renv::snapshot() fails for 1.0.10 version on Windows if run on a package directory that has an exported function called module()

[m7pr]

Thanks for creating renv! This is super useful mechanism that we used in many projects!
We appreciate the time and the energy you invested in building the tool.

ver 1.0.10 issue

Recently we noticed some weird behavior for the newly released renv 1.1.0 on Windows.
If you have an exported function called module and it's used in testthat tests,
then when running renv::snapshot() you can see it fails to create the lockfile.

The reproducible example can be based on this repository https://github.com/m7pr/renv.modules
Just pull it and run renv::snapshot().

That's what I get

> renv::snapshot()
Error in if (package == "R") return() : 
  missing value where TRUE/FALSE needed
In addition: Warning message:
In node[[2L]] : subscript out of bounds

> renv::dependencies()$traceback
Finding R package dependencies ... Done!
 [1] NA                                                                    
 [2] "21: renv::dependencies()"                                            
 [3] "20: renv_dependencies_impl(path = path, root = root, quiet = quiet, "
 [4] "19: renv_dependencies_discover(files, progress, errors)"             
 [5] "15: catch(filebacked(\"dependencies\", path, callback))"             
 [6] "14: tryCatch(withCallingHandlers(expr, error = renv_error_capture), "
 [7] "13: tryCatchList(expr, classes, parentenv, handlers)"                
 [8] " 9: filebacked(\"dependencies\", path, callback)"                    
 [9] " 8: callback(path, ...)"                                             
[10] " 6: renv_dependencies_recurse(expr, function(node) {"                
[11] "        invisible(node)"                                             
[12] "    })"                                                              
[13] " 4: renv_dependencies_recurse_impl(object[[i]], callback)"           
[14] NA                                                                    
[15] NA                                                                    
Warning message:
In node[[2L]] : subscript out of bounds

> sessionInfo()
R version 4.4.0 (2024-04-24 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 11 x64 (build 22621)

Matrix products: default


locale:
[1] LC_COLLATE=English_Europe.utf8  LC_CTYPE=English_Europe.utf8   
[3] LC_MONETARY=English_Europe.utf8 LC_NUMERIC=C                   
[5] LC_TIME=English_Europe.utf8    

time zone: Europe/Berlin
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] renv.modules_0.1.0

loaded via a namespace (and not attached):
[1] compiler_4.4.0    cli_3.6.3         tools_4.4.0       rstudioapi_0.16.0
[5] renv_1.0.10  

ver 1.0.9 no issue

This does not happen on previous version

> remotes::install_github('rstudio/renv@v1.0.9')
> library(renv)
> packageVersion('renv')
[1] ‘1.0.9’
> renv::snapshot()
The following package(s) will be updated in the lockfile:
  
  # CRAN -----------------------------------------------------------------------
- brio        [* -> 1.1.5]
- callr       [* -> 3.7.6]
- crayon      [* -> 1.5.2]
- desc        [* -> 1.4.3]
- diffobj     [* -> 0.3.5]
- digest      [* -> 0.6.35]
- fansi       [* -> 1.0.6]
- fs          [* -> 1.6.4]
- glue        [* -> 1.7.0]
- jsonlite    [* -> 1.8.9]
- lifecycle   [* -> 1.0.4]
- magrittr    [* -> 2.0.3]
- pillar      [* -> 1.9.0]
- pkgbuild    [* -> 1.4.4]
- pkgconfig   [* -> 2.0.3]
- pkgload     [* -> 1.4.0]
- praise      [* -> 1.0.0]
- processx    [* -> 3.8.4]
- ps          [* -> 1.7.6]
- R6          [* -> 2.5.1]
- rematch2    [* -> 2.1.2]
- rlang       [* -> 1.1.4]
- rprojroot   [* -> 2.0.4]
- testthat    [* -> 3.2.1.1]
- tibble      [* -> 3.2.1]
- utf8        [* -> 1.2.4]
- vctrs       [* -> 0.6.5]
- waldo       [* -> 0.5.2]

# GitHub ---------------------------------------------------------------------
- renv        [ver: 1.0.10 -> 1.0.9; ref: v1.0.10 -> v1.0.9; sha: 0687c788 -> ed605b46]

# RSPM -----------------------------------------------------------------------
- evaluate    [* -> 1.0.0]
- withr       [* -> 3.0.1]

# https://packagemanager.posit.co/cran/latest --------------------------------
- cli         [* -> 3.6.3]

Do you want to proceed? [Y/n]: 

Potential places for the fix

Is it a names collision with modules::module that is used inside renv?
Is it some collision with reticulate/python modules?
Is module a reserved name that is omitted or treated differently by renv?

Note

Suprisingly if you call module function with some custom parameter (like x = 1) then this issue no longer appears and you can crate the lockfile.

But if there is no parameter provided, renv::snapshot fails.

[kevinushey]

Thanks for the bug report, and especially for taking the time to provide a reproducible example! This should now be fixed up in the development version of renv; I'll try to get a patch release onto CRAN soon.

[m7pr]

Wow @kevinushey - thanks for the quick fix!