"I get slightly different behavior:" - that's not surprising, given 7
months' difference between compiler versions.
The division's 0 is probably justified, depending on your definition
of epsilon ("Close enough for government work"), but is 1 a reasonable
substitute for "an implausibly large number"?
How far beyond the values of the physical universe does a practical
programming language need to go in pursuit of mathematical
consistency? 10 ** 100 appears to accommodate the number of protons in
the known universe with a good deal to spare.
On 3/18/16, Will Coleda via RT <perl6-bugs-follo...@perl.org> wrote:
> On Sat Aug 29 01:59:19 2015, chr...@gmail.com wrote:
>> 2**10000000000 and 1/(2**10000) both silently return 0.
>>
>> $ perl6 --version
>> This is perl6 version 2015.07.1-177-g5fb81ff built on MoarVM version
>> 2015.07-68-g3240047
>>
>> Christian.
>
> I get slightly different behavior:
>
> $ perl6 -e 'say 2**10000000000'
> 1
> $ perl6 -e 'say 1/(2**10000)'
> 0
> $ perl6 --version
> This is Rakudo version 2016.02 built on MoarVM version 2016.02
> implementing Perl 6.c.
>
> --
> Will "Coke" Coleda
>
