Testing output different between R 3.5.4 and 3.6


I have two development environments - a MacBook Pro and an iMac; the MacBook Pro is currently on 3.5.4 and the iMac on 3.6. The packages themselves are identical - in particular both are running test that 2.1.1.

I notice when I run my unit tests on the 3.6 machine that the output is different: in particular I'm not getting the report of number of tests passed / failed, and the time taken for the tests (which was quite useful as a very rough benchmark).

Is there a way to get the output I'm seeing on the 3.5.4 build (which matches that on the test that 2.0 announcement at https://www.tidyverse.org/articles/2017/12/testthat-2-0-0/ ) on the 3.6 build?

The random number generator changed in R3.6, so that may be the issue.

You can revert to previous behaviour via:

Interesting thought. Just tried it though - and no change.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.