smst-wp-site/swag/www/wordpress/wp-includes/class-wp-block-parser.php

556 lines
15 KiB
PHP
Raw Permalink Normal View History

2023-05-24 01:43:13 +00:00
<?php
/**
* Block Serialization Parser
*
* @package WordPress
*/
/**
* Class WP_Block_Parser_Block
*
* Holds the block structure in memory
*
* @since 5.0.0
*/
class WP_Block_Parser_Block {
/**
* Name of block
*
* @example "core/paragraph"
*
* @since 5.0.0
* @var string
*/
public $blockName;
/**
* Optional set of attributes from block comment delimiters
*
* @example null
* @example array( 'columns' => 3 )
*
* @since 5.0.0
* @var array|null
*/
public $attrs;
/**
* List of inner blocks (of this same class)
*
* @since 5.0.0
* @var WP_Block_Parser_Block[]
*/
public $innerBlocks;
/**
* Resultant HTML from inside block comment delimiters
* after removing inner blocks
*
* @example "...Just <!-- wp:test /--> testing..." -> "Just testing..."
*
* @since 5.0.0
* @var string
*/
public $innerHTML;
/**
* List of string fragments and null markers where inner blocks were found
*
* @example array(
* 'innerHTML' => 'BeforeInnerAfter',
* 'innerBlocks' => array( block, block ),
* 'innerContent' => array( 'Before', null, 'Inner', null, 'After' ),
* )
*
* @since 4.2.0
* @var array
*/
public $innerContent;
/**
* Constructor.
*
* Will populate object properties from the provided arguments.
*
* @since 5.0.0
*
* @param string $name Name of block.
* @param array $attrs Optional set of attributes from block comment delimiters.
* @param array $innerBlocks List of inner blocks (of this same class).
* @param string $innerHTML Resultant HTML from inside block comment delimiters after removing inner blocks.
* @param array $innerContent List of string fragments and null markers where inner blocks were found.
*/
public function __construct( $name, $attrs, $innerBlocks, $innerHTML, $innerContent ) {
$this->blockName = $name;
$this->attrs = $attrs;
$this->innerBlocks = $innerBlocks;
$this->innerHTML = $innerHTML;
$this->innerContent = $innerContent;
}
}
/**
* Class WP_Block_Parser_Frame
*
* Holds partial blocks in memory while parsing
*
* @internal
* @since 5.0.0
*/
class WP_Block_Parser_Frame {
/**
* Full or partial block
*
* @since 5.0.0
* @var WP_Block_Parser_Block
*/
public $block;
/**
* Byte offset into document for start of parse token
*
* @since 5.0.0
* @var int
*/
public $token_start;
/**
* Byte length of entire parse token string
*
* @since 5.0.0
* @var int
*/
public $token_length;
/**
* Byte offset into document for after parse token ends
* (used during reconstruction of stack into parse production)
*
* @since 5.0.0
* @var int
*/
public $prev_offset;
/**
* Byte offset into document where leading HTML before token starts
*
* @since 5.0.0
* @var int
*/
public $leading_html_start;
/**
* Constructor
*
* Will populate object properties from the provided arguments.
*
* @since 5.0.0
*
* @param WP_Block_Parser_Block $block Full or partial block.
* @param int $token_start Byte offset into document for start of parse token.
* @param int $token_length Byte length of entire parse token string.
* @param int $prev_offset Byte offset into document for after parse token ends.
* @param int $leading_html_start Byte offset into document where leading HTML before token starts.
*/
public function __construct( $block, $token_start, $token_length, $prev_offset = null, $leading_html_start = null ) {
$this->block = $block;
$this->token_start = $token_start;
$this->token_length = $token_length;
$this->prev_offset = isset( $prev_offset ) ? $prev_offset : $token_start + $token_length;
$this->leading_html_start = $leading_html_start;
}
}
/**
* Class WP_Block_Parser
*
* Parses a document and constructs a list of parsed block objects
*
* @since 5.0.0
* @since 4.0.0 returns arrays not objects, all attributes are arrays
*/
class WP_Block_Parser {
/**
* Input document being parsed
*
* @example "Pre-text\n<!-- wp:paragraph -->This is inside a block!<!-- /wp:paragraph -->"
*
* @since 5.0.0
* @var string
*/
public $document;
/**
* Tracks parsing progress through document
*
* @since 5.0.0
* @var int
*/
public $offset;
/**
* List of parsed blocks
*
* @since 5.0.0
* @var WP_Block_Parser_Block[]
*/
public $output;
/**
* Stack of partially-parsed structures in memory during parse
*
* @since 5.0.0
* @var WP_Block_Parser_Frame[]
*/
public $stack;
/**
* Empty associative array, here due to PHP quirks
*
* @since 4.4.0
* @var array empty associative array
*/
public $empty_attrs;
/**
* Parses a document and returns a list of block structures
*
* When encountering an invalid parse will return a best-effort
* parse. In contrast to the specification parser this does not
* return an error on invalid inputs.
*
* @since 5.0.0
*
* @param string $document Input document being parsed.
* @return array[]
*/
public function parse( $document ) {
$this->document = $document;
$this->offset = 0;
$this->output = array();
$this->stack = array();
$this->empty_attrs = json_decode( '{}', true );
while ( $this->proceed() ) {
continue;
}
return $this->output;
}
/**
* Processes the next token from the input document
* and returns whether to proceed eating more tokens
*
* This is the "next step" function that essentially
* takes a token as its input and decides what to do
* with that token before descending deeper into a
* nested block tree or continuing along the document
* or breaking out of a level of nesting.
*
* @internal
* @since 5.0.0
* @return bool
*/
public function proceed() {
$next_token = $this->next_token();
list( $token_type, $block_name, $attrs, $start_offset, $token_length ) = $next_token;
$stack_depth = count( $this->stack );
// we may have some HTML soup before the next block.
$leading_html_start = $start_offset > $this->offset ? $this->offset : null;
switch ( $token_type ) {
case 'no-more-tokens':
// if not in a block then flush output.
if ( 0 === $stack_depth ) {
$this->add_freeform();
return false;
}
/*
* Otherwise we have a problem
* This is an error
*
* we have options
* - treat it all as freeform text
* - assume an implicit closer (easiest when not nesting)
*/
// for the easy case we'll assume an implicit closer.
if ( 1 === $stack_depth ) {
$this->add_block_from_stack();
return false;
}
/*
* for the nested case where it's more difficult we'll
* have to assume that multiple closers are missing
* and so we'll collapse the whole stack piecewise
*/
while ( 0 < count( $this->stack ) ) {
$this->add_block_from_stack();
}
return false;
case 'void-block':
/*
* easy case is if we stumbled upon a void block
* in the top-level of the document
*/
if ( 0 === $stack_depth ) {
if ( isset( $leading_html_start ) ) {
$this->output[] = (array) $this->freeform(
substr(
$this->document,
$leading_html_start,
$start_offset - $leading_html_start
)
);
}
$this->output[] = (array) new WP_Block_Parser_Block( $block_name, $attrs, array(), '', array() );
$this->offset = $start_offset + $token_length;
return true;
}
// otherwise we found an inner block.
$this->add_inner_block(
new WP_Block_Parser_Block( $block_name, $attrs, array(), '', array() ),
$start_offset,
$token_length
);
$this->offset = $start_offset + $token_length;
return true;
case 'block-opener':
// track all newly-opened blocks on the stack.
array_push(
$this->stack,
new WP_Block_Parser_Frame(
new WP_Block_Parser_Block( $block_name, $attrs, array(), '', array() ),
$start_offset,
$token_length,
$start_offset + $token_length,
$leading_html_start
)
);
$this->offset = $start_offset + $token_length;
return true;
case 'block-closer':
/*
* if we're missing an opener we're in trouble
* This is an error
*/
if ( 0 === $stack_depth ) {
/*
* we have options
* - assume an implicit opener
* - assume _this_ is the opener
* - give up and close out the document
*/
$this->add_freeform();
return false;
}
// if we're not nesting then this is easy - close the block.
if ( 1 === $stack_depth ) {
$this->add_block_from_stack( $start_offset );
$this->offset = $start_offset + $token_length;
return true;
}
/*
* otherwise we're nested and we have to close out the current
* block and add it as a new innerBlock to the parent
*/
$stack_top = array_pop( $this->stack );
$html = substr( $this->document, $stack_top->prev_offset, $start_offset - $stack_top->prev_offset );
$stack_top->block->innerHTML .= $html;
$stack_top->block->innerContent[] = $html;
$stack_top->prev_offset = $start_offset + $token_length;
$this->add_inner_block(
$stack_top->block,
$stack_top->token_start,
$stack_top->token_length,
$start_offset + $token_length
);
$this->offset = $start_offset + $token_length;
return true;
default:
// This is an error.
$this->add_freeform();
return false;
}
}
/**
* Scans the document from where we last left off
* and finds the next valid token to parse if it exists
*
* Returns the type of the find: kind of find, block information, attributes
*
* @internal
* @since 5.0.0
* @since 4.6.1 fixed a bug in attribute parsing which caused catastrophic backtracking on invalid block comments
* @return array
*/
public function next_token() {
$matches = null;
/*
* aye the magic
* we're using a single RegExp to tokenize the block comment delimiters
* we're also using a trick here because the only difference between a
* block opener and a block closer is the leading `/` before `wp:` (and
* a closer has no attributes). we can trap them both and process the
* match back in PHP to see which one it was.
*/
$has_match = preg_match(
'/<!--\s+(?P<closer>\/)?wp:(?P<namespace>[a-z][a-z0-9_-]*\/)?(?P<name>[a-z][a-z0-9_-]*)\s+(?P<attrs>{(?:(?:[^}]+|}+(?=})|(?!}\s+\/?-->).)*+)?}\s+)?(?P<void>\/)?-->/s',
$this->document,
$matches,
PREG_OFFSET_CAPTURE,
$this->offset
);
// if we get here we probably have catastrophic backtracking or out-of-memory in the PCRE.
if ( false === $has_match ) {
return array( 'no-more-tokens', null, null, null, null );
}
// we have no more tokens.
if ( 0 === $has_match ) {
return array( 'no-more-tokens', null, null, null, null );
}
list( $match, $started_at ) = $matches[0];
$length = strlen( $match );
$is_closer = isset( $matches['closer'] ) && -1 !== $matches['closer'][1];
$is_void = isset( $matches['void'] ) && -1 !== $matches['void'][1];
$namespace = $matches['namespace'];
$namespace = ( isset( $namespace ) && -1 !== $namespace[1] ) ? $namespace[0] : 'core/';
$name = $namespace . $matches['name'][0];
$has_attrs = isset( $matches['attrs'] ) && -1 !== $matches['attrs'][1];
/*
* Fun fact! It's not trivial in PHP to create "an empty associative array" since all arrays
* are associative arrays. If we use `array()` we get a JSON `[]`
*/
$attrs = $has_attrs
? json_decode( $matches['attrs'][0], /* as-associative */ true )
: $this->empty_attrs;
/*
* This state isn't allowed
* This is an error
*/
if ( $is_closer && ( $is_void || $has_attrs ) ) {
// we can ignore them since they don't hurt anything.
}
if ( $is_void ) {
return array( 'void-block', $name, $attrs, $started_at, $length );
}
if ( $is_closer ) {
return array( 'block-closer', $name, null, $started_at, $length );
}
return array( 'block-opener', $name, $attrs, $started_at, $length );
}
/**
* Returns a new block object for freeform HTML
*
* @internal
* @since 3.9.0
*
* @param string $innerHTML HTML content of block.
* @return WP_Block_Parser_Block freeform block object.
*/
public function freeform( $innerHTML ) {
return new WP_Block_Parser_Block( null, $this->empty_attrs, array(), $innerHTML, array( $innerHTML ) );
}
/**
* Pushes a length of text from the input document
* to the output list as a freeform block.
*
* @internal
* @since 5.0.0
* @param null $length how many bytes of document text to output.
*/
public function add_freeform( $length = null ) {
$length = $length ? $length : strlen( $this->document ) - $this->offset;
if ( 0 === $length ) {
return;
}
$this->output[] = (array) $this->freeform( substr( $this->document, $this->offset, $length ) );
}
/**
* Given a block structure from memory pushes
* a new block to the output list.
*
* @internal
* @since 5.0.0
* @param WP_Block_Parser_Block $block The block to add to the output.
* @param int $token_start Byte offset into the document where the first token for the block starts.
* @param int $token_length Byte length of entire block from start of opening token to end of closing token.
* @param int|null $last_offset Last byte offset into document if continuing form earlier output.
*/
public function add_inner_block( WP_Block_Parser_Block $block, $token_start, $token_length, $last_offset = null ) {
$parent = $this->stack[ count( $this->stack ) - 1 ];
$parent->block->innerBlocks[] = (array) $block;
$html = substr( $this->document, $parent->prev_offset, $token_start - $parent->prev_offset );
if ( ! empty( $html ) ) {
$parent->block->innerHTML .= $html;
$parent->block->innerContent[] = $html;
}
$parent->block->innerContent[] = null;
$parent->prev_offset = $last_offset ? $last_offset : $token_start + $token_length;
}
/**
* Pushes the top block from the parsing stack to the output list.
*
* @internal
* @since 5.0.0
* @param int|null $end_offset byte offset into document for where we should stop sending text output as HTML.
*/
public function add_block_from_stack( $end_offset = null ) {
$stack_top = array_pop( $this->stack );
$prev_offset = $stack_top->prev_offset;
$html = isset( $end_offset )
? substr( $this->document, $prev_offset, $end_offset - $prev_offset )
: substr( $this->document, $prev_offset );
if ( ! empty( $html ) ) {
$stack_top->block->innerHTML .= $html;
$stack_top->block->innerContent[] = $html;
}
if ( isset( $stack_top->leading_html_start ) ) {
$this->output[] = (array) $this->freeform(
substr(
$this->document,
$stack_top->leading_html_start,
$stack_top->token_start - $stack_top->leading_html_start
)
);
}
$this->output[] = (array) $stack_top->block;
}
}