This repository has been archived on 2023-10-21. You can view files and clone it, but cannot push or open issues or pull requests.
patchouli/lib/FWIF/FWIF.php
2020-12-24 03:25:10 +00:00

297 lines
12 KiB
PHP

<?php
namespace FWIF;
use DateTime;
use DateTimeInterface;
use DateTimeZone;
class FWIF {
public const CONTENT_TYPE = 'text/plain; charset=us-ascii'; // TODO: come up with a mime type
public const DEFAULT = 0;
public const DISCARD_MILLISECONDS = 0x01; // Always exclude the millisecond component from DateTime
public const TYPE_NULL = 0; // NULL, no data
public const TYPE_INTEGER = 0x01; // LEB128, implicit length
public const TYPE_FLOAT = 0x02; // double precision IEEE 754, fixed length of 8 bytes
public const TYPE_STRING = 0x03; // UTF-8 string, terminated with TRAILER
public const TYPE_BUFFER = 0x04; // Buffer with binary data, prefixed with a LEB128 length
public const TYPE_ARRAY = 0x05; // List of values, terminated with TRAILER
public const TYPE_OBJECT = 0x06; // List of values with ASCII names, terminated with TRAILER
public const TYPE_DATETIME = 0x07; // A gregorian year, month and day as well as an hour, minute and seconds component, fixed length of * bytes
public const TRAILER = 0xFF; // Termination byte
private const CODECS = [
self::TYPE_NULL => 'Null',
self::TYPE_INTEGER => 'Integer',
self::TYPE_FLOAT => 'Float',
self::TYPE_STRING => 'String',
self::TYPE_ARRAY => 'Array',
self::TYPE_OBJECT => 'Object',
self::TYPE_BUFFER => 'Buffer',
self::TYPE_DATETIME => 'DateTime',
];
private const UTF8 = '%^(?:' // https://www.w3.org/International/questions/qa-forms-utf-8.en
. '[\x09\x0A\x0D\x20-\x7E]' // ASCII
. '|[\xC2-\xDF][\x80-\xBF]' // non-overlong 2-byte
. '|\xE0[\xA0-\xBF][\x80-\xBF]' // excluding overlongs
. '|[\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}' // straight 3-byte
. '|\xED[\x80-\x9F][\x80-\xBF]' // excluding surrogates
. '|\xF0[\x90-\xBF][\x80-\xBF]{2}' // planes 1-3
. '|[\xF1-\xF3][\x80-\xBF]{3}' // planes 4-15
. '|\xF4[\x80-\x8F][\x80-\xBF]{2}' // plane 16
. ')*$%xs';
private static function isAssocArray($array): bool {
if(!is_array($array) || $array === [])
return false;
return array_keys($array) !== range(0, count($array) - 1);
}
// apparently this is faster than mb_check_encoding($string, 'utf-8');
// on PHP 7.1 on Windows at least, perhaps investigate this later
private static function isUTF8String(string $string): bool {
return preg_match(self::UTF8, $string) === 1;
}
private static function detectType($data, int $flags): int {
if(is_null($data))
return self::TYPE_NULL;
if(is_int($data))
return self::TYPE_INTEGER;
if(is_float($data))
return self::TYPE_FLOAT;
if(is_string($data))
return self::isUTF8String($data) ? self::TYPE_STRING : self::TYPE_BUFFER;
if(is_object($data) || self::isAssocArray($data)) {
if($data instanceof DateTimeInterface)
return self::TYPE_DATETIME;
return self::TYPE_OBJECT;
}
if(is_array($data))
return self::TYPE_ARRAY;
throw new FWIFUnsupportedTypeException(gettype($data));
}
public static function encode($data, int $flags = self::DEFAULT): string {
if($data instanceof FWIFSerializable)
$data = $data->fwifSerialize();
$type = self::detectType($data, $flags);
return chr($type) . self::{'encode' . self::CODECS[$type]}($data, $flags);
}
public static function decode(string $data, int $flags = self::DEFAULT) {
return self::decodeInternal(new FWIFDecodeStream($data), $flags);
}
private static function decodeInternal(FWIFDecodeStream $data, int $flags) {
$type = $data->readByte();
if(!array_key_exists($type, self::CODECS)) {
$hexType = dechex($type); $hexPos = dechex($data->getPosition());
throw new FWIFUnsupportedTypeException("Unsupported type {$type} (0x{$hexType}) at position {$data->getPosition()} (0x{$hexPos})");
}
return self::{'decode' . self::CODECS[$type]}($data, $flags);
}
private static function encodeNull($data, int $flags): string { return ''; }
private static function decodeNull(FWIFDecodeStream $data, int $flags) { return null; }
private static function encodeInteger(int $number, int $flags): string {
$packed = ''; $more = 1; $negative = $number < 0; $size = PHP_INT_SIZE * 8;
while($more) {
$byte = $number & 0x7F;
$number >>= 7;
if($negative)
$number |= (~0 << ($size - 7));
if((!$number && !($byte & 0x40)) || ($number === -1 && ($byte & 0x40)))
$more = 0;
else
$byte |= 0x80;
$packed .= chr($byte);
}
return $packed;
}
private static function decodeInteger(FWIFDecodeStream $data, int $flags): int {
$number = 0; $shift = 0; $o = 0; $size = PHP_INT_SIZE * 8;
do {
$byte = $data->readByte();
$number |= ($byte & 0x7F) << $shift;
$shift += 7;
} while($byte & 0x80);
if(($shift < $size) && ($byte & 0x40))
$number |= (~0 << $shift);
return $number;
}
private static function encodeFloat(float $number, int $flags): string {
return pack('E', $number);
}
private static function decodeFloat(FWIFDecodeStream $data, int $flags): float {
return unpack('E', $data->readString(8))[1];
}
private static function encodeString(string $string, int $flags): string {
$packed = ''; $string = unpack('C*', mb_convert_encoding($string, 'utf-8'));
foreach($string as $char)
$packed .= chr($char);
return $packed . chr(self::TRAILER);
}
private static function decodeAsciiString(FWIFDecodeStream $data, int $flags): string {
$string = '';
for(;;) {
$byte = $data->readByte();
if($byte === self::TRAILER)
break;
$string .= chr($byte);
}
return $string;
}
private static function decodeString(FWIFDecodeStream $data, int $flags): string { // This should decode based on the utf-8 spec rather than just
return mb_convert_encoding(self::decodeAsciiString($data, $flags), 'utf-8'); // grabbing the FF terminated string representation.
}
private static function encodeArray(array $array, int $flags): string {
$packed = '';
foreach($array as $value)
$packed .= self::encode($value, $flags);
return $packed . chr(self::TRAILER);
}
private static function decodeArray(FWIFDecodeStream $data, int $flags): array {
$array = [];
for(;;) {
if($data->readByte() === self::TRAILER)
break;
$data->stepBack();
$array[] = self::decodeInternal($data, $flags);
}
return $array;
}
private static function encodeObject($object, int $flags): string {
$packed = ''; $array = (array)$object;
foreach($array as $name => $value)
$packed .= $name . chr(self::TRAILER) . self::encode($value, $flags);
return $packed . chr(self::TRAILER);
}
private static function decodeObject(FWIFDecodeStream $data, int $flags): object {
$array = [];
for(;;) {
if($data->readByte() === self::TRAILER)
break;
$data->stepBack();
$name = self::decodeAsciiString($data, $flags);
$array[$name] = self::decodeInternal($data, $flags);
}
return (object)$array;
}
private static function encodeBuffer(string $buffer, int $flags): string {
return self::encodeInteger(strlen($buffer), $flags) . $buffer;
}
private static function decodeBuffer(FWIFDecodeStream $data, int $flags): string {
return $data->readString(self::decodeInteger($data, $flags));
}
private const DATETIME_FLAG_TIME = 0x01;
private const DATETIME_FLAG_MILLI = 0x02;
private const DATETIME_FLAG_MASK = 0x03;
private const DATETIME_YEAR_SIGN = 0x800000;
private const DATETIME_YEAR_MASK = 0x3FFF;
private const DATETIME_YEAR_SHIFT = 9;
private const DATETIME_MONTH_MASK = 0x0F;
private const DATETIME_MONTH_SHIFT = 5;
private const DATETIME_DAY_MASK = 0x1F;
private const DATETIME_HOUR_MASK = 0x1F;
private const DATETIME_HOUR_SHIFT = 12;
private const DATETIME_MINS_MASK = 0x3F;
private const DATETIME_MINS_SHIFT = 6;
private const DATETIME_SECS_MASK = 0x3F;
private const DATETIME_MILLI_MASK = 0xFFFFF;
/* One block Chunked
+--------+ +--------+ +--------+
0 | YYYYYYY| |YYYYYYYY| |FFF.....|
8 |YYYYYYYY| |YYYYYYYM| +--------+
12 | MMMM| |MMMDDDDD|
16 | DDDDD| +--------+
24 |FFFHHHHH| FHmS and U seperate Semicombined FHmSU
32 | mmmmmm| +--------+ +--------+ +--------+ +--------+
40 | SSSSSS| |FFFHHHHH| |UUUUUUUU| |FFFHHHHH| |UUUUUUUU|
48 |uuuuuuuu| |mmmmmmSS| |UUUUUUUU| OR |mmmmmmSS| |UUUUUUUU|
56 |uuuuuuuu| |SSSS....| |UUUU....| |SSSSUUUU| +--------+
64 |uuuuuuuu| +--------+ +--------+ +--------+
+--------+ In both situations the upper byte of FHmS would always be present for the flags.
Meaning the thing would take up at least 4 bytes and at most 9 bytes with 1 unused bit.
Most situations I'd imagine you wouldn't really care about the milliseconds,
in which you end up with 7 bytes with 5 unused bits.
Is there a benefit to this over just using .NET-style ticks?
Yes, it can be entirely handled using just 32-bit integers.
*/
private static function encodeDateTime(DateTimeInterface $dt, int $flags): string {
static $utc = null;
if($utc === null)
$utc = new DateTimeZone('utc');
if($dt->getTimezone()->getOffset($dt) !== 0)
$dt = DateTime::createFromInterface($dt)->setTimezone($utc);
$dtf = 0;
$year = (int)$dt->format('Y');
$month = (int)$dt->format('n');
$day = (int)$dt->format('j');
$subYear = $year < 0;
if($subYear)
$year = ~$year;
$ymd = $subYear ? self::DATETIME_YEAR_SIGN : 0;
$ymd |= ( $year & self::DATETIME_YEAR_MASK) << self::DATETIME_YEAR_SHIFT;
$ymd |= (($month - 1) & self::DATETIME_MONTH_MASK) << self::DATETIME_MONTH_SHIFT;
$ymd |= (($day - 1) & self::DATETIME_DAY_MASK);
var_dump(str_pad(decbin($ymd), 24, '0', STR_PAD_LEFT));
$hours = (int)$dt->format('G');
$mins = (int)$dt->format('i');
$secs = (int)$dt->format('s');
$millis = ($flags & self::DISCARD_MILLISECONDS) ? 0 : (int)$dt->format('u');
// Add year, month, day
if($hours > 0 || $mins > 0 || $secs > 0 || $millis > 0) {
$dtf |= self::DATETIME_FLAG_TIME;
$hms = 0;
$hms |= ($hours & self::DATETIME_HOUR_MASK) << self::DATETIME_HOUR_SHIFT;
$hms |= ($mins & self::DATETIME_MINS_MASK) << self::DATETIME_MINS_SHIFT;
$hms |= ($secs & self::DATETIME_SECS_MASK);
var_dump(str_pad(decbin($hms), 17, '0', STR_PAD_LEFT));
if($millis > 0) {
$dtf |= self::DATETIME_FLAG_MILLI;
$millis &= self::DATETIME_MILLI_MASK;
var_dump(str_pad(decbin($millis), 20, '0', STR_PAD_LEFT));
}
}
echo "\r\n";
return '';
}
private static function decodeDateTime(FWIFDecodeStream $data, int $flags): DateTimeInterface {
return new \DateTime();
}
}