Horton

RE: ecpg command does not warn COPY ... FROM STDIN;

Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com>

View thread View on postgr.es
From: "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
To: 'Tom Lane' <tgl@sss.pgh.pa.us>, Fujii Masao <masao.fujii@oss.nttdata.com>
Cc: Ryo Kanbayashi <kanbayashi.dev@gmail.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Date: 2025-01-09T07:52:57Z
Lists: pgsql-hackers 
Dear Tom, Fujii-san,

> > ISTM that ecpg supports COPY TO STDOUT and includes the regression test
> "copystdout" for it. No?
> 
> Oh right.  (Pokes at it...)  It looks like the backend accepts
> "FROM STDOUT" as a synonym for "FROM STDIN", so that's why this
> is checking for both spellings.  But I agree it'd be better
> for the error message to only use the standard spelling.
> 
> Also I tried
> 
> regression=# copy int4_tbl from program stdin;
> ERROR:  STDIN/STDOUT not allowed with PROGRAM
> 
> So we probably don't need to bother with adjusting the check
> to allow that.

To confirm - I have no objections for the decision. I'm also think it is not
an usual grammar. I checked the doc [1] and I could not find "COPY FROM STDOUT".
I.e., first version is enough.

[1]: https://www.postgresql.org/docs/devel/sql-copy.html

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Commits

  1. ecpg: Restore detection of unsupported COPY FROM STDIN.

  2. plpgsql: pure parser and reentrant scanner

  3. Add support for piping COPY to/from an external program.