Re: [Gretl-users] a simple question about arima command output
On Sun, 11 Nov 2018, Sven Schreiber wrote:
Am 08.11.18 um 17:53 schrieb Allin Cottrell:
> On Thu, 8 Nov 2018, Riccardo (Jack) Lucchetti wrote:
>
>> On Thu, 8 Nov 2018, Allin Cottrell wrote:
>>
>>> Maybe we should offer atan2 as a built-in function?
>>
>> Yes, we probably should.
>
> OK, atan2() is now available and documented in git and snapshots.
Hm, welcome in principle, but I wonder if my delayspectral package still
works correctly, or if loading it would now fail because it uses a function
with the name of a built-in.
Cannot test yet myself, because I haven't updated to current git.
If there is a name clash issue there we would in theory have to check all the
function packages whenever core gretl gets a new function. (Are private
functions OK and only public ones affected? Or none at all?)
Good question. It appears that the delayspectral package is OK, but
that's probably because the included atan2 function is actually
named "atan2m" (in delayspectral 0.6, 2016-07-07).
For future reference, either we need to check packages as you say,
or ensure that we permit "masking" of built-in functions, perhaps
with a warning (as, I think, Matlab allows).
Allin