Nested quasiquotes Marc Nieper-Wißkirchen (31 Aug 2020 09:43 UTC)
Re: Nested quasiquotes Panicz Maciej Godek (31 Aug 2020 09:52 UTC)
Re: Nested quasiquotes Marc Nieper-Wißkirchen (31 Aug 2020 10:00 UTC)
Re: Nested quasiquotes Marc Nieper-Wißkirchen (01 Sep 2020 06:59 UTC)
Re: Nested quasiquotes Felix Thibault (13 Sep 2020 12:57 UTC)
Re: Nested quasiquotes Marc Nieper-Wißkirchen (13 Sep 2020 13:14 UTC)
Re: Nested quasiquotes Felix Thibault (13 Sep 2020 13:27 UTC)
Re: Nested quasiquotes Marc Nieper-Wißkirchen (13 Sep 2020 13:42 UTC)

Re: Nested quasiquotes Marc Nieper-Wißkirchen 01 Sep 2020 06:58 UTC

It's also unclear to me how you want to match (quasiquote <any>) with
any restricted syntax that is a common subset of all three matchers
(here, quasiquote is to be matched literally).

The obvious answer

`(quasiquote ,<any>)

works for the "FHD" matcher, but wouldn't work for the "WCS" matcher
when it understands the nesting level of quasiquoting.

Am Mo., 31. Aug. 2020 um 12:00 Uhr schrieb Marc Nieper-Wißkirchen
<xxxxxx@nieper-wisskirchen.de>:
>
> Panicz,
>
> you will then want to revise your pattern language (just before
> "Incompatible features should not be used") to exclude nested
> quasiquotes.
>
> PS How do you match the symbol ... with the restrictions of SRFI 200?
>
> Without restrictions, the "WCS" matcher can match it as in
>
> (match x (('... <output>)))
>
> And the "FHD" matcher as in
>
> (match x ((,id (guard (eq? '... id)) <output>)
>
> Am Mo., 31. Aug. 2020 um 11:52 Uhr schrieb Panicz Maciej Godek
> <xxxxxx@gmail.com>:
> >
> > As to SRFI-200, I have been considering that a while ago, and decided to leave it unspecified:
> > patterns containing nested quasiquotes are more confusing than they are useful, and there's usually a better way to achieve whatever is meant to be achieved by their means.
> >
> > I can make the "unspecification" explicit for clarity.
> >
> > pon., 31 sie 2020 o 11:43 Marc Nieper-Wißkirchen <xxxxxx@nieper-wisskirchen.de> napisał(a):
> >>
> >> I guess that nested quasiquote patterns are supposed to behave as the
> >> quasiquote syntax of R7RS (as in which unquotes are inserted literally
> >> and which unquotes escape the quasiquote) .
> >>
> >> However, I don't find anything about it in the specification yet.
> >>
> >> This is also relevant for the main thesis in SRFI 200:
> >>
> >> Given the pattern (quasiquote (quasiquote (unquote x))), the SRFI 204
> >> matcher will probably match the pattern (quasiquote (unquote x))
> >> literally. The SRFI 200-modified "FHD" matcher will, however, match
> >> against a more general two-element list whose first element is
> >> quasiquote and whose second element will be bound to x.
> >>
> >> I haven't checked the Racket matcher.
> >>
> >> Marc