Re: mySQL and SQLDriverConnect

Nick,

-----Original Message-----
>From: Nick Gorham <[email protected]>
>Sent: Jun 20, 2008 8:36 AM
>To: Igor Korot <[email protected]>
>Cc: Development issues and topics for unixODBC <[email protected]>
>Subject: Re: [unixODBC-dev] mySQL and SQLDriverConnect
>
>Igor Korot wrote:
>
>>>>      
>>>>
>>>By the way, when you say you explicitly call the non unicode function, 
>>>do you mean you call SQLDriverConnectA, becaue with UNICODE defined 
>>>SQLDriverConnect becomes SQLDreiverConnectW (as it does in windows)
>>>    
>>>
>>
>>retcode = SQLDriverConnect( m_hdbc, parentWnd, (SQLTCHAR FAR *)"DSN=test;", ...);
>>                                                ^^^^^^^^^^^^^^
>>
>>I am doing casting of the parameters.
>>In UNICODE version, this would become (SQLWCHAR FAR *).
>>
>>So, unless the linker is not smart, he should look at the proper signature of the
>>function.
>>
>>  
>>
>Not entirly correct
>
>(SQLTCHAR FAR *)"DSN=test;"
>
>when UNICODE is defined, will create a SQLWCHAR * pointer, but the thing it points to will still be a char string;
>
>I think thats the problem, you are passing the address 8bit string into a function thats expecting a 16bit sequence of values. Casting the pointer to a pointer of another type will not alter the thing thats being pointed to.
>
>Remember C will do what you tell it to, even if thats not what you want it to do :-)

Ok, it is very easy to check. ;-)
It is morning here and I'm at work. When I come home, I will enable ODBC logging and run my
program. Hopefully you are right, I was just banging my head over nothing and simply passing:
"DSN=test;" without casting should fix the issue.

Thank you.

P.S.: Sometimes it's good to do RTFM ;-)
>
>-- 
>Nick
>
>

_______________________________________________
unixODBC-dev mailing list
[email protected]
http://mail.easysoft.com/mailman/listinfo/unixodbc-dev