Re: Avoid possible deference NULL pointer (src/backend/optimizer/path/allpaths.c)

On 05.02.2025 21:56, Tom Lane wrote:
> It's not a bug. Since the call specifies NIL pathkeys (meaning it
> doesn't care about sort order) and does not insist on a parallel-safe
> path, there should always be a path that satisfies it. The only
> way it could fail to find a path is if the rel's pathlist is entirely
> empty, a case already rejected by the sole caller.
>
> Moreover, the argument that we might not have gotten a report is not
> credible. In a production build without an Assert, the next line
> would still dump core just as effectively if the result were NULL.
>
> While the proposed patch doesn't break anything, it's removing a
> logic cross-check that was put there intentionally. So I don't
> find it to be an improvement.
>
> regards, tom lane

Ah, if this Assert was intentionally added to ensure that a path must be
always found under the given conditions, and that any issues with this
can be detected in the right place, then I agree. The patch likely makes
worse.

--
Best regards,
Ilia Evdokimov,
Tantor Labs LLC.