Re: [HACKERS] attlen weirdness?

>
> Quoting The Hermit Hacker (scrappy@hub.org):
> >
> > Bruce?  Guys?  Any word on this one?  I *thought* I remembered reading
> > soething about the 'attlen' becoming -1, but just checked the mailing list
> > archives and couldn't find anything :(
> >
>
>     It appears as if atttypmod now has the field length + 4 ???

I assume you are asking of attlen for char() and varchar() have the max
length.  The answer is now NO.  They are varlena structures like text,
and so have a attlen of -1.  Atttypmod for that pg_attribute row now has
the defined length.

In the old days, pre 6.3, pg_type.typlen != pg_attribute.attlen for
char() and varchar(), causing all sorts of type-conditional wierness in
the backend for these types.  With the new code, pg_type.typlen =
pg_attribute.attlen all the time.

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)