Environment variable MAKE not defined on Windows

On Windows,
Sys.getenv("MAKE")
MAKE 
"" 

when (at a DOS prompt)

X:>set MAKE
Environment variable MAKE not defined

On linux,
Sys.getenv("MAKE")
MAKE 
"make" 

even when (in a shell)

% echo $MAKE

returns nothing.

This has consequences, e.g., in buildVignette where

system(Sys.getenv("MAKE"))

dutifully does nothing on Windows, but evaluates the vignette Makefile
on linux.
sessionInfo()
R version 2.4.0 Under development (unstable) (2006-06-08 r38315)
i386-pc-mingw32
sessionInfo()
R version 2.4.0 Under development (unstable) (2006-05-25 r38197) 
x86_64-unknown-linux-gnu 

Martin
Bioconductor
On Windows,

Sys.getenv("MAKE")
  MAKE 
"" 

when (at a DOS prompt)

X:>set MAKE
Environment variable MAKE not defined

On linux,

Sys.getenv("MAKE")
  MAKE 
"make" 

even when (in a shell)

% echo $MAKE

returns nothing.

This has consequences, e.g., in buildVignette where

system(Sys.getenv("MAKE"))

dutifully does nothing on Windows, but evaluates the vignette Makefile
on linux.

sessionInfo()
R version 2.4.0 Under development (unstable) (2006-06-08 r38315)
i386-pc-mingw32

sessionInfo()
R version 2.4.0 Under development (unstable) (2006-05-25 r38197) 
x86_64-unknown-linux-gnu 
Most Windows systems don't have a working "make" command, and I don't 
know of a reliable test for one.  (I believe the one that ships with MS 
compilers is not compatible with R's makefiles.)  If you do have one, 
you can define MAKE=make on the Rgui or Rterm command line.

Duncan Murdoch
Duncan Murdoch <murdoch at stats.uwo.ca> writes:
On 6/9/2006 4:03 PM, Martin Morgan wrote:
On Windows,

Sys.getenv("MAKE")
  MAKE "" when (at a DOS prompt)
X:>set MAKE
Environment variable MAKE not defined
Most Windows systems don't have a working "make" command, and I don't
know of a reliable test for one.  
I guess on non-Windows systems MAKE gets defined when R is
built. Perhaps a similar thing should happen when I build my own R on
Windows, because then I must have a working make?

I realize that the windows R is distributed in binary form; perhaps
these could be configured not to include MAKE and other environmental
variables? For instance, I notice some unlikely-to-be-generally-useful
variables like PERL5LIB and TEXINPUTS hard-coded in

R_HOME/src/gnuwin32/front-ends/rcmdfn.c

Also, the documentation ?Startup is not correct, at least if 'factory
fresh' means newly built -- there does not appear to be an
etc/Renviron.site on Windows.

Description:

   The name of the site file is the one pointed to by the environment
   variable 'R_ENVIRON'; if this is unset or empty,
   '$R_HOME/etc/Renviron.site' is used (if it exists, which it does
   not in a "factory-fresh" installation except on Windows). 

Finally, the buildVignettes() code might be modified to more
forcefully notify the user of the inadequacies of their system.

Martin
(I believe the one that ships with MS compilers is not compatible
with R's makefiles.)  If you do have one, you can define MAKE=make
on the Rgui or Rterm command line.

Duncan Murdoch
Duncan Murdoch <murdoch at stats.uwo.ca> writes:

On 6/9/2006 4:03 PM, Martin Morgan wrote:
On Windows,

Sys.getenv("MAKE")
  MAKE "" when (at a DOS prompt)
X:>set MAKE
Environment variable MAKE not defined
Most Windows systems don't have a working "make" command, and I don't
know of a reliable test for one.  
I guess on non-Windows systems MAKE gets defined when R is
built. Perhaps a similar thing should happen when I build my own R on
Windows, because then I must have a working make?
The problem is that the natural place to put it is in the shortcut that 
starts R (or the command alias you use to start it from a command line). 
  We don't make a shortcut unless you install R from the binary 
installer, and don't make an alias at all.  If you start R from the 
command line, you're running an .exe directly.

I think the best solution is the one I suggested below:  do it yourself.
I realize that the windows R is distributed in binary form; perhaps
these could be configured not to include MAKE and other environmental
variables? For instance, I notice some unlikely-to-be-generally-useful
variables like PERL5LIB and TEXINPUTS hard-coded in

R_HOME/src/gnuwin32/front-ends/rcmdfn.c

Also, the documentation ?Startup is not correct, at least if 'factory
fresh' means newly built -- there does not appear to be an
etc/Renviron.site on Windows.

Description:

   The name of the site file is the one pointed to by the environment
   variable 'R_ENVIRON'; if this is unset or empty,
   '$R_HOME/etc/Renviron.site' is used (if it exists, which it does
   not in a "factory-fresh" installation except on Windows). 
Right, that is wrong.  The "except on Windows" part is a recent 
addition; I'm not sure why, but I'll remove it.
Finally, the buildVignettes() code might be modified to more
forcefully notify the user of the inadequacies of their system.
I'll add that.

Thanks for the suggestions.

Duncan Murdoch
Martin   

(I believe the one that ships with MS compilers is not compatible
with R's makefiles.)  If you do have one, you can define MAKE=make
on the Rgui or Rterm command line.

Duncan Murdoch

______________________________________________
R-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel