On Thu, Mar 29, 2012 at 12:19 PM, Stuart Dallas <stuart@xxxxxxxx> wrote:
> On 29 Mar 2012, at 17:57, Arno Kuhl wrote:
>
>> I found automatic typecasting can be a bit of a gotcha.
>>
>>
>>
>> $sText = "this.is.a.test.text";
>>
>> if ( $pos = strpos($sText, "test") !== FALSE) {
>>
>> echo substr($sText, 0, $pos)."<".substr($sText, $pos,
>> strlen("test")).">".substr($sText, $pos+strlen("test"));
>>
>> }
>>
>>
>>
>> The code seems logical enough, and the expected result would be:
>>
>> this.is.a.<test>.text
>>
>>
>>
>> In fact it ends up being:
>>
>> t<his.>is.a.test.text
>>
>>
>>
>> The reason is $pos is typecast as TRUE, not int 10, presumably because it's
>> in the same scope as the boolean test.
>>
>> Then when $pos is later used as an int it's converted from TRUE to 1.
>>
>>
>>
>> You have to bracket the $pos setting to move it into its own scope to
>> prevent it being typecast:
>>
>> if ( ($pos = strpos($sText, "test")) !== FALSE) {
>>
>>
>>
>> No doubt it's mentioned somewhere in the php manual, I just never came
>> across it.
>>
>> Just thought I'd highlight one of the gotchas of auto typecasting for any
>> other simpletons like me.
>
> This is not due to typecasting, it's due to operator precedence.
>
> In PHP (and many other languages) the !== comparison operator will get done first, so the result of that comparison will be assigned to $pos. By introducing the brackets you're explicitly specifying the order.
>
> Details here: http://php.net/operators.precedence
>
> -Stuart
>
> --
> Stuart Dallas
> 3ft9 Ltd
> http://3ft9.com/
> --
> PHP General Mailing List (http://www.php.net/)
> To unsubscribe, visit: http://www.php.net/unsub.php
>
Interestingly enough (again due to precedence + LR processing) if you
reverse the order of the test:
if (false !== $pos = strpos($sText, "test"))
it works! But I'm not sure it's a good idea to rely on that, as the
extra bracketing is probably more obvious to the reader.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
