From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>
Cc: Tomas Vondra <tomas.vondra@2ndquadrant.com>
Cc: pgsql-hackers@postgresql.org
Subject: Re: WIP: Faster Expression Processing v4
Date: Wed, 22 Mar 2017 07:52:40 -0700
Message-ID: <20170322145240.vt2iolwt6vpmx33z@alap3.anarazel.de> (raw)
In-Reply-To: <31150.1490193666@sss.pgh.pa.us>
References: <20170307030137.lpo4japv5oqgbqpz@alap3.anarazel.de>
<2090fd00-4dec-7b42-f700-64308958504c@2ndquadrant.com>
<20170313080351.4g5oubxcj4xj5y7t@alap3.anarazel.de>
<fcdaa51b-e905-72e0-39dd-5e22f535a253@2ndquadrant.com>
<20170314065259.ffef4tfhgbsaieoe@alap3.anarazel.de>
<69b7bb0e-8dab-3d39-1a89-3c6ab6409f64@iki.fi>
<20170314224706.nxvtapenky6eom3z@alap3.anarazel.de>
<13062.1489622943@sss.pgh.pa.us>
<20170320062511.hp5qeurtxrwsvfxr@alap3.anarazel.de>
<31150.1490193666@sss.pgh.pa.us>
List-Unsubscribe: <mailto:majordomo@postgresql.org?body=unsub%20pgsql-hackers>
On 2017-03-22 10:41:06 -0400, Tom Lane wrote:
> I've been busily hacking away on this, trying to make things cleaner and
> fix a couple of bugs I stumbled across. (Confusion between ExecQual and
> ExecCheck, for instance - we apparently lack regression tests exercising
> constraints-returning-null in corner cases such as table rewrite.) It
> will probably be a day or two more before I'm done.
Thanks!
> A couple of naming questions:
>
> * I concur with Heikki's dislike for the file name execInterpExpr.c.
> Would it make it any better to switch to execExprInterp.c? I think
> that having all this stuff under a common pattern execExpr*.c would
> be a good thing (and I notice you've already got one comment referring
> to them that way ...)
That works for me.
> * execQual.c doesn't seem to have a unifying reason to exist anymore.
> It certainly has little to do with evaluating typical qual expressions;
> what's left in there seems to be mostly concerned with SRFs. I feel
> like it might be a good idea to rename it, but to what? execExprUtils.c
> perhaps? Or maybe we should destroy it altogether, shoving the SRF
> stuff into nodeFunctionscan.c and moving what little remains into
> execUtils.c.
Yea, I was wondering about that too. What would we do with
GetAttributeByName/Num?
> * I do not like the function name ExecInstantiateExpr(). Webster's
> defines "instantiate" as "to represent (an abstraction) by a concrete
> instance", which does not seem to me to have a lot to do with what this
> function actually does.
It perhaps makes a *bit* more sense if you view it from the POV that,
with the future JIT support (WIP versions of which I posted previously),
it'd actually create a compiled function which'd largely be independent
of the of ->steps (except for the non-hotpath functions, which'd still
end up using it). So one of the above "conrete instances" would be the
interpeted version, another the compiled one. No, not an entirely
convincing argument.
> There's nothing very abstract about its input.
> More, the implication of "instantiate" is that you can instantiate any
> number of representatives of the same abstraction, but this scribbles
> on the input in a one-way fashion. I think perhaps something like
> ExecPrepareExpr or ExecFinishExpr or something along that line would
> be better, but nothing is really standing out as le mot juste.
Either of those work, but they don't strike me as perfect either, but I
can't come up with something better (ExecReadyExprForExec()?).
Greetings,
Andres Freund
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-hackers@postgresql.org
Cc: andres@anarazel.de, tgl@sss.pgh.pa.us, hlinnaka@iki.fi, tomas.vondra@2ndquadrant.com
Subject: Re: WIP: Faster Expression Processing v4
In-Reply-To: <20170322145240.vt2iolwt6vpmx33z@alap3.anarazel.de>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox