I made a number of changes:
1.) made inlinedocs a bit more forgiving with respect to missing parts of the docsstrings which was necessarry for
the package we just work on. Inlinedocs had stopped with weird error messages while pasting the gathered bits
together. (or rather not if some bits where missing...)
2.) R original package.skeleton adds a lot of (duplicated) aliases to method names to a classes Rd file.
These cause R CMD check to complain about the duplication.
I added a new function which is called after modify.Rd.file and comments out the superfluous aliases.
This function does some things that are also done in apply.parsers and modify.Rd.file
which could (and should) be probably factored out.
For instance the new function also evaluates the whole code in a separate environment (like apply parser)
and it also could use the more safeguarded way to compute the final name of an Rd.file (like modify.Rd.file does)
However manipulating environments (explicitly ;-))still scares me a bit.
So I at least wanted to make an intermediate commit to invite discussion and make the changes more reproducible.
3.) I added an additional check before inlinedocs removes the alias to the package name in the package-Rd file.
It had previously removed it from our documentation although there had >> not << been a function of the same name.