On Nov 13, Arthur A.Gleckler wrote:
>
> In <7.1. Script syntax> and <7.1.1. Script header>, I am appalled
> that we're considering putting such a completely Unix-specific
> feature in what is supposed to be a portable specification for the
> language. This belongs in an SRFI targeted at Unix systems, not in
> the language spec. Analogous SRFIs could address the same issue for
> code running on Windows and running on other operating systems. At
> the very least, this feature should come with a caveat saying that
> only Unix-based implementations are required to support this syntax,
> but that would still be a terrible compromise.
I think that unixisms don't belong in the document too, but there's a
very strong point for allowing them in some portable way. AFAICT, the
main problem is with different behaviors for #! that are currently
around (like #! being a comment when its on the first line, or #!...!#
being a balanced comment). Maybe a better solution for this is to
specify some #! behavior if one is used? (This, of course, will not
need to mention unix explicitly, but the rationale will still be
obvious.)
--
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:
http://www.barzilay.org/ Maze is Life!
Received on Tue Nov 14 2006 - 14:12:08 UTC