Content-Type: text/plain; charset=UTF-8
Date: Mon, 22 Jun 2026 07:18:46 -0300
Message-Id: <DJFI2DVOCP4O.1ECNPFUE1B3XC@gmail.com>
Cc: <adoros@starfishstorage.com>, <pgsql-bugs@lists.postgresql.org>
Subject: Re: BUG #19480: PL/Python SRF crashes (SIGSEGV) when function is
 replaced mid-iteration: use-after-free in PLy_funct
From: "Matheus Alcantara" <matheusssilv97@gmail.com>
To: "Tom Lane" <tgl@sss.pgh.pa.us>
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
References: <19480-f1f9fdce30462fc4@postgresql.org>
 <DIS405H05GE0.1OPMN9BEKKMS1@gmail.com> <982975.1779981146@sss.pgh.pa.us>
 <DIY2500FL0UW.Z4M7NYWMGGA4@gmail.com> <2868592.1780356411@sss.pgh.pa.us>
 <DJ1BFHKUC06G.1NO3D0WC9KBV6@gmail.com> <1149447.1781733399@sss.pgh.pa.us>
 <DJC60EHI0Y34.1ZTU7FGUMW6DQ@gmail.com> <381335.1782070804@sss.pgh.pa.us>
In-Reply-To: <381335.1782070804@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/DJFI2DVOCP4O.1ECNPFUE1B3XC%40gmail.com>
Precedence: bulk

On Sun Jun 21, 2026 at 4:40 PM -03, Tom Lane wrote:
> "Matheus Alcantara" <matheusssilv97@gmail.com> writes:
>> Thanks for the review! I've tried to address all your points in the
>> attached v2.
>
> Pushed after a round of review.  I made some mostly-cosmetic changes,
> such as rewriting comments (consolidating some stuff I thought was
> duplicative).  The main thing I fixed that was an actual bug was
> you were careless about lifespan of variables around PG_TRY blocks.
> The rule of thumb is that if a variable is modified inside PG_TRY
> and then used after that block (including in the PG_CATCH) then it
> has to be marked volatile.  Where possible, I avoid using the
> volatile marking by assigning the variable's value before PG_TRY.
>

Noted, thanks for the call.

>> I've also added a regression test, not sure if there is a better way to
>> exercise this fix but this test crash without this patch applied.
>
> Kind of a hokey test, since it doesn't model the likely actual case
> where the CREATE happens in another session, but this is as close as
> we'll get without a much more complex test setup.  I kept it, and
> also added another test that exercises the early-termination path,
> since code coverage showed me that ShutdownPLyFunction() wasn't being
> reached.
>

Thank you for reviewing and committing the patch!

--
Matheus Alcantara
EDB: https://www.enterprisedb.com