← Back to team overview

maria-developers team mailing list archive

Re: b1b38b64598: MDEV-29181 Potential corruption on FK update on a table with vcol index


Hi, Marko,

On Oct 10, Marko Mäkelä wrote:
> On Sun, Oct 9, 2022 at 2:03 PM Sergei Golubchik <serg@xxxxxxxxxxx> wrote:
> > I've spent more time on this issue, and it looks like there is no
> > structure in InnoDB with the life time till the end of the statement
> > (or, at least, nothing easily usable).
> Would ha_innobase::m_prebuilt fill that purpose? Its scope is not
> exactly a statement, because ha_innobase::reset_template() will be
> invoked not only by ha_innobase::reset() at the end of a statement. As
> I can tell, it could be close enough.

prebuilt is only minimally reset, I thought you wouldn't want anything
complex there. And anyway, Nikita's problem is with cascading updates,
they don't have their own ha_innobase::m_prebuilt.

VP of MariaDB Server Engineering
and security@xxxxxxxxxxx