Horton

Re: Extending outfuncs support to utility statements

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@lists.postgresql.org, Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Date: 2022-07-11T00:15:25Z
Lists: pgsql-hackers 
Hi,

On 2022-07-10 19:12:52 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2022-07-09 18:20:26 -0400, Tom Lane wrote:
> >> For my taste, the circa 20K growth in outfuncs.o is an okay
> >> price for being able to inspect utility statements more easily.
> >> However, I'm less thrilled with the 30K growth in readfuncs.o,
> >> because I can't see that we'd get any direct benefit from that.
> >> So I think a realistic proposal is to enable outfuncs support
> >> but keep readfuncs disabled.
> 
> > Another approach could be to mark those paths as "cold", so they are placed
> > further away, reducing / removing potential overhead due to higher iTLB misses
> > etc. 30K of disk space isn't worth worrying about.
> 
> They're not so much "cold" as "dead", so I don't see the point
> of having them at all.  If we ever start allowing utility commands
> (besides NOTIFY) in stored rules, we'd need readfuncs support then
> ... but at least in the short run I don't see that happening.

It would allow us to test utility outfuncs as part of the
WRITE_READ_PARSE_PLAN_TREES check. Not that that's worth very much.

I guess it could be a minor help in making a few more utility commands benefit
from paralellism?

Anyway, as mentioned earlier, I'm perfectly fine not supporting readfuns for
utility statements for now.

Greetings,

Andres Freund

Commits

  1. Avoid copying undefined data in _readA_Const().

  2. Enable WRITE_READ_PARSE_PLAN_TREES of rewritten utility statements

  3. Implement WRITE_READ_PARSE_PLAN_TREES for raw parse trees

  4. Don't lose precision for float fields of Nodes.

  5. Fix write/read of empty string fields in Nodes.

  6. Add read support for some missing raw parse nodes

  7. Fix reading of BitString nodes

  8. Fix reading of most-negative integer value nodes