pow edge cases with rakudo-moar on NetBSD

[p6rt]

Migrated from rt.perl.org#124147 (status was 'open')

Searchable as RT124147$

[p6rt]

From @stmuk

Various libm (power) issues with rakudo-moar on NetBSD which show themselves as roast failures and are known issues with NetBSD itself

see http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=49240

eg.
1**Inf is evaluated as NaN rather than Inf (although 1.1**Inf is correctly evaluated as Inf)
- I fudged this in S03-operators/arith.t

Some other other issues are fixable with _LIB_VERSION = _IEEE_ details to follow

[p6rt]

From @stmuk

On Sun Mar 22 11​:54​:48 2015, steve.mynott+bitcard@​gmail.com wrote​:

Various libm (power) issues with rakudo-moar on NetBSD which show
themselves as roast failures and are known issues with NetBSD itself

see http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=49240

eg.
1**Inf is evaluated as NaN rather than Inf (although 1.1**Inf is
correctly evaluated as Inf)

Typo correct behaviour is

$ perl6
say> 1**Inf
1

- I fudged this in S03-operators/arith.t

Some other other issues are fixable with _LIB_VERSION = _IEEE_ details
to follow

[p6rt]

From @stmuk

Also see discussion in #​85750​: Test failures on OpenBSD
(although note this no longer seems to be a problem on OpenBSD)

[p6rt]

@usev6 - Status changed from 'new' to 'open'

[p6rt]

From @usev6

There was a report for similar test failures on Solaris on #moarvm today​: https://irclog.perlgeek.de/moarvm/2017-10-20#i_15328975

Also, we had ticket #​69464 (rejected) with a similar report for OpenSolaris.

I'm going to add solaris as a platform to this ticket.

[p6rt]

From @usev6

There was a report for similar test failures on Solaris on #moarvm today​: https://irclog.perlgeek.de/moarvm/2017-10-20#i_15328975

Also, we had ticket #​69464 (rejected) with a similar report for OpenSolaris.

I'm going to add solaris as a platform to this ticket.