Chris@0: <?php
Chris@4: 
Chris@0: namespace Masterminds\HTML5\Parser;
Chris@0: 
Chris@4: use Masterminds\HTML5\Exception;
Chris@4: 
Chris@0: /**
Chris@4:  * The scanner scans over a given data input to react appropriately to characters.
Chris@0:  */
Chris@0: class Scanner
Chris@0: {
Chris@0:     const CHARS_HEX = 'abcdefABCDEF01234567890';
Chris@0:     const CHARS_ALNUM = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ01234567890';
Chris@0:     const CHARS_ALPHA = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';
Chris@0: 
Chris@4:     /**
Chris@4:      * The string data we're parsing.
Chris@4:      */
Chris@4:     private $data;
Chris@0: 
Chris@4:     /**
Chris@4:      * The current integer byte position we are in $data.
Chris@4:      */
Chris@4:     private $char;
Chris@4: 
Chris@4:     /**
Chris@4:      * Length of $data; when $char === $data, we are at the end-of-file.
Chris@4:      */
Chris@4:     private $EOF;
Chris@4: 
Chris@4:     /**
Chris@4:      * Parse errors.
Chris@4:      */
Chris@4:     public $errors = array();
Chris@0: 
Chris@0:     /**
Chris@0:      * Create a new Scanner.
Chris@0:      *
Chris@4:      * @param string $data     Data to parse.
Chris@4:      * @param string $encoding The encoding to use for the data.
Chris@4:      *
Chris@4:      * @throws Exception If the given data cannot be encoded to UTF-8.
Chris@0:      */
Chris@4:     public function __construct($data, $encoding = 'UTF-8')
Chris@0:     {
Chris@4:         if ($data instanceof InputStream) {
Chris@4:             @trigger_error('InputStream objects are deprecated since version 2.4 and will be removed in 3.0. Use strings instead.', E_USER_DEPRECATED);
Chris@4:             $data = (string) $data;
Chris@4:         }
Chris@4: 
Chris@4:         $data = UTF8Utils::convertToUTF8($data, $encoding);
Chris@4: 
Chris@4:         // There is good reason to question whether it makes sense to
Chris@4:         // do this here, since most of these checks are done during
Chris@4:         // parsing, and since this check doesn't actually *do* anything.
Chris@4:         $this->errors = UTF8Utils::checkForIllegalCodepoints($data);
Chris@4: 
Chris@4:         $data = $this->replaceLinefeeds($data);
Chris@4: 
Chris@4:         $this->data = $data;
Chris@4:         $this->char = 0;
Chris@4:         $this->EOF = strlen($data);
Chris@4:     }
Chris@4: 
Chris@4:     /**
Chris@4:      * Check if upcomming chars match the given sequence.
Chris@4:      *
Chris@4:      * This will read the stream for the $sequence. If it's
Chris@4:      * found, this will return true. If not, return false.
Chris@4:      * Since this unconsumes any chars it reads, the caller
Chris@4:      * will still need to read the next sequence, even if
Chris@4:      * this returns true.
Chris@4:      *
Chris@4:      * Example: $this->scanner->sequenceMatches('</script>') will
Chris@4:      * see if the input stream is at the start of a
Chris@4:      * '</script>' string.
Chris@4:      *
Chris@4:      * @param string $sequence
Chris@4:      * @param bool   $caseSensitive
Chris@4:      *
Chris@4:      * @return bool
Chris@4:      */
Chris@4:     public function sequenceMatches($sequence, $caseSensitive = true)
Chris@4:     {
Chris@4:         $portion = substr($this->data, $this->char, strlen($sequence));
Chris@4: 
Chris@4:         return $caseSensitive ? $portion === $sequence : 0 === strcasecmp($portion, $sequence);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the current position.
Chris@0:      *
Chris@0:      * @return int The current intiger byte position.
Chris@0:      */
Chris@0:     public function position()
Chris@0:     {
Chris@4:         return $this->char;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Take a peek at the next character in the data.
Chris@0:      *
Chris@0:      * @return string The next character.
Chris@0:      */
Chris@0:     public function peek()
Chris@0:     {
Chris@4:         if (($this->char + 1) <= $this->EOF) {
Chris@4:             return $this->data[$this->char + 1];
Chris@4:         }
Chris@4: 
Chris@4:         return false;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the next character.
Chris@0:      * Note: This advances the pointer.
Chris@0:      *
Chris@0:      * @return string The next character.
Chris@0:      */
Chris@0:     public function next()
Chris@0:     {
Chris@4:         ++$this->char;
Chris@4: 
Chris@4:         if ($this->char < $this->EOF) {
Chris@4:             return $this->data[$this->char];
Chris@0:         }
Chris@0: 
Chris@0:         return false;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the current character.
Chris@0:      * Note, this does not advance the pointer.
Chris@0:      *
Chris@0:      * @return string The current character.
Chris@0:      */
Chris@0:     public function current()
Chris@0:     {
Chris@4:         if ($this->char < $this->EOF) {
Chris@4:             return $this->data[$this->char];
Chris@0:         }
Chris@0: 
Chris@0:         return false;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Silently consume N chars.
Chris@4:      *
Chris@4:      * @param int $count
Chris@0:      */
Chris@0:     public function consume($count = 1)
Chris@0:     {
Chris@4:         $this->char += $count;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Unconsume some of the data.
Chris@0:      * This moves the data pointer backwards.
Chris@0:      *
Chris@4:      * @param int $howMany The number of characters to move the pointer back.
Chris@0:      */
Chris@0:     public function unconsume($howMany = 1)
Chris@0:     {
Chris@4:         if (($this->char - $howMany) >= 0) {
Chris@4:             $this->char -= $howMany;
Chris@4:         }
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the next group of that contains hex characters.
Chris@0:      * Note, along with getting the characters the pointer in the data will be
Chris@0:      * moved as well.
Chris@0:      *
Chris@0:      * @return string The next group that is hex characters.
Chris@0:      */
Chris@0:     public function getHex()
Chris@0:     {
Chris@4:         return $this->doCharsWhile(static::CHARS_HEX);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the next group of characters that are ASCII Alpha characters.
Chris@0:      * Note, along with getting the characters the pointer in the data will be
Chris@0:      * moved as well.
Chris@0:      *
Chris@0:      * @return string The next group of ASCII alpha characters.
Chris@0:      */
Chris@0:     public function getAsciiAlpha()
Chris@0:     {
Chris@4:         return $this->doCharsWhile(static::CHARS_ALPHA);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the next group of characters that are ASCII Alpha characters and numbers.
Chris@0:      * Note, along with getting the characters the pointer in the data will be
Chris@0:      * moved as well.
Chris@0:      *
Chris@0:      * @return string The next group of ASCII alpha characters and numbers.
Chris@0:      */
Chris@0:     public function getAsciiAlphaNum()
Chris@0:     {
Chris@4:         return $this->doCharsWhile(static::CHARS_ALNUM);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the next group of numbers.
Chris@0:      * Note, along with getting the characters the pointer in the data will be
Chris@0:      * moved as well.
Chris@0:      *
Chris@0:      * @return string The next group of numbers.
Chris@0:      */
Chris@0:     public function getNumeric()
Chris@0:     {
Chris@4:         return $this->doCharsWhile('0123456789');
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Consume whitespace.
Chris@4:      * Whitespace in HTML5 is: formfeed, tab, newline, space.
Chris@0:      *
Chris@4:      * @return int The length of the matched whitespaces.
Chris@0:      */
Chris@0:     public function whitespace()
Chris@0:     {
Chris@4:         if ($this->char >= $this->EOF) {
Chris@4:             return false;
Chris@4:         }
Chris@4: 
Chris@4:         $len = strspn($this->data, "\n\t\f ", $this->char);
Chris@4: 
Chris@4:         $this->char += $len;
Chris@4: 
Chris@4:         return $len;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Returns the current line that is being consumed.
Chris@0:      *
Chris@0:      * @return int The current line number.
Chris@0:      */
Chris@0:     public function currentLine()
Chris@0:     {
Chris@4:         if (empty($this->EOF) || 0 === $this->char) {
Chris@4:             return 1;
Chris@4:         }
Chris@4: 
Chris@4:         // Add one to $this->char because we want the number for the next
Chris@4:         // byte to be processed.
Chris@4:         return substr_count($this->data, "\n", 0, min($this->char, $this->EOF)) + 1;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Read chars until something in the mask is encountered.
Chris@4:      *
Chris@4:      * @param string $mask
Chris@4:      *
Chris@4:      * @return mixed
Chris@0:      */
Chris@0:     public function charsUntil($mask)
Chris@0:     {
Chris@4:         return $this->doCharsUntil($mask);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Read chars as long as the mask matches.
Chris@4:      *
Chris@4:      * @param string $mask
Chris@4:      *
Chris@4:      * @return int
Chris@0:      */
Chris@0:     public function charsWhile($mask)
Chris@0:     {
Chris@4:         return $this->doCharsWhile($mask);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Returns the current column of the current line that the tokenizer is at.
Chris@0:      *
Chris@0:      * Newlines are column 0. The first char after a newline is column 1.
Chris@0:      *
Chris@0:      * @return int The column number.
Chris@0:      */
Chris@0:     public function columnOffset()
Chris@0:     {
Chris@4:         // Short circuit for the first char.
Chris@4:         if (0 === $this->char) {
Chris@4:             return 0;
Chris@4:         }
Chris@4: 
Chris@4:         // strrpos is weird, and the offset needs to be negative for what we
Chris@4:         // want (i.e., the last \n before $this->char). This needs to not have
Chris@4:         // one (to make it point to the next character, the one we want the
Chris@4:         // position of) added to it because strrpos's behaviour includes the
Chris@4:         // final offset byte.
Chris@4:         $backwardFrom = $this->char - 1 - strlen($this->data);
Chris@4:         $lastLine = strrpos($this->data, "\n", $backwardFrom);
Chris@4: 
Chris@4:         // However, for here we want the length up until the next byte to be
Chris@4:         // processed, so add one to the current byte ($this->char).
Chris@4:         if (false !== $lastLine) {
Chris@4:             $findLengthOf = substr($this->data, $lastLine + 1, $this->char - 1 - $lastLine);
Chris@4:         } else {
Chris@4:             // After a newline.
Chris@4:             $findLengthOf = substr($this->data, 0, $this->char);
Chris@4:         }
Chris@4: 
Chris@4:         return UTF8Utils::countChars($findLengthOf);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get all characters until EOF.
Chris@0:      *
Chris@0:      * This consumes characters until the EOF.
Chris@0:      *
Chris@0:      * @return int The number of characters remaining.
Chris@0:      */
Chris@0:     public function remainingChars()
Chris@0:     {
Chris@4:         if ($this->char < $this->EOF) {
Chris@4:             $data = substr($this->data, $this->char);
Chris@4:             $this->char = $this->EOF;
Chris@4: 
Chris@4:             return $data;
Chris@4:         }
Chris@4: 
Chris@4:         return ''; // false;
Chris@4:     }
Chris@4: 
Chris@4:     /**
Chris@4:      * Replace linefeed characters according to the spec.
Chris@4:      *
Chris@4:      * @param $data
Chris@4:      *
Chris@4:      * @return string
Chris@4:      */
Chris@4:     private function replaceLinefeeds($data)
Chris@4:     {
Chris@4:         /*
Chris@4:          * U+000D CARRIAGE RETURN (CR) characters and U+000A LINE FEED (LF) characters are treated specially.
Chris@4:          * Any CR characters that are followed by LF characters must be removed, and any CR characters not
Chris@4:          * followed by LF characters must be converted to LF characters. Thus, newlines in HTML DOMs are
Chris@4:          * represented by LF characters, and there are never any CR characters in the input to the tokenization
Chris@4:          * stage.
Chris@4:          */
Chris@4:         $crlfTable = array(
Chris@4:             "\0" => "\xEF\xBF\xBD",
Chris@4:             "\r\n" => "\n",
Chris@4:             "\r" => "\n",
Chris@4:         );
Chris@4: 
Chris@4:         return strtr($data, $crlfTable);
Chris@4:     }
Chris@4: 
Chris@4:     /**
Chris@4:      * Read to a particular match (or until $max bytes are consumed).
Chris@4:      *
Chris@4:      * This operates on byte sequences, not characters.
Chris@4:      *
Chris@4:      * Matches as far as possible until we reach a certain set of bytes
Chris@4:      * and returns the matched substring.
Chris@4:      *
Chris@4:      * @param string $bytes Bytes to match.
Chris@4:      * @param int    $max   Maximum number of bytes to scan.
Chris@4:      *
Chris@4:      * @return mixed Index or false if no match is found. You should use strong
Chris@4:      *               equality when checking the result, since index could be 0.
Chris@4:      */
Chris@4:     private function doCharsUntil($bytes, $max = null)
Chris@4:     {
Chris@4:         if ($this->char >= $this->EOF) {
Chris@4:             return false;
Chris@4:         }
Chris@4: 
Chris@4:         if (0 === $max || $max) {
Chris@4:             $len = strcspn($this->data, $bytes, $this->char, $max);
Chris@4:         } else {
Chris@4:             $len = strcspn($this->data, $bytes, $this->char);
Chris@4:         }
Chris@4: 
Chris@4:         $string = (string) substr($this->data, $this->char, $len);
Chris@4:         $this->char += $len;
Chris@4: 
Chris@4:         return $string;
Chris@4:     }
Chris@4: 
Chris@4:     /**
Chris@4:      * Returns the string so long as $bytes matches.
Chris@4:      *
Chris@4:      * Matches as far as possible with a certain set of bytes
Chris@4:      * and returns the matched substring.
Chris@4:      *
Chris@4:      * @param string $bytes A mask of bytes to match. If ANY byte in this mask matches the
Chris@4:      *                      current char, the pointer advances and the char is part of the
Chris@4:      *                      substring.
Chris@4:      * @param int    $max   The max number of chars to read.
Chris@4:      *
Chris@4:      * @return string
Chris@4:      */
Chris@4:     private function doCharsWhile($bytes, $max = null)
Chris@4:     {
Chris@4:         if ($this->char >= $this->EOF) {
Chris@4:             return false;
Chris@4:         }
Chris@4: 
Chris@4:         if (0 === $max || $max) {
Chris@4:             $len = strspn($this->data, $bytes, $this->char, $max);
Chris@4:         } else {
Chris@4:             $len = strspn($this->data, $bytes, $this->char);
Chris@4:         }
Chris@4: 
Chris@4:         $string = (string) substr($this->data, $this->char, $len);
Chris@4:         $this->char += $len;
Chris@4: 
Chris@4:         return $string;
Chris@0:     }
Chris@0: }