COMMENT ON

Started by Tatsuo Ishiiover 24 years ago4 messages
#1Tatsuo Ishii
t-ishii@sra.co.jp

In comment.sgml:

COMMENT ON AGGREGATE my_aggregate (double precision) IS 'Computes
sample variance';

this raises error. However,

COMMENT ON AGGREGATE my_aggregate double precision IS 'Computes
sample variance';

works but looks strange syntax. Should we fix the program or docs?
--
Tatsuo Ishii

#2Peter Eisentraut
peter_e@gmx.net
In reply to: Tatsuo Ishii (#1)
Re: COMMENT ON

Tatsuo Ishii writes:

In comment.sgml:

COMMENT ON AGGREGATE my_aggregate (double precision) IS 'Computes
sample variance';

this raises error. However,

COMMENT ON AGGREGATE my_aggregate double precision IS 'Computes
sample variance';

works but looks strange syntax. Should we fix the program or docs?

I vote for fixing the program.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

#3Tom Lane
tgl@sss.pgh.pa.us
In reply to: Peter Eisentraut (#2)
Re: COMMENT ON

Peter Eisentraut <peter_e@gmx.net> writes:

COMMENT ON AGGREGATE my_aggregate double precision IS 'Computes
sample variance';

works but looks strange syntax. Should we fix the program or docs?

I vote for fixing the program.

If we fix this, we should also change DROP AGGREGATE, which also uses
the paren-less syntax. (I think the COMMENT ON syntax was modeled on
DROP.)

I'd be in favor of changing, but we do need to maintain consistency.

Another issue is that pg_dump knows about using both of these
commands... we'll have a compatibility problem if we don't continue
to accept the old syntax for awhile.

regards, tom lane

#4Tom Lane
tgl@sss.pgh.pa.us
In reply to: Tom Lane (#3)
Re: COMMENT ON

Peter Eisentraut <peter_e@gmx.net> writes:
COMMENT ON AGGREGATE my_aggregate double precision IS 'Computes
sample variance';

works but looks strange syntax. Should we fix the program or docs?

I vote for fixing the program.

If we fix this, we should also change DROP AGGREGATE, which also uses
the paren-less syntax. (I think the COMMENT ON syntax was modeled on
DROP.)

I'd be in favor of changing, but we do need to maintain consistency.

Another issue is that pg_dump knows about using both of these
commands... we'll have a compatibility problem if we don't continue
to accept the old syntax for awhile.

All fixed ...

regards, tom lane