samba-4.3.6 and talloc-2.1.6

Stefan Metzmacher metze at
Mon Mar 14 02:05:00 UTC 2016

Hi Andrew,

>> And have a method similar to talloc.enable_null_tracking(),
>> maybe called talloc.pytalloc_enable_strict_type_checking()
>> that will prevent the fallback.
> I really don't want this wart on the API forever.  It is bad enough
> what we have with talloc.Object being left around, I would have
> preferred not to have to do that. 
> As far as I'm aware, the only use case for this is that call in dbcheck
> (here called by provision).  While undesirable, the failure message is
> clear (to us, and will quickly find this thread in google), and is not
> an abort(), which is what was happening in the same area for some
> versions previously.  
>> Then we should also include the type name (including info about
>> tp_base
>> and tp_basicsize)
>> in the raised exception, the following doesn't give enough
>> information:
>> pytalloc_reference_ex() called for object type not based on talloc
> I'm happy for you to extend the exception, but I still strongly prefer
> that we patch the older Samba versions and move on.  Leaving an
> unchecked cast in our ABI forever just doesn't sit well with me at all.
> Once we patch even just 4.4 and 4.3, the number of users running the
> most recent talloc with an older Samba, as an AD DC, will be
> vanishingly small.

Can you please prepare the backports for 4.4 and 4.3,
we need the 4.4 patches by tomorrow in order to get them in 4.4.0.

and its referenced bugs:


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <>

More information about the samba-technical mailing list