On Sun, 20 Jun 2021, Artur T. wrote:
Am 20.06.21 um 12:40 schrieb Riccardo (Jack) Lucchetti:
> On Sun, 20 Jun 2021, Sven Schreiber wrote:
>
>> Am 20.06.2021 um 12:08 schrieb Artur T.:
>>> Am 19.06.21 um 20:40 schrieb Riccardo (Jack) Lucchetti:
>>>> On Sat, 19 Jun 2021, Allin Cottrell wrote:
>>>>
>>>>> ? set shell_ok
>>>>> shell_ok: boolean (on/off), currently on
>>>>>
>>>>> But that's not programmatic of course. You do have the --to-file
>>>>> option to "set". And I suppose we could offer a bundle of
all set
>>>>> variables as, say, $settings.
>>>>
>>>> This would be quite nice. Or we could add a key to $sysinfo.
>>>
>>> Yes, that's also useful and would avoid having another accessor.
>>
>> I don't think $sysinfo is suitable, because it collects static
>> information about the host machine and the gretl build. In contrast, the
>> settings are dynamic.
>
> Well, to be pedantic I should have to point out that we now have
> $sysinfo.mem, which is (partly) dynamic.
I think the distinction should not between static vs. dynamic information but
rather reflect the purpose of an accessor. And re-thinking Sven's argument,
it looks to me that retrieving system information ($sysinfo) is different
from getting user-settings ($settings).
I agree with Artur on this; I don't think user settings info belongs
under $sysinfo (and I think "$info" is too generic to be useful as
an identifier for anything).
BTW, the "mem" member of $sysinfo is documented, but it still needs
to be noted in ChangeLog (as do a bunch of other new things).
Allin