[Rd] R-devel Digest, Vol 105, Issue 22

[Terry Therneau]

On Tue, 2011-11-22 at 12:00 +0100, r-devel-request at r-project.org wrote:
> How are the package authors supposed to develop their own NAMESPACEd
> packages efficiently? And what are the directions R is taking in order
> to
> facilitate the development cycle?
> 

This is my strategy.
I have a separate directory "test.local" in my tree, not exported to
Rforge.  It has a Makefile which loads all the sources from ../R, copies
the C files from ../src and makes a local loadable S.so.  I then do all
my development there, without a name space.  I can overwrite functions,
trace, add browser() calls --- all the things you want to do --- with
standard semantics. I obviously don't "load" my package.
  
I think this is the simplest route.  Namespaces are a great idea, but
during testing I don't want all those protections.  It took me a half
hour to set up the Makefile; I've recouped that effort many times over.

Terry Therneau