<?php
namespace MaxMind\Db\Reader;
// @codingStandardsIgnoreLine
// We subtract 1 from the log to protect against precision loss.
\define(__NAMESPACE__ . '\_MM_MAX_INT_BYTES', (log(PHP_INT_MAX, 2) - 1) / 8);
class Decoder
{
private $fileStream;
private $pointerBase;
private $pointerBaseByteSize;
// This is only used for unit testing
private $pointerTestHack;
private $switchByteOrder;
const _EXTENDED = 0;
const _POINTER = 1;
const _UTF8_STRING = 2;
const _DOUBLE = 3;
const _BYTES = 4;
const _UINT16 = 5;
const _UINT32 = 6;
const _MAP = 7;
const _INT32 = 8;
const _UINT64 = 9;
const _UINT128 = 10;
const _ARRAY = 11;
const _CONTAINER = 12;
const _END_MARKER = 13;
const _BOOLEAN = 14;
const _FLOAT = 15;
public function __construct(
$fileStream,
$pointerBase = 0,
$pointerTestHack = false
) {
$this->fileStream = $fileStream;
$this->pointerBase = $pointerBase;
$this->pointerBaseByteSize = $pointerBase > 0 ? log($pointerBase, 2) / 8 : 0;
$this->pointerTestHack = $pointerTestHack;
$this->switchByteOrder = $this->isPlatformLittleEndian();
}
public function decode($offset)
{
list(, $ctrlByte) = unpack(
'C',
Util::read($this->fileStream, $offset, 1)
);
++$offset;
$type = $ctrlByte >> 5;
// Pointers are a special case, we don't read the next $size bytes, we
// use the size to determine the length of the pointer and then follow
// it.
if ($type === self::_POINTER) {
list($pointer, $offset) = $this->decodePointer($ctrlByte, $offset);
// for unit testing
if ($this->pointerTestHack) {
return [$pointer];
}
list($result) = $this->decode($pointer);
return [$result, $offset];
}
if ($type === self::_EXTENDED) {
list(, $nextByte) = unpack(
'C',
Util::read($this->fileStream, $offset, 1)
);
$type = $nextByte + 7;
if ($type < 8) {
throw new InvalidDatabaseException(
'Something went horribly wrong in the decoder. An extended type '
. 'resolved to a type number < 8 ('
. $type
. ')'
);
}
++$offset;
}
list($size, $offset) = $this->sizeFromCtrlByte($ctrlByte, $offset);
return $this->decodeByType($type, $offset, $size);
}
private function decodeByType($type, $offset, $size)
{
switch ($type) {
case self::_MAP:
return $this->decodeMap($size, $offset);
case self::_ARRAY:
return $this->decodeArray($size, $offset);
case self::_BOOLEAN:
return [$this->decodeBoolean($size), $offset];
}
$newOffset = $offset + $size;
$bytes = Util::read($this->fileStream, $offset, $size);
switch ($type) {
case self::_BYTES:
case self::_UTF8_STRING:
return [$bytes, $newOffset];
case self::_DOUBLE:
$this->verifySize(8, $size);
return [$this->decodeDouble($bytes), $newOffset];
case self::_FLOAT:
$this->verifySize(4, $size);
return [$this->decodeFloat($bytes), $newOffset];
case self::_INT32:
return [$this->decodeInt32($bytes, $size), $newOffset];
case self::_UINT16:
case self::_UINT32:
case self::_UINT64:
case self::_UINT128:
return [$this->decodeUint($bytes, $size), $newOffset];
default:
throw new InvalidDatabaseException(
'Unknown or unexpected type: ' . $type
);
}
}
private function verifySize($expected, $actual)
{
if ($expected !== $actual) {
throw new InvalidDatabaseException(
"The MaxMind DB file's data section contains bad data (unknown data type or corrupt data)"
);
}
}
private function decodeArray($size, $offset)
{
$array = [];
for ($i = 0; $i < $size; ++$i) {
list($value, $offset) = $this->decode($offset);
array_push($array, $value);
}
return [$array, $offset];
}
private function decodeBoolean($size)
{
return $size === 0 ? false : true;
}
private function decodeDouble($bits)
{
// This assumes IEEE 754 doubles, but most (all?) modern platforms
// use them.
//
// We are not using the "E" format as that was only added in
// 7.0.15 and 7.1.1. As such, we must switch byte order on
// little endian machines.
list(, $double) = unpack('d', $this->maybeSwitchByteOrder($bits));
return $double;
}
private function decodeFloat($bits)
{
// This assumes IEEE 754 floats, but most (all?) modern platforms
// use them.
//
// We are not using the "G" format as that was only added in
// 7.0.15 and 7.1.1. As such, we must switch byte order on
// little endian machines.
list(, $float) = unpack('f', $this->maybeSwitchByteOrder($bits));
return $float;
}
private function decodeInt32($bytes, $size)
{
switch ($size) {
case 0:
return 0;
case 1:
case 2:
case 3:
$bytes = str_pad($bytes, 4, "\x00", STR_PAD_LEFT);
break;
case 4:
break;
default:
throw new InvalidDatabaseException(
"The MaxMind DB file's data section contains bad data (unknown data type or corrupt data)"
);
}
list(, $int) = unpack('l', $this->maybeSwitchByteOrder($bytes));
return $int;
}
private function decodeMap($size, $offset)
{
$map = [];
for ($i = 0; $i < $size; ++$i) {
list($key, $offset) = $this->decode($offset);
list($value, $offset) = $this->decode($offset);
$map[$key] = $value;
}
return [$map, $offset];
}
private function decodePointer($ctrlByte, $offset)
{
$pointerSize = (($ctrlByte >> 3) & 0x3) + 1;
$buffer = Util::read($this->fileStream, $offset, $pointerSize);
$offset = $offset + $pointerSize;
switch ($pointerSize) {
case 1:
$packed = (pack('C', $ctrlByte & 0x7)) . $buffer;
list(, $pointer) = unpack('n', $packed);
$pointer += $this->pointerBase;
break;
case 2:
$packed = "\x00" . (pack('C', $ctrlByte & 0x7)) . $buffer;
list(, $pointer) = unpack('N', $packed);
$pointer += $this->pointerBase + 2048;
break;
case 3:
$packed = (pack('C', $ctrlByte & 0x7)) . $buffer;
// It is safe to use 'N' here, even on 32 bit machines as the
// first bit is 0.
list(, $pointer) = unpack('N', $packed);
$pointer += $this->pointerBase + 526336;
break;
case 4:
// We cannot use unpack here as we might overflow on 32 bit
// machines
$pointerOffset = $this->decodeUint($buffer, $pointerSize);
$byteLength = $pointerSize + $this->pointerBaseByteSize;
if ($byteLength <= _MM_MAX_INT_BYTES) {
$pointer = $pointerOffset + $this->pointerBase;
} elseif (\extension_loaded('gmp')) {
$pointer = gmp_strval(gmp_add($pointerOffset, $this->pointerBase));
} elseif (\extension_loaded('bcmath')) {
$pointer = bcadd($pointerOffset, $this->pointerBase);
} else {
throw new \RuntimeException(
'The gmp or bcmath extension must be installed to read this database.'
);
}
}
return [$pointer, $offset];
}
private function decodeUint($bytes, $byteLength)
{
if ($byteLength === 0) {
return 0;
}
$integer = 0;
for ($i = 0; $i < $byteLength; ++$i) {
$part = \ord($bytes[$i]);
// We only use gmp or bcmath if the final value is too big
if ($byteLength <= _MM_MAX_INT_BYTES) {
$integer = ($integer << 8) + $part;
} elseif (\extension_loaded('gmp')) {
$integer = gmp_strval(gmp_add(gmp_mul($integer, 256), $part));
} elseif (\extension_loaded('bcmath')) {
$integer = bcadd(bcmul($integer, 256), $part);
} else {
throw new \RuntimeException(
'The gmp or bcmath extension must be installed to read this database.'
);
}
}
return $integer;
}
private function sizeFromCtrlByte($ctrlByte, $offset)
{
$size = $ctrlByte & 0x1f;
if ($size < 29) {
return [$size, $offset];
}
$bytesToRead = $size - 28;
$bytes = Util::read($this->fileStream, $offset, $bytesToRead);
if ($size === 29) {
$size = 29 + \ord($bytes);
} elseif ($size === 30) {
list(, $adjust) = unpack('n', $bytes);
$size = 285 + $adjust;
} elseif ($size > 30) {
list(, $adjust) = unpack('N', "\x00" . $bytes);
$size = ($adjust & (0x0FFFFFFF >> (32 - (8 * $bytesToRead))))
+ 65821;
}
return [$size, $offset + $bytesToRead];
}
private function maybeSwitchByteOrder($bytes)
{
return $this->switchByteOrder ? strrev($bytes) : $bytes;
}
private function isPlatformLittleEndian()
{
$testint = 0x00FF;
$packed = pack('S', $testint);
return $testint === current(unpack('v', $packed));
}
}