Re: [PATCH 2/2] tools/python: Fix memory leak on error path

[Julien Grall <julien@xxxxxxx>]

Hi Marek,

On 27/06/2023 13:29, Marek Marczykowski-Górecki wrote:
> On Mon, Jun 26, 2023 at 08:07:46PM +0200, Marek Marczykowski-Górecki wrote:
> > On Thu, Jun 08, 2023 at 02:59:13PM +0100, Luca Fancellu wrote:
> > > Commit 56a7aaa16bfe introduced a memory leak on the error path for a
> > > Py_BuildValue built object that on some newly introduced error path
> > > has not the correct reference count handling, fix that by decrementing
> > > the refcount in these path.
> > >
> > > Fixes: 56a7aaa16bfe ("tools: add physinfo arch_capabilities handling for Arm")
> > > Signed-off-by: Luca Fancellu <luca.fancellu@xxxxxxx>
> > > Reported-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> > Acked-by: Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>
> Oh, and BTW, in relation to the discussion on the summit about
> committing process, the buggy version was committed without my ack,
> after waiting for my review for about two weeks.
I was the committer for the series. In this case, this was not a case of 
committing because I thought the patch was uncontroversial. Instead, 
this was a lack of my side to properly check the acks on the patch. 
Sorry for that.
In general, as I mentioned during the design session, I don't commit 
with missing acks without explicitly saying so on the mailing list and 
this is often after consulting with others on IRC.
On a similar topic. So far, checking the ack is a manual process for me. 
So this is not entirely an error-free process (in particular for patch 
requiring multiple maintainers acks). I would love to have a script that 
could tell me if a patch is fully approved and/or what acks are missing.
Cheers,

--
Julien Grall