Horton

Re: ExecRTCheckPerms() and many prunable partitions (checkAsUser)

Alvaro Herrera <alvherre@alvh.no-ip.org>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Amit Langote <amitlangote09@gmail.com>
Cc: Justin Pryzby <pryzby@telsasoft.com>, Ian Lawrence Barwick <barwick@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, David Rowley <dgrowleyml@gmail.com>, Greg Stark <stark@mit.edu>, Julien Rouhaud <rjuju123@gmail.com>, pgsql-hackers@lists.postgresql.org
Date: 2023-01-19T12:25:20Z
Lists: pgsql-hackers 
On 2023-Jan-19, Amit Langote wrote:

> It seems that, with the test as written, it's not the partitioned
> table referenced in the view's query that becomes a child of the UNION
> ALL parent subquery, but the subquery itself.  The bug being fixed in
> 0002 doesn't affect the planning of this query at all, because child
> subquery is planned independently of the main query involving UNION
> ALL because of it being unable to be pushed up into the latter.  We
> want the partitioned table referenced in the child subquery to become
> a child of the UNION ALL parent subquery for the bug to be relevant.
> 
> I tried rewriting the test such that the view's subquery does get
> pulled up such that the partitioned table becomes a child of the UNION
> ALL subquery.  By attaching a debugger, I do see the bug affecting the
> planning of this query, but still not in a way that changes the plan.
> I will keep trying but in the meantime I'm attaching 0001 to show the
> rewritten query and the plan.

Thanks for spending time tracking down a test case.  I'll try to have a
look later today.

> > As for 0001, it seems simpler to me to put the 'userid' variable in the
> > same scope as 'onerel', and then compute it just once and don't bother
> > with it at all afterwards, as in the attached.
> 
> That sounds better.  Attached as 0002.

Pushed this one, thank you.

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/

Commits

  1. Add a test case for a316a3bc

  2. Correctly set userid of subquery relations' child rels

  3. Fix buggy recursion in flatten_rtes_walker().

  4. Remove some dead code in selfuncs.c

  5. Update outdated comment in ApplyRetrieveRule

  6. Rework query relation permission checking

  7. Generalize ri_RootToPartitionMap to use for non-partition children

  8. Stop accessing checkAsUser via RTE in some cases

  9. Add 'missing_ok' argument to build_attrmap_by_name