ntdll/file.c: Fix a possible NULL dereferencing.

[Nikolay Sivov]

Rob Shearman wrote:
> 2008/8/27 Nikolay Sivov <bunglehead at gmail.com>:
>   
>> Henri Verbeet wrote:
>>     
>>> 2008/8/27 Nikolay Sivov <bunglehead at gmail.com>:
>>>
>>>       
>>>> Markus Hitter wrote:
>>>>
>>>>         
>>>>> if (!attr || !attr->ObjectName) return STATUS_INVALID_PARAMETER;
>>>>>
>>>>>           
>>>> Shouldn't this be splitted? It isn't safe to rely on evaluation order.
>>>> Or is it a default compiler setting for us?
>>>>
>>>>
>>>>         
>>> This is well defined in C. !attr will be evaluated first, and if it
>>> evaluates to true, the rest of the condition won't be evaluated. This
>>> is also known as short-circuit evaluation.
>>>
>>>
>>>       
>> I thought it's possible to override this: not a evaluation order of
>> course but the completeness of
>> evaluation for debug purpose.
>>     
>
> All C89/C99-compliant compilers will always respect the order of
> evaluation of the operands of the logical AND and OR operators. I
> don't know what "debug purpose" you are referring to, but the code
> generated for "if (!attr || !attr->ObjectName)" is guaranteed by the C
> standard to never crash on a NULL pointer.
>   
By "debug purpose" I meant this situation: if a second operand of OR 
operator is a return value
of a some call could I force it to be called regardless of first operand 
value for example to
get some debug output inside this call.

P.S. I was confused because some years ago using Delphi compiler I've 
found an option called
something like "complete boolean evaluation" or something like that.

In a C89 draft I've found just now:
"...If the first operand compares unequal to 0, the second operand is 
not evaluated..."
So this question is closed now.