[jcifs] Unmapped NetServerEnum2Response code

[Del Merritt]

On 03/11/2010 05:36 PM, Michael B Allen wrote:
> On Thu, Mar 11, 2010 at 4:40 PM, Del Merritt <del at alum.mit.edu> wrote:
>   
>> I'm using JCIFS 1.3.14*.  I have code that does this:
>>
>> public class NetServerEnum2Ext extends SmbFile
>> {
>>   myMethod() {
>>     int Flags = 0x00000200;
>>     NetServerEnum2 req = new NetServerEnum2(host, Flags);
>>     NetServerEnum2Response resp = new NetServerEnum2Response();
>>     send(req, resp);
>>     // Check status
>>     if(resp.status != SmbException.ERROR_SUCCESS && resp.status !=
>> SmbException.ERROR_MORE_DATA)
>>                     throw new SmbException(resp.status, true);
>>   ...
>>   }
>> }
>>
>> Usually it works fine.  Sometimes it will throw because it doesn't recognize
>> the response status.  The value I see in the response status is "2".  The
>> same host, flags, and authentication info (not shown for brevity) are used
>> both when it works and when it doesn't.
>>
>> jcifs.smb.SmbException is not figuring out a text translation.  Where should
>> I look to see what the status means?
>>     
> Hi Del,
>
> I don't even know where to begin to decipher what you're doing here.
> The "2" could be a status code or it could be garbage.
>
> Note that the Net* calls are RAP calls which are totally obsolete and
> will not work in most environments. 

Would you define "obsolete"?  I am not trying to be obtuse; instead, I
am trying to put together an application that will play well in a
sandbox that includes XP, Vista, and 7 systems, along with a couple of
Linux systems running samba and CUPS.  At least one - and arguably two -
of those OSs are considered obsolete.  For the moment I am willing to
constrain the discovery mechanism to be running in a "typical"
domain-based environment.

> In fact, that is probably the
> problem you're having. I think this particular RAP call would only
> work with a NetBIOS "master browser". There is a special lookup for
> that (see the NbtAddress class) but again, not all environments will
> have a master browser and if they do it's not going to have a good
> consistent list of servers. 

I'm still puzzling as to why it "works" most of the time.  What I
haven't done is log the host that is responding to my request(s);
perhaps multiple master browsers are responding, and one is not
responding well.  Hmm. 

Thanks for the tip to use the NbtAddress class; I may be back with some
more questions on that.

> Getting a proper list of servers on the
> subnet is a rather difficult think to do actually.
>   

Understood.  I'm happy to try to do it the Right Way.  I also want it to
work in legacy networks.

Thanks,
-Del