Re: [PECL-DEV] Re: [PDO] Breaking the PDO API

To me Wez's fix seems like the simplest solution.


Ilia Alshanetsky




On 15-Apr-09, at 7:52 PM, Wez Furlong wrote:

> Any objection to fixing this in the drivers themselves (as I  
> suggested)?
> It seems like the easiest fix, and doesn't introduce a massive  
> change to the code.
>
> --Wez.
>
> On Apr 15, 2009, at 7:47 PM, Matteo Beccati wrote:
>
>> Wez Furlong ha scritto:
>>> I found the patch... can you explain the nature of the crash and  
>>> which
>>> driver(s) it affects?
>>> Your patch is huge, and it seems to me that the right fix is to  
>>> prevent
>>> drivers from returning empty strings when the expectation is a LOB;
>>> either return NULL or make a LOB with no content.
>>
>> The pgsql driver is surely affected. The oci driver might be  
>> affected,
>> although I don't have enough experience with it to verify.
>>
>> From what I can see the problem is that the get_col API can set value
>> either to a (php_stream *) or a (char *). The check on value[0] was
>> added to fix http://bugs.php.net/bug.php?id=46274, to prevent the  
>> stream
>> related code to be run when an empty string is returned by the  
>> driver.
>>
>> As you've already seen my proposal was to an is_stream parameter to  
>> the
>> get_col API to make sure PDO knows what it's expected to handle.
>>
>>
>> Cheers
>>
>> -- 
>> Matteo Beccati
>>
>> OpenX - http://www.openx.org
>
>
> -- 
> PECL development discussion Mailing List (http://pecl.php.net/)
> To unsubscribe, visit: http://www.php.net/unsub.php
>