lSM lICE: not found

Hubert Palme
Tue, 5 May 1998 16:44:09 +0200 (MDT)

Paul Gilbert: 
 > Ross> The intent of "make clean" is to remove the temporaries but leave a
 > Ross> complete runnable R.  It is quite useful when you have installed
 > Ross> versions of R.  Maybe the "make clean" / "make distclean" distinction
 > Ross> needs to be explicit in the "INSTALL" file.
 > I had assumed "make clean" would leave a complete runnable R, that is my
 > runnable R wouldn't get messed up until I did "make" which does "install", which
 > can potentially mess up my running R. But I had also assumed that when I did
 > "make" I would be starting clean, which is not completely true it seems - at
 > least as far as the architecture and other stuff as determined by configure are
 > concerned.
 > I'm not sure what would be the best approach to make things easy for novices. I
 > don't use "make clean" very often for R and it is probably more of a utility for
 > developers - but I did get caught this time. My usual approach is just to delete
 > everything and start over with the tar file. I think, however, that anything
 > like "make distclean" or "make realclean" or "make acclean" is going to be even
 > more of a utility for developers - they all assume you are familiar with the
 > Makefile, or, if they are documented, that you read the documentation
 > (regularly, since these things change).
 > Probably the best thing to make it easy for novices is to have some "install" or
 > "setup" script which does all of the appropriate steps. This should also make
 > certain that everything is sufficiently clean that there won't be any mistakes.

I agree!

The (good!) tradition Peter Dalgaard addresses means

- "make" -- OK, we all know...

- "make clean" returns to the state after configure

- "make distclean" returns to the state before configure

- "make install" installs the whole at the standard place /usr/local
which may be redirected by "configure --prefix=..."

Everyone knows it and one may be irritated if there is another
arrangement (OK, may be I become old :-)

I allready wanted to suggest you authors to add that install target

- I don't like to build the whole stuff in /usr/local
- I don't like to have the R script point to any build directory
- I don't want to think about which files have to be moved to
/usr/local for the program to run correctly. The author knows that (I
hope :-)

As for R, there is no clean separation between the build stuff and the
run stuff. I would like to have it.

But anyway: Thanks for that program!

Hubert Palme         Bergische Universitaet-Gesamthochschule Wuppertal
                                      Computing  Center
                                      D-42097 Wuppertal
Email:             (Germany)
r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: