Why not to use another function instead of modifying system function?
class Test
{
    public function toString($bDebug = FALSE)
    {
        return $bDebug ? 'This is the debugging output' : 'This is the default 
output';
    }
    public function __toString()
    {
        return $this->toString();
    }
}

Because there is already a method for that and because this is an extra call. When you are converting thousands of objects into strings this becomes noticeable. Using your example:

```
$iterations = 1000000;
$test       = new Test();
$this->runBenchmark('toString', function () use ($iterations, $test) {
    for ($i = 0; $i < $iterations; $i++) {
        $result = $test->toString();
    }
});
$this->runBenchmark('__toString', function () use ($iterations, $test) {
    for ($i = 0; $i < $iterations; $i++) {
        $result = $test->__toString();
    }
});
$this->runBenchmark('(string)', function () use ($iterations, $test) {
    for ($i = 0; $i < $iterations; $i++) {
        $result = (string) $test;
    }
});
```

testToString(toString)      Time: 2.008s Memory: 5.50Mb
testToString(__toString)    Time: 3.875s Memory: 5.50Mb
testToString((string))      Time: 4.173s Memory: 5.50Mb

The "why not to use another function instead" is almost twice slower. While the performance implication is always debatable, it's clear that allowing magic methods, __toString in particular, to accept default arguments has a real value for people who care to write dry and performant code.

This feature request doesn't make any sense to me.  Users are not
supposed to call magic methods explicitly (except maybe to
override them and call the parent's method).  Thus I'm closing as
WONTFIX.

If anybody still wants to see optional arguments for __toString(),
please pursue the RFC process[1].

[1] <https://wiki.php.net/rfc/howto>