dragosmg commented on code in PR #13620:
URL: https://github.com/apache/arrow/pull/13620#discussion_r930031860
##########
.github/workflows/r.yml:
##########
@@ -327,6 +327,13 @@ jobs:
shell: Rscript {0}
working-directory: r
run: |
+ Sys.setenv(
+ RWINLIB_LOCAL = file.path(Sys.getenv("GITHUB_WORKSPACE"), "r",
"windows", "libarrow.zip"),
+ MAKEFLAGS = paste0("-j", parallel::detectCores()),
+ ARROW_R_DEV = TRUE,
+ "_R_CHECK_FORCE_SUGGESTS_" = FALSE
+ )
+ pak::pak("local::.")
Review Comment:
AFAIU `pak::pak()` and `pak::pkg_install()` are equivalent
(https://pak.r-lib.org/reference/pak.html#details-1) and so are
`pak::local_install()` and `pak::pkg_install("local::.")`
(https://pak.r-lib.org/reference/local_install.html#details-1)
=> `pak::local_install()` and `pak::pak("local::.")` are equivalent.
Are you thinking that `pak::local_install()` is a bit more readable? I would
think people modifying this code are somewhat familiar with `pak`, and thus
they encountered the `"local::."` incantation. For me, the `local_` pattern is
less readable as there is a bit of overlap and different meaning than
`withr::local_`. I read `local_` as we're doing something locally (which
involves a temporary state of something), whereas in this context it actually
means we are installing the package located in the root directory (more about
location). From this point of view, I think `"local::."` is less ambiguous.
But happy to change if you think otherwise.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
