This is the mail archive of the
kawa@sourceware.org
mailing list for the Kawa project.
Re: Confusing error message
* Per Bothner [2011-04-17 15:46] writes:
> On 04/17/2011 05:58 AM, Helmut Eller wrote:
>> * Per Bothner [2011-04-17 07:04] writes:
>>
>>> Technically the "integer" type should probably be renamed "exact-integer"
>>> - that would be more compatible with R6RS and the just-released R7RS draft.
>>
>> In Kawa, all integers are exact, right? If so, then "exact-integer" is
>> IMO unnecessary confusing.
>
> Well, following RnRS terminology 3.0 is an inexact integer.
>
> (integer? 3.0) ==> #t
>
> The new R7RS draft has a separate exact-integer? procedure to capture
> the intersection of integer and exact.
>
> Generally in Kawa, a type predicate (T? x) is equivalent to
> (instance? x T). However, that is not the case for T==integer.
> That's a little unfortunate. We could fix that irregularity by
> renaming the integer type to exact-integer, and changing integer
> to have the RnRS meaning. However, that means people would have to
> update their code to change integer to exact-integer (or suffer some
> performance-loss). Plus it's not clear why you'd ever want to
> use integer in the RnRS sense as a type-specifier.
>
> So my inclination is to just note the inconsistency.
OK, I see. That's indeed very awkward.
Helmut