Heikki Linnakangas <heikki@enterprisedb.com> writes:
> ISTM supporting "somefunc.ambiguous" just gives us another way to
> reference the parameter, and there still isn't any way to refer the column.
Sure. All this will do is let us remove a noted incompatibility with
Oracle, which seems worth doing if it's a one-line change that doesn't
break anything.
Further down the road, we could imagine some option in plpgsql that
prevents substitution of variables *unless* they are qualified with
the appropriate block name --- in which case we'd better make sure
there is a way to qualify function parameter names. So this might
be a necessary component of a solution that tightens up the
substitution behavior, but it's not the solution by itself.
regards, tom lane