[perl #134171] ext/POSIX/t/math.t failure

On Mon, 03 Jun 2019 19:18:10 GMT, rares.aioanei@gmail.com wrote:
> Subject: ext/POSIX/t/math.t failure
> Message-Id: <5.30.0_11724_1559578277@openindiana>
> From: rares.aioanei@gmail.com
> Reply-To: rares.aioanei@gmail.com
> Cc: arares@openindiana.local
> To: perlbug@perl.org
> 
> 
> This is a bug report for perl from rares.aioanei@gmail.com,
> generated with the help of perlbug 1.41 running under perl 5.30.0.
> 
> 
> -----------------------------------------------------------------
> ext/POSIX/t/math.t fails on Solaris (OpenIndiana Hipster) with :
> 
> #   Failed test 'tanh(-1)'
> #   at t/math.t line 55.
> #     '0.761594155955765'
> #      <=
> #     '-0.76'
> 
> #   Failed test 'tanh(1) == -tanh(-1)'
> #   at t/math.t line 56.
> #       got: 0.761594155955765
> #     expected: -0.761594155955765
> 
> 

Are you in a position to run a test where you add the following option to Perl's ./Configure program?

#####
-Duse64bitall
#####

That is, configure and build like this:

#####
sh ./Configure -des -Dusedevel -Duse64bitall && make test_prep
#####

... and then run:

#####
cd t;./perl harness -v ../ext/POSIX/t/math.t;cd -
#####

I ask, because if the test in question PASSes when perl is configured with '-Duse64bitall', that would be consistent with some smoke test reports we have been receiving, e.g.,

http://perl5.test-smoke.org/report/88747
http://perl5.test-smoke.org/report/88826

Unfortunately, while we have been receiving these smoke-test reports, we haven't been in position to fully diagnose this on Solaris.

Thank you very much.

-- 
James E Keenan (jkeenan@cpan.org)

---
via perlbug:  queue: perl5 status: new
https://rt.perl.org/Ticket/Display.html?id=134171

• [perl #134171] ext/POSIX/t/math.t failure by sisyphus@cpan.org via RT
• [perl #134171] ext/POSIX/t/math.t failure by James E Keenan via RT