Testing results for Bert's suggestion...
I tested this with a different -- much smaller -- package: Put the S3method()
construct in NAMESPACE but the generic package is not mentioned at all in any
DESCRIPTION field. I built the package and checked it --as-cran. Interesting
results:
* Calling multcomp::cld(object) worked just fine without my doing anything to
register the method.
* In checking, there was no objection to my not importing the generic for cld.
So this looks like a potentially good solution, except for the documentation
issue.
BTW, note this parallel:
* In NAMESPACE, the construct S3method(pkg::gen, cls)
* In Rd files, the macro \S3method{pkg::gen}{cls}(args)
The latter is what I am asking about, that creates a warning. However, the
parallel between these just cries out to be valid!
Russ
From: Bert Gunter <bgunter.4...@gmail.com>
Sent: Thursday, August 6, 2020 12:22 PM
To: Lenth, Russell V <russell-le...@uiowa.edu>
Subject: Re: [R-pkg-devel] Documenting courtesy S3 methods
Probably unnecessary to say, but if/when you determine whether the package must
be installed, let us all know.
Bert
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
