[RESEND PATCH] DsBind and DsGetDomainControllerInfo to return 2k8 structs complete

[Garming Sam]

Hi,

The test looks fairly reasonable, I'll probably double check it later.
But there's still a // comment left-over. We'd definitely appreciate the
extra patch to clean up the duplicated code, but we can try to get this
in first.

Also, to the end of the commit message can you add links to the bug
reports with the following lines, and your signed off by.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=9971
BUG: https://bugzilla.samba.org/show_bug.cgi?id=9976

Signed-off-by: ...


Cheers,

Garming Sam

On 22/04/16 09:23, Dirk Godau wrote:
> Hi,
>
> I've done a little cleaning, and wrote a new test.
>
> DsGetDomainControllerInfo still contains a lot of redundant code, but aside from a lot of work cleaning this up, is level 3 identical to level 2 with the only addition of is_rodc. Since it fills the level 3 answer, and I did not want to mess too much around, I would leave that for another patch. 
>
> DsBind should be in a good state. It now sets DRSUAPI_SUPPORTED_EXTENSION_LH_BETA2, which is needed for level 3 response from DsGetDomainControllerInfo.
>
> Since I also didn't want to mess with existing tests, there is now source4/torture/rpc/drsuapi_w2k8.c, testing DsBind and DsGetDomainControllerInfo for W2K8 Clients / AD Functional level. It was also the easiest way of doing this, since source4/torture/rpc/drsuapi.c worked perfectly as a template.
>
> Regards,
> Dirk Godau
>
> On 21.04.2016 09:03, Andrew Bartlett wrote:
>> On Wed, 2016-04-20 at 21:15 +0200, voidswitch wrote:
>>> Hi,
>>>
>>> Currently I've done some cleanup, and look into writing tests. Is it 
>>> preferable to write them in Python, or should it be added to 
>>> source4/torture/rpc/drsuapi.c?
>> Either is fine.  I think python is easier to write this kind of test
>> in.  
>>
>> Just a note of caution:  We spent some time today on some debugging yet
>> again some of the insane non-determinism that plagues some areas of our
>> pidl-generated python.  (A variable, not re-assigned, went from
>> representing a DN to representing "").
>>
>> That is, if you find yourself questioning if 1+1 still is 2, don't feel
>> alone, and that it is just you :-).  At that point you may indeed feel
>> like retreating to the 'safety' of C.
>>
>> Thanks,
>>
>> Andrew Bartlett 
>>