Re: "COPY foo FROM STDOUT" and ecpg - Mailing list pgsql-hackers

From Tom Lane
Subject Re: "COPY foo FROM STDOUT" and ecpg
Date
Msg-id 2841.1361895814@sss.pgh.pa.us
Whole thread Raw
In response to "COPY foo FROM STDOUT" and ecpg  (Heikki Linnakangas <hlinnakangas@vmware.com>)
Responses Re: "COPY foo FROM STDOUT" and ecpg  (Heikki Linnakangas <hlinnakangas@vmware.com>)
List pgsql-hackers
Heikki Linnakangas <hlinnakangas@vmware.com> writes:
> While looking at Fujita Etsuro's patch to allow copy to/from a shell 
> command, I noticed that the grammar currently allows these:

> COPY foo FROM STDOUT
> COPY foo TO STDIN

> In other words, STDIN and STDOUT can be used completely interchangeably. 
> However, the ecpg grammar is more strict about that:

> ERROR: COPY TO STDIN is not possible

> Any particular reason for ecpg to check that, while the backend doesn't 
> care? I think we should just remove those checks from the ecpg grammar.

Agreed, but your draft patch doesn't do that completely.  It should only
make tests that correspond to what the error message says.  (I assume
the backend will bounce the other cases at some post-grammar stage.)
        regards, tom lane



pgsql-hackers by date:

Previous
From: Dimitri Fontaine
Date:
Subject: Re: pg_xlogdump
Next
From: Andres Freund
Date:
Subject: Re: pg_xlogdump compile error