Re: Bug#310113: [VIRTUAL PACKAGE] SRFI 22 names for Scheme implementations
Bill Allombert <allomber@math.u-bordeaux.fr> writes:
>> I'm not sure I understand what you mean with "testsuite" for this?
>
> Probably Junichi was refering to a set of tests that would allow to check if
> a scheme interpretor come close to actually accept the SRFI 7 language,
> and so is a valid candidate for 'scheme-srfi-7'.
'scheme-srfi-7' is the program name for an interpreter which
accepts programs written in the language specified in SRFI 7 -
this is _not_ Scheme. Either your implementation provides a binary
which does so (Scheme 48 does, for example), or it does not. If
that binary does not interpret the language specified in SRFI 7
correctly, then it's a bug.
A simple test for that would like this:
-----8<-----------------------8<-----
#! /usr/bin/env scheme-srfi-7
(program
(code
(define (main args)
(display "Hello, World!")
(newline)
0)
))
----->8----------------------->8-----
It's difficult to test the whole language of SRFI 7 without
requiring at least some SRFI to be available as well.
If that's the kind of test Junichi was looking for, I can provide
them for the different kinds of interpreters specified by SRFI 22,
but I'm not sure this is that useful, especially since the ones
for scheme-{r4rs,r5rs,ieee-11878-1900} will be the same (the r5rs
example I posted already), and the differences are difficult to
check.
Greetings,
-- Jorgen
--
Debian GNU/Linux Developer
forcer@debian.org
http://www.forcix.cx/
Reply to: