search

statnet / ergm.ego

Fit, Simulate and Diagnose Exponential-Family Random Graph Models to Egocentrically Sampled Network Data https://statnet.org
Other
14 stars 4 forks source link

`data("fmhfit")` #72

Closed romainfrancois closed 2 years ago

romainfrancois commented 2 years ago

We're in the process of releasing dplyr 1.0.8 and our revdep checks have picked up this potential problem, which however feels unrelated to dplyr: 

W  checking contents of ‘data’ directory (3.9s)
   Output for data("fmhfit", package = "ergm.ego"):
     Search path was changed

I'm not sure what is the problem though. The error comes from this function that is run as part of R CMD check : 

.check_package_datasets2 <-
function(fileName, pkgname)
{
    oldSearch <- search()
    dataEnv <- new.env(hash = TRUE)
    suppressMessages(utils::data(list = fileName, package = pkgname,
                                 envir = dataEnv))
    if (!length((ls(dataEnv)))) message("No dataset created in 'envir'")
    if (!identical(search(), oldSearch)) message("Search path was changed")
    invisible(NULL)
}
mbojan commented 2 years ago

Thanks @romainfrancois. Does making the data loading lazy solve it? That would be strange indeed...

romainfrancois commented 2 years ago

yeah, #73 solves this for some reason

mbojan commented 2 years ago

The package passes all the checks on all platforms without it. Is it possible that some thing is not right with your revdep testing infrastructure?

mbojan commented 2 years ago

@romainfrancois I'll try to have a look myself. Where's the dplyr version you are planning to release? rstudio/dplyr@master?

romainfrancois commented 2 years ago

yeah, except the default branch is now main, you can get it with: pak::pkg_install("tidyverse/dplyr")

mbojan commented 2 years ago

Thanks.

PS. So is remotes::install_github() falling out of favor? ;-)

romainfrancois commented 2 years ago

They serve different purpose, pak::pak() is more for interactive use, and remotes for use in e.g. GitHub actions

mbojan commented 2 years ago

OK, thanks. I have some difficulties with getting the dev version of dplyr to work, so perhaps I'll just merge it...