Closed richelbilderbeek closed 7 years ago
nLTTstat_exact
function, nLTTstat_exact_classic
Fails just the same!
Would there be a back-end problem?
Oops.
Experiment: re-run older successful version!
Before:
Yuck, that one fails due to HMisc
being absent. Tme to Go Back In Time and create a branch called 'patch26'.
Older version also fails:
Also locally, the old version fails:
@thijsjanzen : it appears something has changed beyond the scope of our GitHub
I can fix this by replacing
testthat::expect_equal(
mean(A),
ML$minimum[[1]],
tolerance = 0.05
)
by
testthat::expect_equal(
mean(A),
ML$minimum[[1]],
tolerance = 0.052 # Higher tolerance
)
Do I have a go?
I'm unsure why this fails, but a tolerance of 0.052 seems reasonable (however, with the same seeds, this should never fail). Also, can't the HMisc issue be solved other than reverting to an old version?
OK, I'll increase the tolerance.
Fixing Hmisc
is just removing it from DESCRIPTION
.
Feel free to transition the package to packrat @richelbilderbeek :D
Last commit that works: c3e44977d5f7cd3b0a223cfce914c312090bfdf3
I think this is due to a changed
nLTTstat_exact
function.