There is a problem with parglmVS_test.go. It is too large, so GitHub workflow exceeds execution time, as we can see in the picture below:

Since in commit d80248e code passed the tests that do not use the nested parameter and after that commit there are no changes in the parglmVS.R code, I am going to deactivate that part of the tests, so that the part associated with the nested parameter passes and we finish testing the code.

The problem with this test is that if I do not use enough permutations, p-values of the R version and Matlab version will differ since they depends on random permutations (their random number generator are different even if I set a seed) and it will not pass the test because of differences in the results. On the other hand, if I use too much permutations, GitHub Workflow will exceed the execution time.

Anyways, at weekend I will execute it in local, so I can upgrade a capture of all the test passed together; but by the moment I will deactivate the part of the test already passed on commit d80248e