Re: anyone care to explain the logic behind this output ....
| From: | David Tulloh | Date: | Sat, 18 Feb 2006 00:18:04 +0000 |
| Subject: | Re: anyone care to explain the logic behind this output .... | ||
| References: | 1 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-230673@lists.php.net to get a copy of this message | ||
(Don't you hate it when people forget to post back to the list...)
The secret is actually hidden in the docs,
http://php.net/manual/en/language.operators.comparison.php
When comparing $a < $b, you compare the first value of $a (0) to the
value with the same key in $b (1). 0 < 1 --> true
When you compare $b < $a, the first value of $b is 0, with a key of 1.
The value in $a is 1, 0 < 1 --> true
The discussion on php-dev, where you found this example, reveals that
PHP makes the assumption that $a > $b == $b < $a. This means that they
can reuse the less than function for greater than calls.
In this case the assumption isn't actually valid. $a > $b should be 0 >
1 --> false.
David
Jochem Maas wrote:
> THIS CODE
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> php -r '
> $a = array(0, 1);
> $b = array(1 => 0, 0 => 1);
> var_dump($a < $b); // true
> var_dump($a > $b); // true
> var_dump($b < $a);
> var_dump($b > $a);
>
> echo "\n\$a:\n"; var_dump((bool)$a, (int)$a, (string)$a, intval($a),
> strval($a));
> echo "\n\$b:\n"; var_dump((bool)$b, (int)$b, (string)$b, intval($b),
> strval($b));
> '
>
> OUTPUTS (on php5.0.4):
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> bool(true)
> bool(true)
> bool(true)
> bool(true)
>
> $a:
> bool(true)
> int(1)
> string(5) "Array"
>
> $b:
> bool(true)
> int(1)
> string(5) "Array"
>
> WHICH MEANS THAT:
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> one the one hand $a is greater than AND less than $b
> but on the other hand casting $a OR $b to either a boolean,
> integer or string results in the exact same value. ie:
>
> php -r '
> $a = array(0, 1); $b = array(1 => 0, 0 => 1);
> var_dump( ((($a > $b) === ($b > $a)) === ((int)$a === (int)$b)) ); //
> WTF IT'S TRUE
> '
>
> weird? I think so - but then again I'd never test that array $a is
> greater than array $b because this is meaningless to me (in what way is $a
> greater - how is this quantified, what 'rules' determine 'greatness' in
> this context?)
>
> PS - changing the $b array to something else (anything else as far as i
> can tell)
> causes the weirdness to not occur - which gives me the impression this
> could be a bug,
> anyone else get this impression? for instance try changing the second
> line of
> the code above to (merely switching the order or the defined array
> elements):
>
> $b = array(0 => 1, 1 => 0);
>