possible test failure in upcoming Julia version 1.5

[rfourquet]

A PkgEval run for a PR which changes the generated numbers for randn! indicates that the tests of this package might fail in Julia 1.5 (and on Julia current master). Apologies if this is a false positive.
cf. https://github.com/JuliaCI/NanosoldierReports/blob/7de24e455342298cbef56826b5827f0d7640d2c1/pkgeval/by_hash/b89e35c_vs_098ef24/logs/MIRT/1.5.0-DEV-71a4a114c2.log

[JeffFessler]

Thanks for alerting me. It turns out that I had a tolerance that was too small there, so I have changed it to use isapprox and I suspect it will be fine now. It passes Julia 1.4 tests. I tried to test 1.5 on TravisCI but I guess they are not yet supporting 1.5. So I downloaded the latest nightly build of 1.5 and tested it locally and it passed.
It's wonderful that the Julia community has these tools that help us be proactive! Thanks!

[rfourquet]

You're welcome! Actually the PR was merged just before opening the issue, so I guess the "latest nightly build" doesn't contain yet the change. But yes, using isapprox will probably be enough (but the change also broke some tests which had isapprox).

[JeffFessler]

I tested with a newer nightly and it passed, after fixing some other new issues. Thanks again.