I'd be interested to see 2 reports indicative of that.
The two reports linked in my original post: both showed linux 3.16.0-4-amd64 debian 3.16.7-ckt25-2 (2016-04-08) with gccversion='4.9.2', libc=libc-2.19.so, and gnulibc_version='2.19', the only difference being 5.20.1 vs 5.22.0. But as stevieb replied, it's probably the improved NaN handling in 5.22 which fixed the bug. (Which then surprises me that my ancient linux perl v5.8.5 passed... but then again, there were probably other NaN handling changes in Perl or differences between the systems/compilers).
excerpts from the fail reports in the spoiler
In reply to Re^4: Advice wanted for debugging CPAN Testers failures
by pryrt
in thread Advice wanted for debugging CPAN Testers failures
by pryrt
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |