New python PIDL checks cause 2221 new Coverity warnings

Kai Blin kai at
Thu Aug 27 08:19:58 UTC 2015

On 2015-08-27 04:44, Andrew Bartlett wrote:

Hi Andrew,

>> I see no point in crying over spilled milk. So now that the plan has
>> collided with reality, how do we proceed to make our Coverity useful
>> again?
> I'm really not sure.  This really wasn't the quick-and-dirty solution,
> it was the best we could come up with (I worked with Douglas on it, as
> he had started on some 32 bit patches in this area).  I originally
> tried generating different code for different cases, and we ended up
> with a tangled mess of duplicated PIDL generators, and still much the
> same problem due to the 32 bit python integer limit.

I've never claimed it was the quick-and-dirty solution, but right now it 
is causing considerable noise in the static code analysis tools, 
rendering them pretty useless. I'm not sure this is a price we're just 
willing to pay and move on.

> Productive (ie, please look at the generators, this isn't a simple
> area) suggestions, and patches are most welcome, and gladly this area
> is now tested so we have a hope of validating the result.

Realistically, I don't have time to do that, and likely won't have any 
time soon. I mainly spoke up because "fix a coverity issue" was a nice 
thing I could do when I had a bit of time and felt like hacking on 
Samba, and now it looks like I'll have to data mine for a true issue first.


Kai Blin
Samba Developer

More information about the samba-technical mailing list