Re: define anywhere

[Damien Mattei <damien.mattei@gmail.com>]

On Sun, Jun 6, 2021 at 3:36 PM Linus Björnstam <linus.internet@fastmail.se>
wrote:

> Oh, my version is an error. I implemented the macro before definitions in
> (some) expression contexts were a thing.
>

do you talk about what? the linked code?

>
> Andy did a marvellous thing implementinf the new letrec: if it is
> possible, letrec will have no overhead. It also automatically handle
> dependant clauses: even though it might seem counterintuitive, the
> guile-proper version will do the right thing always. The syntactic
> differentiation mine does is just there for our feeble human minds, at
> least with regards to what guile-proper is doing.
>
> So TL/DR: guiles way is the correct way.
>
> --
>   Linus Björnstam
>
> On Sun, 6 Jun 2021, at 14:02, Dr. Arne Babenhauserheide wrote:
> >
> > Linus Björnstam <linus.internet@fastmail.se> writes:
> >
> > > becomes ONE letrec under gulie3, whereas my library turns it into
> > > (letrec ((a 2) (b 3))
> > >   (display "hej")
> > >   (letrec ((c 3))
> > >     (+ a b c)))
> > >
> > > That should be an easy fix, again if there is any interest.
> >
> > I’m not sure which approach I prefer. Your approach is more precise, but
> > I slightly lean towards the Guile3-version, because it does not change
> > behaviour when I put a pretty-print between two defines.
> >
> > Though I would want a precise approach inside other forms (like when:
> > separating inside when and outside when).
> >
> > Best wishes,
> > Arne
> > --
> > Unpolitisch sein
> > heißt politisch sein
> > ohne es zu merken
> >
> > Attachments:
> > * signature.asc
>