编写可移植的Scheme代码的最佳方法？

发布于 2024-08-05 08:58:27 字数 804 浏览 2 评论 0原文

在 Common Lisp 中，我可以有条件地排除或包含不同实现的代码，如下所示：

#+sbcl (use-sbcl-cool-feature)
#-sbcl (use-my-own-not-so-cool-version)

这样我可以通过隔离不可移植位来编写可移植代码。

在Scheme中如何做到这一点？有没有办法像这样询问Scheme解释器或编译器的名称（甚至它实现的标准）？

(cond ((r6rs?)
       (make-eq-hashtable))
      ((gambit?)
       (make-table))
      ;; other cases here
      (#t (make-my-inefficient-hash-table))

我知道我可以将所有可能的不可移植代码片段包装在不同的过程中，然后具有如下所示的兼容性包：

;; in my-libs-gambit.scm:
(define make-hash-table make-table)

同样对于 my-libs-[other-schemes-here].scm

这是唯一的方法吗？

我不会尝试在Scheme中制作一个完全可移植的应用程序，但如果我能让我的程序在两个或三个不同的实现中运行，那就太好了。

原文

In Common Lisp I can conditionally exclude or include code for different implementations like this:

#+sbcl (use-sbcl-cool-feature)
#-sbcl (use-my-own-not-so-cool-version)

This way I can write portable code by isolating the non-portable bits.

How can this be done in Scheme? Is there a way to ask the Scheme interpreter or compiler its name (or even which standard it implements) like this?

(cond ((r6rs?)
       (make-eq-hashtable))
      ((gambit?)
       (make-table))
      ;; other cases here
      (#t (make-my-inefficient-hash-table))

I know I could wrap all possible non-portable pieces of code in different procedures and then have compatibility packages like this:

;; in my-libs-gambit.scm:
(define make-hash-table make-table)

And simlarly for my-libs-[other-schemes-here].scm

Is that the only way to do it?

I wouldn't try to make a completely portable application in Scheme, but it would be nice if I could make my progams run in two or maybe three different implementations.

分享到QQ

分享到微博