[apparmor] [PATCH 2/2] libapparmor: Make swig aware of the new public libapparmor functions

John Johansen john.johansen at canonical.com
Tue Jun 16 06:36:33 UTC 2015

On 06/15/2015 08:47 AM, Tyler Hicks wrote:
> On 2015-06-13 12:00:52, John Johansen wrote:
>> On 06/13/2015 10:48 AM, Tyler Hicks wrote:
>>> Signed-off-by: Tyler Hicks <tyhicks at canonical.com>
>> Acked-by: John Johansen <john.johansen at canonical.com>
> After giving it more thought, I'd like to withdraw this patch from
> consideration for two reasons:
> 1) It breaks the Python wrapper build. Shame on me for not even build
>    testing that before sending this patch.
>    The build breaks because we have a snippet in libapparmor.i that
>    injects some code into all wrapper functions to set an exception from
>    the errno if 'result' is less than 0. That's a problem for some of
>    the void return type functions that are exported with this change
>    because their wrapper functions don't have the 'result' variable
>    declared.
> 2) More thought should go into wrapping the aa_*_new() functions and
>    aa_*_unref() functions. They should be considered as
>    constructors/destructors in the languages that we provide bindings
>    for instead of having, for example, Python code that explicitly calls
>    those functions.
> There is one valid hunk in this patch, though. It is the hunk that adds
> aa_spitcon() to the libapparmor.i file. I plan on keeping John's ack and
> reducing this patch to only include that one hunk and then pushing that
> to trunk. Let me know if that's ok, John.
yeah, that is fine

More information about the AppArmor mailing list