Re: [Scheme-reports] "include" filename resolution Eli Barzilay (12 Aug 2011 20:38 UTC)
Re: [Scheme-reports] "include" filename resolution Andy Wingo (12 Aug 2011 20:53 UTC)
Re: [Scheme-reports] "include" filename resolution Eli Barzilay (12 Aug 2011 20:59 UTC)
Re: "include" filename resolution Arthur A. Gleckler (12 Aug 2011 21:41 UTC)
Re: [Scheme-reports] "include" filename resolution Eli Barzilay (12 Aug 2011 21:42 UTC)
Re: "include" filename resolution Arthur A. Gleckler (12 Aug 2011 21:44 UTC)
Re: [Scheme-reports] "include" filename resolution Andy Wingo (12 Aug 2011 21:43 UTC)
Re: [Scheme-reports] "include" filename resolution Ray Dillinger (13 Aug 2011 15:29 UTC)

Re: [Scheme-reports] "include" filename resolution Andy Wingo 12 Aug 2011 21:43 UTC

On Fri 12 Aug 2011 22:58, Eli Barzilay <eli@barzilay.org> writes:

> you'd expect `load' to be a
> completely run-time thing (open+read+eval), so the CWD makes perfect
> sense.  This way you have a single behavior rather than something that
> changes based on whether it's used in some file or on the REPL.
>
> And given that the latter will be popular for a while longer, it's
> especially surprising that it's different from inside a module.  Given
> that, it might make more sense to have it use the CWD when used on the
> REPL, and throw a syntax error from inside a module.

The problem is that in the Guile 1.8, read and expansion and evaluation
all happened at the same time, so our users were not accustomed to the
subtleties.  This terrible thing that Guile 2.0 does is the best that we
can do to support old code, it seems.

I respect racket's "just make a new language" model, but Guile is trying
to incrementally migrate old code over to new idioms.  We have to make
it easy and mostly compatible though, or users will revolt.

Andy
--
http://wingolog.org/

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports