( '/\((?!\s*select)[^(]*?\)/is', '()', substr( $query, 0, 1000 ) );
+ // Strip everything between parentheses except nested selects.
+ $query = preg_replace( '/\((?!\s*select)[^(]*?\)/is', '()', $query );
// Quickly match most common queries.
if ( preg_match( '/^\s*(?:'
. '|REPLACE(?:\s+LOW_PRIORITY|\s+DELAYED)?(?:\s+INTO)?'
. '|UPDATE(?:\s+LOW_PRIORITY)?(?:\s+IGNORE)?'
. '|DELETE(?:\s+LOW_PRIORITY|\s+QUICK|\s+IGNORE)*(?:\s+FROM)?'
- . ')\s+`?([\w-]+)`?/is', $query, $maybe ) ) {
- return $maybe[1];
+ . ')\s+((?:[0-9a-zA-Z$_.`]|[\xC2-\xDF][\x80-\xBF])+)/is', $query, $maybe ) ) {
+ return str_replace( '`', '', $maybe[1] );
}
// SHOW TABLE STATUS and SHOW TABLES
if ( preg_match( '/^\s*(?:'
. 'SHOW\s+TABLE\s+STATUS.+(?:LIKE\s+|WHERE\s+Name\s*=\s*)'
. '|SHOW\s+(?:FULL\s+)?TABLES.+(?:LIKE\s+|WHERE\s+Name\s*=\s*)'
- . ')\W([\w-]+)\W/is', $query, $maybe ) ) {
- return $maybe[1];
+ . ')\W((?:[0-9a-zA-Z$_.`]|[\xC2-\xDF][\x80-\xBF])+)\W/is', $query, $maybe ) ) {
+ return str_replace( '`', '', $maybe[1] );
}
// Big pattern for the rest of the table-related queries.
. '|LOAD\s+DATA.*INFILE.*INTO\s+TABLE'
. '|(?:GRANT|REVOKE).*ON\s+TABLE'
. '|SHOW\s+(?:.*FROM|.*TABLE)'
- . ')\s+\(*\s*`?([\w-]+)`?\s*\)*/is', $query, $maybe ) ) {
- return $maybe[1];
+ . ')\s+\(*\s*((?:[0-9a-zA-Z$_.`]|[\xC2-\xDF][\x80-\xBF])+)\s*\)*/is', $query, $maybe ) ) {
+ return str_replace( '`', '', $maybe[1] );
}
return false;