On Wed, 22 Apr 2009, Allin Cottrell wrote:
On Wed, 22 Apr 2009, Marcin B�^Bażejowski wrote:
> suppose we have two functions: Foo and foo.
> We pass a named list to function Foo, which body contains line with call
> of function foo and foo takes as an argument series not list, for which
> computes simple ols (example code below). Where is the problem? Function
> foo doesn't even run ols, so no accesors are available and we get error.
The second function was not reporting the OLS results because it
was bracketed by the fact that estimation occurred within a loop,
and models are treated specially within loops.
However, this was not correct in context since the loop in
question was at a different level of function execution from the
ols command. I think this is now handled correctly in CVS; your
script now runs OK here. But I'm not certain that I have this
totally right, so please post again if you find anomalies in this
sort of context.
Sorry, it seems I managed to break my patch before I applied it
to CVS last night. I believe this should be fixed correctly
now (and in the Windows snapshot too).
Allin.