Re: Decoding speculative insert with toast leaks memory

On Thu, Jun 10, 2021 at 2:12 PM Dilip Kumar <dilipbalaut@gmail.com> wrote:
>
> On Wed, Jun 9, 2021 at 8:59 PM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> >
> > May I suggest to use a different name in the blurt_and_lock_123()
> > function, so that it doesn't conflict with the one in
> > insert-conflict-specconflict?  Thanks
>
> Renamed to blurt_and_lock(), is that fine?
>

I think a non-conflicting name should be fine.

> I haved fixed other comments and also prepared patches for the back branches.
>

Okay, I have verified the fix on all branches and the newly added test
was giving error without patch and passes with code change patch. Few
minor things:
1. You forgot to make the change in ReorderBufferChangeSize for v13 patch.
2. I have made a few changes in the HEAD patch, (a) There was an
unnecessary cleanup of spec insert at one place. I have replaced that
with Assert. (b) I have added and edited few comments both in the code
and test patch.

Please find the patch for HEAD attached. Can you please prepare the
patch for back-branches by doing all the changes I have done in the
patch for HEAD?

-- 
With Regards,
Amit Kapila.