Re: [SQL/PGQ] Native executor for Graph query

From: Henson Choi <assam258(at)gmail(dot)com>
To: Henri GASC <henri(dot)gasc(at)airbus(dot)com>
Cc: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: [SQL/PGQ] Native executor for Graph query
Date: 2026-09-08 11:18:35
Message-ID: CAAAe_zAr4JsHX95RicJ5cVZbF9gosmrX0fNBgtMvMm6sgjuoVQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Henri,

Glad it was useful.

> I think I will either reimplement only what I need, or build upon
> the parent (like I did before).

I would build upon the parent. The open items behind the revert are
about catalogs and dependencies, not about execution, so the shape
your node sits on should come back much as it was. Reimplementing a
subset means answering the catalog questions that are exactly the ones
still open, and your executor patch would then have to carry them.

Reverting the revert gives you a commit, and that commit is your
development base:

git checkout -b pgq-restore master
git revert --no-commit b1f106c80
git commit -m "Restore SQL/PGQ as a cfbot base; not for review"

Then build your work on top of it, and post both commits in order:

git format-patch -o . master..<your branch>

That gives 0001 for the base and 0002 for your work, and cfbot has
something to apply.

One thing to watch: the base is a large patch touching many files, so
with your own on top the mail may be big enough to be held for
moderation.

Once Ashutosh posts the patch series again, use that as the base
instead and drop 0001.

Best regards,
Henson

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Palak Chaturvedi 2026-09-08 11:24:37 Re: Changing shared_buffers without restart
Previous Message Jakub Wartak 2026-09-08 10:58:39 Re: pg_threads.h take II