[PATCH] fix segfault in pyldb

[Noel Power]

On 21/04/2019 23:34, Douglas Bagnall via samba-technical wrote:
> hi Lukas,
>
>>> diff --git a/python/samba/tests/segfault.py b/python/samba/tests/segfault.py
>>> index 38dcb5c7036..b95c1ef5fa9 100644
>>> --- a/python/samba/tests/segfault.py
>>> +++ b/python/samba/tests/segfault.py
>> It seems to be pure ldb test.
> That is true.
>
>> Is there any reason wh it was not added into ldb tests ?
>> lib/ldb/tests/python
> A test has to do special tricks to turn a segfault into a failure rather
> than a disappearance, which is needed to be able to label the test as a
> known failure. python/samba/tests/segfault.py does those tricks, while I
> know of nothing on the ldb side that does. It would be easy to copy
> across, but then I also don't know if the ldb tests have a mechanism for
> known failures.
>
> There are other ldb tests in main samba tests -- for example the ones
> called samba4.ldb.simple.* that run testprogs/blackbox/test_ldb_simple.sh.
> Perhaps they are left there so they can be listed as knownfails.
>
> I think it would be reasonable in this case to omit the test altogether.
> We may introduce practically unnoticeable segfaults into pyldb in the
> future, but it would be very good luck to reintroduce exactly the one we
> test for.
>
> So I think ldb should be happy. It doesn't need to test, but a friendly
> external project --samba-- is testing on its behalf anyway.
>
>> Otherwise nice catch.
> thanks!
>
> Any team members want to review?

RB+

pushed to autobuild

Noel