On Sat, Sep 7, 2013 at 5:25 AM, Per Bothner <[email protected]> wrote:

> I've attached FYI the diffs between the r7rs-tsest.scm versions
> in the Chibi repository and what I just checked into the Kawa repository.
> The latter now runs "cleanly" in that non-working functionality
> is either disabled or marked as "failure expected".
>
> The patches I think you might want to incorporate are the
> write and write-shared tests with datum labels.
> Kawa starts with #1, while Chibi starts with #0.
> I could change Kawa if starting with #0 is considered "normal".
> (I just have this vague feeling that #0 is similar to
> regex back-references that refer to the whole string,
> and so perhaps should be used for that conventionally.
> Not a very good reason, I admit.)
>
> There are also a few places where I changed the expected
> value from integer to floating-point, like for the
> result of (sqrt 4).


Thanks!  I've applied the inexact and datum labels changes.

I don't have any particular preference on labels starting
with #0 or #1, but most implementations (at least Chez,
Chicken, Ikarus, Racket) use #0.

-- 
Alex
_______________________________________________
Scheme-reports mailing list
[email protected]
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports

Reply via email to