Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Experiment: Run through esc_attr() in a single optimized pass. #5337

Draft
wants to merge 10 commits into
base: trunk
Choose a base branch
from
197 changes: 197 additions & 0 deletions src/wp-includes/class-wp-token-set.php
Original file line number Diff line number Diff line change
@@ -0,0 +1,197 @@
<?php

class WP_Token_Set {
const KEY_LENGTH = 2;

const MAX_LENGTH = 256;

/**
* Stores an optimized form of the word set, where words are grouped
* by first two letters and then collapsed into a string.
*
* @var array
*/
private $large_words = array();

/**
* Stores an optimized row of short words, where every entry is two
* bytes long and zero-extended if the word is only a single byte.
*
* @var string
*/
private $small_words = '';

public static function from_array( $words ) {
$set = new WP_Token_Set();

// Start by grouping words.

$groups = array();
$shorts = array();
foreach ( $words as $word ) {
if ( ! is_string( $word ) || self::MAX_LENGTH <= strlen( $word ) ) {
return null;
}

$length = strlen( $word );

if ( self::KEY_LENGTH >= $length ) {
$shorts[] = $word;
} else {
$group = substr( $word, 0, self::KEY_LENGTH );

if ( ! isset( $groups[ $group ] ) ) {
$groups[ $group ] = array();
}

$groups[ $group ][] = substr( $word, self::KEY_LENGTH );
}
}

// Sort the words by longest-first, then alphabetical.

usort( $shorts, array( self::class, 'longest_first_then_alphabetical' ) );
foreach ( $groups as $group_key => $group ) {
usort( $groups[ $group_key ], array( self::class, 'longest_first_then_alphabetical' ) );
}

// Finally construct the optimized lookups.

foreach ( $shorts as $word ) {
$set->small_words .= str_pad( $word, self::KEY_LENGTH, "\x00" );
}

foreach ( $groups as $group => $group_words ) {
$group_string = '';

foreach ( $group_words as $word ) {
$group_string .= chr( strlen( $word ) ) . $word;
}

$set->large_words[ $group ] = $group_string;
}

return $set;
}

public static function from_precomputed_table( $large_words, $small_words ) {
$set = new WP_Token_Set();

$set->large_words = $large_words;
$set->small_words = $small_words;

return $set;
}

public function contains( $word ) {
if ( self::KEY_LENGTH >= strlen( $word ) ) {
return str_contains( $this->small_words, str_pad( $word, self::KEY_LENGTH, "\x00" ) );
}

$group_key = substr( $word, 0, self::KEY_LENGTH );
if ( ! isset( $this->large_words[ $group_key ] ) ) {
return false;
}

$group = $this->large_words[ $group_key ];
$slug = substr( $word, self::KEY_LENGTH );
$length = strlen( $slug );
$at = 0;
while ( $at < strlen( $group ) ) {
$token_length = ord( $group[ $at++ ] );
if ( $token_length === $length && 0 === substr_compare( $group, $slug, $at, $token_length ) ) {
return true;
}

$at += $token_length;
}

return false;
}

public function read_token( $text, $offset ) {
$text_length = strlen( $text );

// Search for a long word first, if the text is long enough, and if that fails, a short one.
if ( self::KEY_LENGTH < $text_length ) {
$group_key = substr( $text, $offset, self::KEY_LENGTH );

if ( ! isset( $this->large_words[ $group_key ] ) ) {
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

bug! if the group doesn't exist, it doesn't also imply that the short word doesn't exist. this should break but not return

return false;
}

$group = $this->large_words[ $group_key ];
$group_length = strlen( $group );
$at = 0;
while ( $at < $group_length ) {
$token_length = ord( $group[ $at++ ] );
$token = substr( $group, $at, $token_length );

if ( 0 === substr_compare( $text, $token, $offset + self::KEY_LENGTH, $token_length ) ) {
return $group_key . $token;
}

$at += $token_length;
}
}

// Perhaps a short word then.
$small_text = str_pad( substr( $text, $offset, self::KEY_LENGTH ), self::KEY_LENGTH, "\x00" );
$at = strpos( $this->small_words, $small_text );

return false !== $at
? rtrim( substr( $this->small_words, $at, self::KEY_LENGTH ), "\x00" )
: false;
}

public function precomputed_php_source_table( $indent = "\t" ) {
$i1 = $indent;
$i2 = $indent . $indent;

$output = self::class . "::from_precomputed_table(\n";
$output .= $i1 . "array(\n";

foreach ( $this->large_words as $prefix => $group ) {
$comment_line = "{$i2}//";
$data_line = "{$i2}'{$prefix}' => \"";
$at = 0;
while ( $at < strlen( $group ) ) {
$length = ord( $group[ $at++ ] );
$digits = str_pad( dechex( $length ), 2, '0', STR_PAD_LEFT );
$token = substr( $group, $at, $length );
$at += $length;

$comment_line .= " &{$prefix}{$token}";
$data_line .= "\\x{$digits}{$token}";
}
$comment_line .= "\n";
$data_line .= "\",\n";

$output .= $comment_line;
$output .= $data_line;
}

$output .= "{$i1}),\n";
$small_text = str_replace( "\x00", '\x00', $this->small_words );
$output .= "{$i1}'{$small_text}'\n";
$output .= ");\n";

return $output;
}

private static function longest_first_then_alphabetical( $a, $b ) {
if ( $a === $b ) {
return 0;
}

$la = strlen( $a );
$lb = strlen( $b );

// Longer strings are less-than for comparison's sake.
if ( $la !== $lb ) {
return $lb - $la;
}

return strcmp( $a, $b );
}
}
Loading
Loading