forcing R CMD COMPILE

Let me tell an example.

I found the bayesSurv package in CRAN. It has some .cpp and .h files, and
doesn't have a makefile. I did the following steps:

1. compile the package (R CMD INSTALL)
2. write something in /src/List.cpp (only add an empty line)
3. recompile package (INSTALL)
  it works, OK.
4. write something in /src/List.h
5. recompile
  it does nothing. List.cpp was not changed and so R says List.o is up
  to date, right?

Is there any way to tell R that List.o depends on List.h???

I think it is not an R problem, because Makefiles can check it. My
question is: I know that List.o is not up to dated, because I changed
List.h, and my Makefile recognizes it. But if I tell R to compile List.cpp
again using COMPILE, it will say that the file is up to date, and won't
compile it. So, the only way to check dependencies, as I can see, is using a
Makefile, and compiling .cpp files without any R help. Is it true?

On Wed, 23 Mar 2005, Prof Brian Ripley wrote:

On Wed, 23 Mar 2005, Pedro Ribeiro de Andrade Neto wrote:

On Wed, 23 Mar 2005, Prof Brian Ripley wrote:

On Wed, 23 Mar 2005, Pedro Ribeiro de Andrade Neto wrote:

I am developing a package with a lot of C++ code, and I have a question
about R CMD COMPILE. As I can see, when the package's Makefile calls

R CMD COMPILE foo.cpp

Why does your package have a Makefile?  And why is it calling COMPILE and
not SHLIB?

My package has a Makefile because I need to compile 16 .cpp files. After
it, I call R CMD SHLIB.

R CMD SHLIB would do that for you.  Many packages using multiple .cpp
files have no Makefile.  Please study the many examples on CRAN.

--
Brian D. Ripley,                  ripley@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