Re: [HACKERS] For review: Server instrumentation patch

Tom Lane wrote:
> I wrote:
> > I'll see about installing an initdb-time kluge to make it use OUT
> > parameters.
>
> Done:
>
> regression=# SELECT * FROM pg_stat_file('postgresql.conf');
> length | atime | mtime | ctime | isdir
> --------+------------------------+------------------------+------------------------+-------
> 12578 | 2005-08-13 14:51:03-04 | 2005-08-13 14:50:32-04 | 2005-08-13 14:50:32-04 | f
> (1 row)

Great.

> I removed the separate pg_file_length() function, as it doesn't have any
> significant notational advantage anymore; you can do

Perfect. I was going to suggest that could be removed once pg_stat_file
was more usable.

> regression=# select (pg_stat_file('postgresql.conf')).length;
> length
> --------
> 12578
> (1 row)

Great. I was also wondering if that would work. One more closed item!

> BTW, \df is no real help when it comes to stuff with OUT parameters;
> it still says
>
> regression=# \df pg_stat_file
> List of functions
> Schema | Name | Result data type | Argument data types
> ------------+--------------+------------------+---------------------
> pg_catalog | pg_stat_file | record | text
> (1 row)
>
> Possibly we should try to improve that.

Good point. Let's see if people ask for it. Because they don't need to
know the data types to use the function, we might be fine.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073