[Samba] Problem after update version 4.15.0

[Ingo Asche]

I think I figured it out myself. Don't have seen the wood for the trees...

I applied the patch to the released 4.15.0 and compiled it.

With this patched version I could login to my test machine.

Regards
Ingo

Ingo Asche via samba schrieb am 20.10.2021 um 10:46:
> By the way, is this patch to be used in the same way as described in 
> the Samba wiki?
>
> Ingo Asche via samba schrieb am 20.10.2021 um 10:36:
>> Ah, sorry, now I understand... My bad...
>>
>> I try the patch and come back to you...
>>
>> Andrew Bartlett via samba schrieb am 20.10.2021 um 10:33:
>>>
>>> On Wed, 2021-10-20 at 10:24 +0200, Ingo Asche wrote:
>>>> Hi Andrew,
>>>>
>>>> then the original revert is not necessary anymore?
>>> I'm not sure - if things just work, then all you have found is that
>>> things break mid-stream with GD's patches (not unusual), but the real
>>> bug was later.  This happens.
>>>
>>>> And I presume Günther is of the hook? ;-)
>>> Not sure, it is possible there is more than one bug.  Try the patch and
>>> see.
>>>
>>>> If I should test a new git revert what has to be the correct command?
>>> I attached the patch to apply, we developed a fix in master after we
>>> noticed a suspicious behaviour in further Kerberos testing.
>>>
>>> git cherry-pick -x 8e1efd8bd3bf698dc0b6ed2081919f49b1412b53 will get it
>>> otherwise (if the list ate the attachement).
>>>
>>> Andrew Bartlett
>>>
>>>> Regards
>>>> Ingo
>>>>
>>>> Andrew Bartlett schrieb am 20.10.2021 um 10:18:
>>>>> On Wed, 2021-10-20 at 10:11 +0200, Ingo Asche wrote:
>>>>>> Message after last commit tested:
>>>>>> b3ee034b4d457607ef25a5b01da64e1eaf5906dd is the first bad commit
>>>>>> commit b3ee034b4d457607ef25a5b01da64e1eaf5906dd
>>>>>> Author: Stefan Metzmacher <metze at samba.org>
>>>>>> Date:   Fri Apr 10 23:10:28 2020 +0200
>>>>>>
>>>>>>       s4:kdc: prefer newer enctypes for preauth responses
>>>>>>
>>>>>>       This matches Windows KDCs, which was demonstrated by the
>>>>>>       krb5.as_req_tests tests.
>>>>>>
>>>>>>       Signed-off-by: Stefan Metzmacher <metze at samba.org>
>>>>>>       Reviewed-by: Andrew Bartlett <abartlet at samba.org>
>>>>>>
>>>>>>       Autobuild-User(master): Stefan Metzmacher <metze at samba.org>
>>>>>>       Autobuild-Date(master): Thu Jul  1 18:37:14 UTC 2021 on sn-
>>>>>> devel-
>>>>>> 184
>>>>>>
>>>>>> :040000 040000 85b894b1c0893bd9b280bf09f405704ef9263061
>>>>>> 82f81bb5ff6a8428dd90d29c088970017850f4ab M      selftest
>>>>>> :040000 040000 4f166f3da6fb271df4965ac05666f67728492a6c
>>>>>> f862aeebe5e0f6df34529701237b33fe86735ab8 M      source4
>>>>> Awesome.  We have a winner!
>>>>>
>>>>> This is what I was expecting the issue to be, see
>>>>> https://bugzilla.samba.org/show_bug.cgi?id=14864
>>>>>
>>>>> The commit to cherry-pick from master
>>>>> is 8e1efd8bd3bf698dc0b6ed2081919f49b1412b53 and is attached.
>>>>>
>>>>> Sorry for not mentioning this earlier, I've been a bit swamped and
>>>>> the
>>>>> earlier bisect seemed not to support this.
>>>>>
>>>>> Andrew Bartlett
>>>>>
>>
>>
>
>