Hi, I might be wrong but I'm using testthat to test a package and this issue has made me think the issue I'm facing might be due to rprojroot via testthat::test_check ->> rprojroot::is_testthat or something similar.
When I submit the package rtweet to devtools::checks_win_oldrelease(), it fails (temporary link) the checks because root directory is not found. This only happen with oldrelease in windows (with the current pre-release version) not on CRAN or other R versions of the presubmission checks:
Running the tests in 'tests/testthat.R' failed.
Complete output:
> # Use conditionally testthat
> # if (requireNamespace("rtweet", quietly = TRUE)) {
> library(testthat)
> library(rtweet)
>
> test_check("rtweet")
Error: No root directory found in d:/RCompile/CRANguest/R-oldrelease/rtweet.Rcheck/tests_i386/testthat or its parent directories. Root criterion: contains a directory "tests"
Execution halted
** running tests for arch 'x64' ... [2s] ERROR
Running 'testthat.R' [1s]
Running the tests in 'tests/testthat.R' failed.
Complete output:
> # Use conditionally testthat
> # if (requireNamespace("rtweet", quietly = TRUE)) {
> library(testthat)
> library(rtweet)
>
> test_check("rtweet")
Error: No root directory found in d:/RCompile/CRANguest/R-oldrelease/rtweet.Rcheck/tests_x64/testthat or its parent directories. Root criterion: contains a directory "tests"
Execution halted
I'm not sure why is this happening and why hasn't happened before (that I could find) or if it is worth fixing if R will soon(?) drop support for 32 bits systems. But I guess it was better to report it
Hi, I might be wrong but I'm using testthat to test a package and this issue has made me think the issue I'm facing might be due to rprojroot via
testthat::test_check
->>rprojroot::is_testthat
or something similar.When I submit the package rtweet to
devtools::checks_win_oldrelease()
, it fails (temporary link) the checks because root directory is not found. This only happen with oldrelease in windows (with the current pre-release version) not on CRAN or other R versions of the presubmission checks:I'm not sure why is this happening and why hasn't happened before (that I could find) or if it is worth fixing if R will soon(?) drop support for 32 bits systems. But I guess it was better to report it