FreeTDS Development Group

[jklowden AT schemamania.org]

On Fri, Sep 09, 2011 at 04:18:37PM -0400, Bob Hetzel wrote:
> James,
> 
> Regarding re-trying the login with each protocol until it finds a good 
> one... is there any chance that'll cause an account lock-out if a place has 
> really tight security settings?
> 
> I guess the real question I'm asking is... how far will it get when it 
> fails due to the wrong tds version for that server?  If it fails before 
> sending the username/password then there would probably not be a problem. 
> After that point, it might be.  

http://msdn.microsoft.com/en-us/library/dd304019(v=PROT.13).aspx

I can't prove it, Bob, but I just don't think so.  

If you just telnet to a TDS host, type ~40 bytes and press ENTER, the server
will close the connection.  It knows it's not a valid login packet long before
the username arrives.  

The first 32 bits of a login packet is the length; the next 32 bits is the TDS
version.  A server receiving a wrong (to it) length or a version higher than
expected is pretty much obliged to quit.  It has no way to know where the 
username and password lie in the packet.  It's not like they're highlighted in
yellow or something.  ;-)

--jkl