On fre, 2012-03-16 at 13:47 -0400, Tom Lane wrote:
> I'm a bit concerned about whether that's actually going to be useful.
> A quick check shows that in the regression database, the proposed patch
> produces 3246 possible completions, which suggests that by the time you
> get down to a unique match you're going to have typed most of the name
> anyway.
Well, the regression test database is not really an example of real-life
object naming, I think. I tried this out on a couple of real databases
and found it quite handy.
>
> BTW, you should at least exclude dropped columns, I think.
>
Yes.
