Re: Brainstorming ideas for define-configuration

[Joshua Branson]

Bruno Victal <mirai@makinata.eu> writes:

> Co-authored-by: Felix Lechner
>
>
> After spending some time with old and new Guix services, I'd like to
> suggest some potential improvements to our define-configuration macro:
>
>
> User-specified sanitizer support
> ===============================================================================
>

Yes please!  Thanks for working on these improvements!  It would make my
opensmtpd-service updates much easier (though I haven't really touched
the code in a while).

https://issues.guix.gnu.org/56046

>
> The sanitizers should be integrated with the type. Otherwise, they are
> tedious to use and appear verbose when repeatedly applied to multiple fields.
>
> ;; Suggestion #1
> ;; The procedure could return a sanitized value. Upon failure, there are
> ;; the following options:
> ;;    - The procedure returns only a special value (akin to %unset-value)
> ;;      and an error message, as a pair.
> ;;      Exception raising is done by define-sanitized macro behind the scenes
> ;;      which makes the procedure easier to write.
> ;;    - The procedure raises an exception. There would be no consistency
> ;;      on the message formats, however, except for any agreed convention.
> ;;      This would involve some code duplication.
> ;; Cons: too specific, not extensible.
>
> (define-sanitized typename procedure
>   (prefix ...))
>
>
> ;; Suggestion #2
> ;; A user-supplied procedure ('procname' below) would work just like the
> ;; procedure in option #1.
> ;; There is some similiarity to the Guix record-type*.
> ;; This could be extended more easily in the future should it be required.
> (define-type typename        ; maybe call this 'define-configuration-type' ?
>   (sanitizer procname)
>   (maybe-type? #t)
>   ;; The properties below are service specific.
>   ;; If this is implemented with Guix record-type* then we could have a
>   ;; module containing generic types and do something along the lines of:
>   ;; (define-type foo-ip-address
>   ;;    (inherit generic-ip-address)
>   ;;    (serializer ...))
>   (serializer procname)      ; define-type/no-serialization = sets this field 
> to #f ?
>   (prefix ...))
>
>
> Record Validator
> ===============================================================================
>
> There is also a need to validate records. Matching fields alone do not 
> actually
> ensure that the configuration is coherent and usable. For example, some fields
> may be mutually incompatible with others.
>
> We could provide procedures that validate each record type within
> define-configuration itself instead of validating the value at runtime (i.e.
> within the body of the service-type).

This is also a great idea!

> Cheers,
> Bruno