[PATCH] switch smbcli_open over to NTCreateAndX
jamespeach at me.com
Wed Jul 28 14:11:32 MDT 2010
On Jul 26, 2010, at 9:56 AM, Volker Lendecke wrote:
> On Mon, Jul 26, 2010 at 09:33:14AM -0700, James Peach wrote:
>>> Everybody else uses ntcreate, so this might be a
>>> little regression test for the "old" code.
>> Well the end result of that argument is that smbtorture should use
>> random ancient calls in unpredictable circumstances :) I counter by
>> claiming that smbtorture should use the calls that are in use by real
>> clients, since those are the ones that matter to users.
> Well, it *does* use the modern calls in other circumstances.
> I would argue pretty stronly that we keep the old calls
> active. I had more than one day of bug-fixing of weird
> clients against NetApp filers, because those do exactly what
> you describe: They only serve the clients that are in their
> supported client matrix, they don't go an inch beyond what
> their clients demand. They don't act like Samba tries to
> act, namely try to do everything a Windows server would do.
> But that's really a matter of philosophy.
Yes, there's definitely valid arguments in both directions.
In general, I'm arguing that the default operation for smbtorture should be to use the modern protocol. That is, if the test author did not go to the trouble of deliberately specifying which operation to use, then it should use NT LM 0.12.
I'm not arguing that we should not keep the old calls active, just that they should not be the default.
> If you apply that patch, please make sure that the
> corresponding call is still excercised in another test.
OK, I will hold off until I get a chance to add an explicit SMBOpenAndX test. I'll post for review before committing.
More information about the samba-technical