[Rd] proposal for new flag to R CMD INSTALL

[Prof Brian Ripley]

Why do you think dependencies are not needed for the help system? At 
some point they will be needed to resolve cross-references: at install 
time if you are generating static HTML.

I've not seen any issues that cannot be solved by --install=fake: I 
have fake installs of a few packages (e.g. ROracle) to allow others to 
be tested (including their help pages).


On Mon, 29 Nov 2010, Kjetil Halvorsen wrote:

> For the purpose of helping in installing only parts of a package
> (in my case, the help system), R CMD INSTALL
> should accept a flag   --no-check-deps
> Below is a diff for R-devel, svn revision    53672
>
>
>
> kjetil at kjetil:~/R/R-devel/src/library/tools/R$ diff install.R.old install.R
> 116a117
>>             "      --no-check-deps      skip test if installed depends/imports",
> 506c507
> <             if (length(miss) > 1)
> ---
>>             if ((length(miss) > 1) && check_deps)
> 510c511
> <             else if (length(miss))
> ---
>>             else if (length(miss) && check_deps)
> 1025a1027
>>     check_deps <- TRUE
> 1133a1136,1137
>>         } else if (a == "--no-check-deps") {
>>             check_deps <- FALSE
>
>
>
> kjetil
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595