[Rd] R CMD check problem with R 3.0.2

[Duncan Murdoch]

On 25/10/2013 11:37 AM, Sanford Weisberg wrote:
> Using SUSE Linux, Windows 32 bit and Windows 64 bit R 3.0.2 , I am unable
> to use R CMD check successfully.  Here is the Windows 64 bit report:

Both checking and installing code are really designed to work on 
tarballs, as John said.  Some parts of them work on directories, because 
that can be a lot quicker (e.g. you may not need to recompile a ton of 
files), but you should expect to get some extra warnings or errors.  
Looks like you did.

Since there's an easy workaround for this (build before checking), I 
don't think it will be a high priority to fix, but feel free to submit 
it as a bug report, and someone will likely deal with it eventually.  
Please attach a tiny package that illustrates the error to your bug 
report (e.g. pkg with just about everything taken out).

Duncan Murdoch
>
>
> Z:\R\source\effects>R CMD check pkg
> * using log directory 'Z:/R/source/effects/pkg.Rcheck'
> * using R version 3.0.2 (2013-09-25)
> * using platform: x86_64-w64-mingw32 (64-bit)
> * using session charset: ISO8859-1
> * checking for file 'pkg/DESCRIPTION' ... ERROR
> Required fields missing or empty:
>    'Author' 'Maintainer'
>
> The file DESCRIPTION looks like this:
>
> Package: effects
> Version: 2.3-0
> Date: 2013/10/22
> Title: Effect Displays for Linear, Generalized Linear, Multinomial-Logit,
> Proportional-Odds Logit Models and Mixed-Effects Models
> Authors at R: c(person("John", "Fox", role = c("aut", "cre"), email = "
> jfox at mcmaster.ca"),
>      person("Sanford", "Weisberg", role = "aut", email = "sandy at umn.edu"),
>      person("Michael", "Friendly", role = "aut", email = "friendly at yorku.ca
> "),
>      person("Jangman", "Hong", role = "aut"),
>      person("Robert", "Andersen", role = "ctb"),
>      person("David", "Firth", role = "ctb"),
>      person("Steve", "Taylor", role = "ctb"))
> Depends: lattice, grid, colorspace
> Suggests: nlme, lme4, MASS, nnet, poLCA, heplots
> LazyLoad: yes
> LazyData: yes
> Description:
>    Graphical and tabular effect displays, e.g., of interactions, for linear
>    generalized linear, multinomial-logit, proportional-odds logit models,
>    mixed-effect models,  polytomous latent-class models and multivariate
> linear models.
> License: GPL (>= 2)
> URL: http://www.r-project.org, http://socserv.socsci.mcmaster.ca/jfox/
>
> The 'Author' and 'Maintainer' fields should be automatically generated.
> With version 3.0.1, I had no such problem, and my coauthors who use Eciplse
> and R-studio have no problems with R 3.0.2.  John Fox suggested the
> following:
>
> When R CMD build creates a package tarball it writes the information from
> Authors at R into the
> Author and Maintainer fields. I think that Sandy's
> problem is produced by checking the package source directory rather than a
> package source tarball. I use RStudio to check packages, and it
> automatically builds the tarball first, which is the recommended procedure.
> R-Forge does that too. Michael uses Eclipse, and if I remember right, it too
> creates a tarball (but I haven't used it in quite some time).
>
> Is this a bug in R CMD check?
>
>