Thread: info is a reserved word?

I have two identical functions below, the only difference is I  
declared my variable name to be 'info' instead of 'stuff'. I could  
not find anywhere in the docs that 'info' has any special meaning.  
Did I miss it?

create type my_info as (a text,b text
);


-- this works
create or replace function my_stuff ()
returns my_info as $$
declarestuff my_info;
beginstuff.a := 'hi';stuff.b := 'there';return stuff;
end;
$$ language plpgsql;


create or replace function my_stuff ()
returns my_info as $$
declareinfo my_info;
begininfo.a := 'hi';info.b := 'there';return info;
end;
$$ language plpgsql;

Evaluating this definition gives:

psql:16: ERROR:  syntax error at or near "info" at character 71
psql:16: LINE 4:  info my_info;
psql:16:          ^

pg 8.1.1, OS X 10.4.3



John DeSoi, Ph.D.
http://pgedit.com/
Power Tools for PostgreSQL

On Thu, Jan 12, 2006 at 08:14:42PM -0500, John DeSoi wrote:
> I have two identical functions below, the only difference is I  
> declared my variable name to be 'info' instead of 'stuff'. I could  
> not find anywhere in the docs that 'info' has any special meaning.  

'info' only seems special in PL/pgSQL, presumably because it's one
of the possible RAISE levels.  You should also get an error if you
try 'exception', 'warning', etc.

-- 
Michael Fuhr

Michael Fuhr <mike@fuhr.org> writes:
> 'info' only seems special in PL/pgSQL, presumably because it's one
> of the possible RAISE levels.  You should also get an error if you
> try 'exception', 'warning', etc.

plpgsql is not very good about reserving words "minimally", ie, not
treating a word as a keyword outside the context where the keyword
is meaningful.

This could probably be fixed, or at least greatly reduced, with some
flex/bison hacking.  Anyone up for it?
        regards, tom lane

On Thu, Jan 12, 2006 at 11:21:28PM -0500, Tom Lane wrote:
> plpgsql is not very good about reserving words "minimally", ie, not
> treating a word as a keyword outside the context where the keyword
> is meaningful.
> 
> This could probably be fixed, or at least greatly reduced, with some
> flex/bison hacking.  Anyone up for it?

Possibly.  Would it involve much more than what the main parser's
grammar does with unreserved_keyword and friends?  I suppose this
ought to move to pgsql-hackers.

-- 
Michael Fuhr