parse( $file );
}
}
/**
* The Regex parser is the only parser we have found that
* doesn't break serialized data. It does have two bugs
* that can break serialized data. Those are calling rtrim
* on each $importline and adding a newline to each $importline.
* This class fixes those bugs.
*
* @since 1.8
*/
class FLBuilderImportParserRegex extends WXR_Parser_Regex {
/**
* @since 1.8
* @return array
*/
function parse( $file ) {
// @codingStandardsIgnoreLine
$wxr_version = $in_post = false;
$fp = $this->fopen( $file, 'r' );
if ( $fp ) {
while ( ! $this->feof( $fp ) ) {
$importline = $this->fgets( $fp );
if ( ! $wxr_version && preg_match( '|(\d+\.\d+)|', $importline, $version ) ) {
$wxr_version = $version[1];
}
if ( false !== strpos( $importline, '' ) ) {
preg_match( '|(.*?)|is', $importline, $url );
$this->base_url = $url[1];
continue;
}
if ( false !== strpos( $importline, '' ) ) {
preg_match( '|(.*?)|is', $importline, $category );
if ( isset( $category[1] ) ) {
$this->categories[] = $this->process_category( $category[1] );
}
continue;
}
if ( false !== strpos( $importline, '' ) ) {
preg_match( '|(.*?)|is', $importline, $tag );
if ( isset( $tag[1] ) ) {
$this->tags[] = $this->process_tag( $tag[1] );
}
continue;
}
if ( false !== strpos( $importline, '' ) ) {
preg_match( '|(.*?)|is', $importline, $term );
if ( isset( $term[1] ) ) {
$this->terms[] = $this->process_term( $term[1] );
}
continue;
}
if ( false !== strpos( $importline, '' ) ) {
preg_match( '|(.*?)|is', $importline, $author );
if ( isset( $author[1] ) ) {
$a = $this->process_author( $author[1] );
}
$this->authors[ $a['author_login'] ] = $a;
continue;
}
if ( false !== strpos( $importline, '- ' ) ) {
$post = '';
$in_post = true;
continue;
}
if ( false !== strpos( $importline, '
' ) ) {
$in_post = false;
$this->set_pcre_limit( apply_filters( 'fl_builder_importer_pcre', '23001337' ) );
$this->posts[] = $this->process_post( $post );
$this->set_pcre_limit( 'default' );
continue;
}
if ( $in_post ) {
$post .= $importline;
}
}
$this->fclose( $fp );
// Try to fix any broken builder data.
foreach ( $this->posts as $post_index => $post ) {
if ( ! isset( $post['postmeta'] ) || ! is_array( $post['postmeta'] ) ) {
continue;
}
foreach ( $post['postmeta'] as $postmeta_index => $postmeta ) {
if ( stristr( $postmeta['key'], '_fl_builder_' ) ) {
$this->posts[ $post_index ]['postmeta'][ $postmeta_index ]['value'] = FLBuilderImporterDataFix::run( $postmeta['value'] );
}
}
}
}
if ( ! $wxr_version ) {
return new WP_Error( 'WXR_parse_error', __( 'This does not appear to be a WXR file, missing/invalid WXR version number', 'fl-builder' ) );
}
return array(
'authors' => $this->authors,
'posts' => $this->posts,
'categories' => $this->categories,
'tags' => $this->tags,
'terms' => $this->terms,
'base_url' => $this->base_url,
'version' => $wxr_version,
);
}
/**
* Try increasing PCRE limit to avoid failing of importing huge postmeta data.
*
* @since 1.10.9
* @param string $value
*/
function set_pcre_limit( $value ) {
if ( ! isset( $this->default_backtrack_limit ) ) {
$this->default_backtrack_limit = @ini_get( 'pcre.backtrack_limit' ); // @codingStandardsIgnoreLine
$this->default_recursion_limit = @ini_get( 'pcre.recursion_limit' ); // @codingStandardsIgnoreLine
}
if ( 'default' != $value ) {
@ini_set( 'pcre.backtrack_limit', $value ); // @codingStandardsIgnoreLine
@ini_set( 'pcre.recursion_limit', $value ); // @codingStandardsIgnoreLine
} else {
// Reset limit back to default.
if ( is_numeric( $this->default_backtrack_limit ) ) {
@ini_set( 'pcre.backtrack_limit', $this->default_backtrack_limit ); // @codingStandardsIgnoreLine
}
if ( is_numeric( $this->default_recursion_limit ) ) {
@ini_set( 'pcre.recursion_limit', $this->default_recursion_limit ); // @codingStandardsIgnoreLine
}
}
}
}
/**
* Portions borrowed from https://github.com/Blogestudio/Fix-Serialization/blob/master/fix-serialization.php
*
* Attempts to fix broken serialized data.
*
* @since 1.8
*/
final class FLBuilderImporterDataFix {
/**
* @since 1.8
* @return string
*/
static public function run( $data ) {
// return if empty
if ( empty( $data ) ) {
return $data;
}
$data = maybe_unserialize( $data );
// return if maybe_unserialize() returns an object or array, this is good.
if ( is_object( $data ) || is_array( $data ) ) {
return $data;
}
return preg_replace_callback( '!s:(\d+):([\\\\]?"[\\\\]?"|[\\\\]?"((.*?)[^\\\\])[\\\\]?");!', 'FLBuilderImporterDataFix::regex_callback', $data );
}
/**
* @since 1.8
* @return string
*/
static public function regex_callback( $matches ) {
if ( ! isset( $matches[3] ) ) {
return $matches[0];
}
return 's:' . strlen( self::unescape_mysql( $matches[3] ) ) . ':"' . self::unescape_quotes( $matches[3] ) . '";';
}
/**
* Unescape to avoid dump-text issues.
*
* @since 1.8
* @access private
* @return string
*/
static private function unescape_mysql( $value ) {
return str_replace( array( '\\\\', "\\0", "\\n", "\\r", '\Z', "\'", '\"' ),
array( '\\', "\0", "\n", "\r", "\x1a", "'", '"' ),
$value );
}
/**
* Fix strange behaviour if you have escaped quotes in your replacement.
*
* @since 1.8
* @access private
* @return string
*/
static private function unescape_quotes( $value ) {
return str_replace( '\"', '"', $value );
}
}