[Samba] Problem after update version 4.15.0

[Ingo Asche]

This are the messages from before the error:

[2918/4430] Linking 
bin/default/python/python-glue.cpython-37m-arm-linux-gnueabi hf.so
../../lib/audit_logging/tests/audit_logging_error_test.c: In function 
‘__wrap_ta lloc_named_const’:
../../lib/audit_logging/tests/audit_logging_error_test.c:243:14: error: 
cast to  pointer from integer of different size 
[-Werror=int-to-pointer-cast]
   void *ret = (void *)mock();
               ^
../../lib/audit_logging/tests/audit_logging_error_test.c: In function 
‘__wrap_ta lloc_strdup’:
../../lib/audit_logging/tests/audit_logging_error_test.c:259:14: error: 
cast to  pointer from integer of different size 
[-Werror=int-to-pointer-cast]
   void *ret = (void *)mock();
               ^
In file included from 
../../lib/audit_logging/tests/audit_logging_error_test.c:4 6:
../../lib/audit_logging/tests/audit_logging_error_test.c: In function 
‘test_json _add_address’:
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:629:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_named_const, REAL_TALLOC);
   ^~~~~~~~~~~
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:650:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_named_const, REAL_TALLOC);
   ^~~~~~~~~~~
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:665:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_named_const, NULL);
   ^~~~~~~~~~~
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:677:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_named_const, REAL_TALLOC);
   ^~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c: In function 
‘test_json _to_string’:
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:754:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_named_const, REAL_TALLOC);
   ^~~~~~~~~~~
../../third_party/cmocka/cmocka.h:111:6: error: cast from pointer to 
integer of  different size [-Werror=pointer-to-int-cast]
      ((LargestIntegralType)(value))
       ^
../../third_party/cmocka/cmocka.h:313:18: note: in expansion of macro 
‘cast_to_l argest_integral_type’
                   cast_to_largest_integral_type(value), 1)
                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../lib/audit_logging/tests/audit_logging_error_test.c:771:2: note: in 
expansi on of macro ‘will_return’
   will_return(__wrap_talloc_strdup, NULL);
   ^~~~~~~~~~~
cc1: all warnings being treated as errors



Andrew Bartlett via samba schrieb am 10.10.2021 um 23:40:
> Yes, you start from 4.14.0rc1 as this is the branch point where master
> split into 4.14 (working) and what would eventually be 4.15.
>
> If you do it on a new test DC then the paths will all line up with the
> source based install, otherwise an existing test DC is fine as long as
> you fix Samba up to find the DBs etc.  A new DC may be simpler.
>
> Thanks,
>
> Andrew
>
> On Sun, 2021-10-10 at 22:30 +0200, Ingo Asche via samba wrote:
>> Hi Andrew,
>>
>> I will try that tomorrow. I hope I can get you the needed
>> information,
>> never done something like that before.
>>
>> Just one question: Samba 4.14.0rc1 is really correct? Not 4.15.0rc1?
>>
>> And, that may be a dumb question, should I do that on the existing
>> test
>> DC or should I install it new?
>>
>> Regards
>> Ingo
>>
>> Andrew Bartlett via samba schrieb am 10.10.2021 um 21:35:
>>> On Sat, 2021-10-09 at 16:37 +0200, Ingo Asche via samba wrote:
>>>> To be on the sure side: I joined the test workstation to my
>>>> 4.14.8
>>>>
>>>> domain and tried to login. This worked without problem. So it
>>>> seems
>>>> on
>>>>
>>>> the first sight that the Windows installation is not the problem.
>>>> By
>>>> the
>>>>
>>>> way it's an actual Windows 10 21H1.
>>>>
>>>>
>>>>
>>>> Rejoining to the test domain worked but again no longer login is
>>>> possible.
>>> So, what we really need is for a git bisect in a git
>>> checkout.  Build
>>> Samba from source on Samba 4.14.0rc1 with:
>>>
>>> git clone https://git.samba.org/samba.git
>>> cd samba
>>> ./configure.developer && make -j && make install
>>> (join to your domain or point at your existing DB)
>>>
>>> Confirm this fails (if it works then we will instead bisect back to
>>> find where it started working, but for now I assume it fails), then
>>> run:
>>>
>>> git checkout samba-4.14.0rc1
>>>
>>> ./configure.developer && make -j && make install
>>> (join to your domain or
>>> point at your existing DB)
>>>
>>> Confirm this works (if it fails we have bigger problems), then run:
>>>
>>> git bisect good HEAD
>>> git bisect bad origin/master
>>>
>>> The run
>>>
>>> ./configure.developer && make -j && make install
>>> for each step, answering with
>>>
>>> git bisect [good|bad]
>>>
>>> for each decision point.
>>>
>>> Use 'git bisect skip' if it all blows up, but do 'rm -rf bin' first
>>> and
>>> perhaps blow away the binaries in your install.  You should also,
>>> after
>>> the first join, be able to run 'bin/samba' rather than the slower
>>> 'make
>>> install' step.
>>>
>>> You don't need to re-join each time as the DB is upgrade/downgrade
>>> compatible on these versions.
>>>
>>> Looking into this for us would save us a lot of time, we are quite
>>> busy
>>> on some other things and need your help.
>>>
>>> Andrew Bartlett
>>>
>>>> Regards
>>>>
>>>> Ingo
>>