This is inside a block!"
*
* @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;
/**
* 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();
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