# Having only tested on macOS, because Stan results will only be exactly reproducible # if all of the following components are identical: # Stan version # Stan interface (RStan, PyStan, CmdStan) and version, plus version of interface language (R, Python, shell) # versions of included libraries (Boost and Eigen) # operating system version # computer hardware including CPU, motherboard and memory # C++ compiler, including version, compiler flags, and linked libraries # same configuration of call to Stan, including random seed, chain ID, initialization and data skip_on_cran() skip_on_appveyor() test_that("Test-all, for-loop", { options(warn=-1) for (method in 0:6) { for (var.equal in c(T, F)) { for (design in c("within", "between")) { for (treat in c("random", "fixed")) { for (diagnostics in c(F, T)) { test <- rmHDI(data.wide = recall.wide, seed = 277, method = method, var.equal = var.equal, design = design, treat = treat, diagnostics = diagnostics, chain = 1, warmup = 5, iter = 10) expect_gte(length(test),3) } } } } } options(warn=0) })