[Rd] Roxygen: question regarding 'use.Rd2' and creation of DESCRIPTION

Janko Thyson janko.thyson at ku-eichstaett.de
Wed Sep 15 15:29:21 CEST 2010

Dear List,

I ran into the following two problems while using the package 'roxygen':

I split the relevant R-Code for my package into the following scripts:
classes.R (S4), functions.R ('standard' functions), generics.R (S4),
methods.R (S4). Function package.skeleton() generates Rd-files for all class
defs, function defs etc. in dir 'pkg/man'. So far, so good. Now, I'd like to
run 'roxygenize()' on dir 'pkg' and face the problem that the argument
'use.Rd2' only works for parts of the Rd-files, no matter how it is
specified. Setting it FALSE works for all non-S4 defs, setting it TRUE works
for all S4-defs ('works' in the sense of ending up with non-empty \title{}
in the Rd-files). However, both types of defs are in 'pkg/man' or 'pkg/R',

I implemented a workaround in which each of the four scripts is addressed
separately by package.skeleton() and roxygenize(). The resulting Rd-files
are stored in temp dirs (e.g. 'pkg/tmp_classes', 'pgk/tmp_functions' etc.)
and are merged to 'pkg/man' at the very end of the process. But I figured
there must be a better or at least more elegant way to do this. Any hints? 

I should mention that not all of my defs (be it non-S4 or S4) are already
prepended with roxygen-code (related to this post
ions), yet the workaround creates non-empty \title{} for all Rd-files. Is
this also possible by running package.skelleton() and roxygenize() just

Even though I followed the roxygen vignette closely and also had a look at
the webvis package as suggested here
umentation, I still cannot reproduce a DESCRIPTION file that reflects the
specifications I made via the part of the roxygen-code that is related to
stating the package title, author etc. The outcome in the Rd file
'<pkg_name>-package.Rd' is fine, but the DESCRIPTION file itself still looks
like the dummy created by package.skelleton. Am I doing something wrong

Thanks very much for any suggestions,

More information about the R-devel mailing list