Re: Mutating existing persistent data

[Joost Kremers <joostkremers@fastmail.fm>]

On Tue, Dec 07 2021, Joost Kremers wrote:
> On Tue, Dec 07 2021, Lars Ingebrigtsen wrote:
>> Uhm...  anybody else got any ideas for a word for a variable that has a
>> value that survives an Emacs shutdown?
>
> persist-defvar?
>
> Oh, wait, that already exists... 😲 (Well, OK, it's on GNU ELPA, but still.)
>
> I noticed you hadn't listed "permanent" yet, or "durable".
>
> But I do think 'persistent' is the usual term to describe this sort of thing, so
> personally I'd find a way to use it.

Actually, what I wanted to add: If this is to be a persistent storage mechanism
for Emacs, I do think there should be an automatic fall-back option for people
that can't or don't want to use sqlite3, e.g., because they prefer their
persistent storage to be in plain text. Even more so if this persistent storage
is going to be used by 2nd and 3rd party packages (which I assume is the
intention). You don't want to leave unsuspecting users hanging when they install
some package that doesn't work as advertised because it uses a persistent storage
mechanism that doesn't work on the user's installation.

OTOH if no such fall-back is planned, it would IMHO be better to reflect that in
the name of the library and its functions. So preferably `sqlite-defvar` rather
than `persistent-defvar` or something. That would stand a better chance of
alerting package authors to the fact that they cannot count on this
functionality being available on every Emacs install.

Just my €0.02, of course.


-- 
Joost Kremers
Life has its moments