Re: [Gretl-devel] weekdays (was Re: [Gretl-users] problems with daily data)
On Sat, 24 May 2014, Riccardo (Jack) Lucchetti wrote:
On Fri, 23 May 2014, Allin Cottrell wrote:
> 2) I've added something Jack suggested: a --permanent option to
"smpl",
> which substitutes the restricted dataset for the original. This will need
> some testing. Right now it's not available in the context of just moving
> the sample end-points: it's valid only in association with the options
> --restrict, --dummy, --no-missing or --random.
I very welcome the change, but it had a very nasty side effect. The options
--restrict and --dummy are now marked as incompatible, whereas they happily
worked together before. This change broke some of my scripts, which I'm now
working to update. I don't know what the best thing to do is: this isn't,
strictly speaking, a backwards-incompatible change, but the effect it has is
the same.
Ah, marking --restrict and --dummy as incompatible didn't really
have anything inherently to do with adding the --permanent option,
it's just that while I was at it I noticed that we didn't have a
check in place for incompatible options so I added one.
But if using --dummy and --restrict together hasn't caused any
problems to date then we should probably relax the new check in that
regard. (I guess the --restrict option simply gets ignored if
--dummy is given.)
Allin