[Rcpp-devel] Error creating package with RcppArmadillo

Dirk Eddelbuettel edd at debian.org
Sat Oct 13 20:35:50 CEST 2018


Hi Bill,

On 13 October 2018 at 11:13, William Dunlap wrote:
| But do replace pkgKitten's boilerplate before submitting the package to
| CRAN, lest CRAN people decide to insist on meaningful content.
| 
| ./RApiDatetime/man/RApiDatetime-package.Rd:  This section should provide a
| more detailed overview of how to use the
| ./SimplifyStats/man/SimplifyStats-package.Rd:  This section should provide
| a more detailed overview of how to use the
| ./googlePolylines/man/sfencode-package.Rd:  This section should provide a
| more detailed overview of how to use the
| ./multistateutils/man/des-package.Rd:  This section should provide a more
| detailed overview of how to use the
| ./mrbsizeR/man/mrbsizer-package.Rd:  This section should provide a more
| detailed overview of how to use the
| ./xtensor/man/xtensor-package.Rd:  This section should provide a more
| detailed overview of how to use the

Fair point that you raise here. I hope you don't mind if I reply:

1. Pointing users to a helper tool which produces non-working output is
counter-productive and a slap in their face.  I tried many times to convince
Duncan Murdoch (who, while an R Core member, would at least at times listen
to proles like myself).  I failed, and all I could do is write pkgKitten.
But this is, as the saying goes, "a hill I willing to die on". What R does
here is wrong.

2. Your critique is valid, but I have neither carrot nor stick to push on. In
my mind R CMD check could (and should !!) check for this not-replaced-stanza
text.  But you know how it goes about changes in CRAN / R Core.

3. You list several packages. One is mine. It should not even exists if R
Core was not so paranoid about about their API.  Another is one I helped with
so guilty as charged.

4. You list six packages. Rights now Rcpp has 1460 reverse dependencies. I
take six out of 1460 (surely upward biased as not all use package creators)
if it means the first point above is taken care of.

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org


More information about the Rcpp-devel mailing list