[PATCH] user32: check for valid pointer in CURSORICON_CreateIconFromBMI

[Wolfram Sang]

On 23/06/10 15:32, Alexandre Julliard wrote:
> Wolfram Sang <wolfram at the-dreams.de> writes:
> 
>> @@ -884,6 +884,10 @@ static HICON CURSORICON_CreateIconFromBMI( BITMAPINFO *bmi,
>>          return 0;
>>      }
>>  
>> +    /* A few Windows versions crash getting NULL, but most simply return it */
>> +    if (!bmi)
>> +        return 0;
> 
> That's not the right place for it, this is an internal function.

OK. Shall I remove the similar check in CURSORICON_CreateIconFromANI()
then in the second version of the patch?