[Rd] Depends: R (>= 4.1) for packages that use |> and \(...)

[Kurt Hornik]

>>>>> Ivan Krylov via R-devel writes:

Currently, we have 146 CRAN packages for which the check code finds 4.1
or later syntax additions in the Rd file examples only.  

Ivan kindly tried the effect of checking these with 4.0 (thanks again: I
cannot "easily" do this as Debian testing has libcurl 8 and the 4.0
configure code insists on libcurl 7).  This had 140 packages fail with
parse errors, 3 more for other reasons, and packages

  gmailr httr2 purrr

still pass as using the configure plus tools/examples.R renewcommand for
\examples trick.

So in c87900 I just committed I have simply stoplisted these three in
the check code, and the false positive NOTEs for these should be gone
when this gets picked up.

Thanks for spotting and helping to fix :-)

Best
-k

> В Wed, 5 Mar 2025 07:47:04 -0600
> Hadley Wickham <h.wickham using gmail.com> пишет:

>> Unfortunately your test generates a false positive for httr2 (
>> https://cran.r-project.org/web/checks/check_results_httr2.html) and
>> other tidyverse packages where we use the base pipe in examples, but
>> carefully disable them for older versions of R.

> Please accept my apologies. Indeed, the script [1] called at
> configuration time does replace the examples section to avoid parse
> errors for example(...). And having example() do nothing (with all code
> and an explanation in the help page) is a small cost for having the
> rest of the package work on R versions as old as 3.5.

> I don't see a way to take this into account, since the workaround is
> completely invisible to sufficiently new versions of R. It might be
> that the best way forward is to revert the Rd example check.

> -- 
> Best regards,
> Ivan

> [1] https://github.com/r-lib/httr2/blob/main/tools/examples.R

> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel