Skip to content

Conversation

eitsupi
Copy link
Contributor

@eitsupi eitsupi commented May 31, 2025

Same as #1780

Copy link
Member

@shikokuchuo shikokuchuo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @eitsupi looks good to me!

Comment on lines +68 to +75
if (is.null(version)) {
loaded <- lapply(pkg, function(x) {
is.character(x) && nzchar(x) && is.environment(.getNamespace(x))
})
if (all(as.logical(loaded))) {
return(TRUE)
}
}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this a typical intended interface? is_installed(list("dplyr", "dbplyr"))

If not, we can move the checks "outside the loop" toget early exits for any of the !is.character(pkg) and !all(nzchar(pkg)) cases.

Next, why use is.environment(.getNamespace(x)) instead of isNamespaceLoaded(x)? The former is described as an "internal" (hence the leading .), the latter is more readable (and nanosecond-scale-faster).

Suggested change
if (is.null(version)) {
loaded <- lapply(pkg, function(x) {
is.character(x) && nzchar(x) && is.environment(.getNamespace(x))
})
if (all(as.logical(loaded))) {
return(TRUE)
}
}
if (is.null(version) && is.character(pkg) && all(nzchar(pkg)) && all(vapply(pkg, isNamespaceLoaded, NA))) {
return(TRUE)
}

Copy link
Member

@shikokuchuo shikokuchuo Jun 11, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The argument is expected to be a character vector c("dplyr", "dbplyr") rather than a list, so your suggestion would work well in that case. A list would just fail the fast path but continue, so is not disastrous.

If we make the change here, we should do the same at check_installed().

Thanks for the suggestion of isNamespaceLoaded(x). It is indeed faster as it shares the same C level code path as .getNamespace(x) but just returns a logical value rather than the environment itself.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@MichaelChirico sorry somehow I missed your entire code suggestion, so my inital reply was totally off the mark (now amended above).

Copy link
Contributor Author

@eitsupi eitsupi Jun 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the suggestions!

Would it be better to refactor this with check_installed after merging this since my intention of this PR is just to bring the same changes here as check_installed in #1780?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes that's fine thanks! If Lionel adopts the suggestion here, I will follow up with the corresponding change to check_installed().

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants