similar to: R-1.5.0: compiling package foreign on WinNT failed

Has anyone had success building a package on a windows machine? I am trying the following command with no luck (perhaps my flags are incorrect?). At this point I am stuck. Any help greatly appreciated. Niels Waller (BTW - I have no idea why it cannot find sh.exe since I have it is a directory that is specified in a path statement) H:\R\bin>rcmd build --binary --use-zip h:\taxon2 * checking

I just installed the R-patched 1.4.0 and for some reason I now have trouble compiling R packages - I could compile it a few weeks ago with R 1.4.0 released in December. This only happens if I put the --binary option in. For example: Rcmd build --binary cw returns the error: make: Entering directory `/cygdrive/c/PROGRA~1/R/RW1040/src/gnuwin32' make DLLNM= \ -C C:/MyRWork/cw PKG=cw

I'm using R 1.4.1 under Windows 2000 and am experiencing difficulty installing a package. I've included an example output using Rcmd build, but the same problem occurs with Rcmd check and Rcmd INSTALL. The error (make[1]: /bin/sh.exe: Command not found) is not a new one - it's referenced in readme.packages among other locations. I've placed a copy of sh.exe in C:\bin, but it

Hello, happy new year to all. The new Year gave me a new computer with Win98 and promptly I installed R on it. I created a directory R, with subdirectories gcc, perl,bin, helpwrk and rw1040. I got rw1040 from BDR place, the rest were gotter from the internet in the last week of the year, i.e. they are the newest versions. I modified autoexec.bat to put gcc\bin, perl\bin, etc in the path. I put the

..many thanks, no it works until he didn't find hhc.exe (..but the path to HTMLHelpWorkshop should be correct !?) and break down, ok i just check all path settings and find a solution for my self ! But here another problem: ...modify without the MkRules in my current binary R.1.6.1\src\gnuwin32 C:\DataMining\rw1061\library>Rcmd INSTALL Oop make: Entering directory

I observed the same behavior as Edward this morning with R-1.2.2 on Red Hat Linux 7.0. There appears to be a problem with the library in netpbm-9.5-5, even after hunting down the missing shhopt.h that pbmplus.h is looking for. I worked around the problem by compiling e1071 without pbm support (moved pbm.c to pbm.c.bak). Matt Matthew R. Nelson, Ph.D. Director, Information Systems Esperion

I observed the same behavior as Edward this morning with R-1.2.2 on Red Hat Linux 7.0. There appears to be a problem with the library in netpbm-9.5-5, even after hunting down the missing shhopt.h that pbmplus.h is looking for. I worked around the problem by compiling e1071 without pbm support (moved pbm.c to pbm.c.bak). Matt Matthew R. Nelson, Ph.D. Director, Information Systems Esperion

I am not sure if this is already known. I checked BUGS and found some references to similar behavior when R_HOME is set? Here is what I get: (1) Fresh build of 1.4.1 under Linux/RH-7.2. All tests pass. > R CMD check ctest * checking for working latex ... OK * using log directory `/home/andyj/stat/R-1.4.1/library/ctest.Rcheck' Installing *source* package `ctest'

Hi! My Platform: Windows 2000 R version >1.5 Active Perl 5.6 Rtools at the start of the PATH. I am trying to build a package. The functions are written only in R (no C code). I am trying to proceed as described in R-exts.pdf But instead to get a build package i am getting (see below). At first glance it seems to me that i understand the output. At something similar i am looking most of my

Full_Name: David Firth Version: 1.2.3 OS: Windows NT4 Submission from: (NULL) (163.1.103.107) Two things that look like bugs: 1. RCMD INSTALL does not seem to cope with quotes in the Windows path. With my path set to PATH="C:\R\Rtools";"C:\R\rw1023\bin";C:\Perl\bin\;C:\WINNT\system32;C:\WINNT;C:\ mingw2\bin I get, for example, C:\R>RCMD INSTALL boot RCMD

The trailing backslash seems to be the culprit. After looking at rw1020\bin\INSTALL, I noticed $pkg =~ s/\/$//; which I assume is to strip the trailing slash from a package name before obtaining the package name and cleaning out its directory I added $pkg =~ s/\\$//; immediately after this line. Rerunning Rcmd INSTALL C:\rw1011\src\library\bqtl\ with that change seemed to work OK

4. Noticed this error go by in the r-devel build in Windows 95: F:/R/TOOLS/BIN/MAKE.EXE -C ./help RHOME=F:/R/R-devel PKGDIR=F:/R/R-devel/src/lib rary RLIB=F:/R/R-devel/library ziponlyhelp-base MAKE.EXE[2]: Entering directory `F:/R/R-devel/src/gnuwin32/help' zip -jqmX help/Rhelp help/* -x help/AnIndex 2> nul MAKE.EXE[3]: [zipup] Error 12 (ignored) My guess is that it's a problem with

Dear helpers! I tried to build a package I have written, and which I have built up to now under R v .1.2.1, Windows 98 and Cygwin. The plan is to migrate to R v 1.5.0, Windows 2000 and Rtools. The package fits the decription of Simple Ports in readme.packages. The command C:\R\rw1050\src\gnuwin32>make pkg-sag gave the output make DLLNM= \ -C C:/R/rw1050/src/library/sag PKG=sag

I want to share this knowledge to save you the hours it took me to troubleshoot the following problem. Maybe you already know about this, but I certainly didn't. In the end I also suggest an update fix. PROBLEM: I have been building packages for a couple of months now without any problems. However, a couple of hours ago I ran into a strange problem, which I first couldn't understand or

Dear R-help, (This is the first time I tried compiling 1.6.x from source on WinNT4, so please bear with me...) I tried following the instruction in src/gnuwin32/INSTALL as closely as I could, but encounter the following errors at the make step. Can some kind soul please help? I'm using pcre-3.9 and MinGW-2.0.0-3. I checked and libRpcre.a _is_ in src/extra/pcre. TIA, Andy ... gcc

>I a missing in R extension a more detailed step by step description with=20 >some examples. >To whom send proposals of change to the R extensions manual. I am also=20 >willing to writ one step by step descriptin for windows 2000. If you would like to draft a detailed section on building packages under windows, that would be great. Once you're done, please email it to me (as

Hello! I tried to build the RMySQL package by myself on an NT machine. I have a NT machine with cygwin compiler and activestate Perl installed on it. Perl details: D:\prog\rw1011\src\gnuwin32>perl -v This is perl, version 5.005_03 built for MSWin32-x86-object (with 1 registered patch, see perl -V for more detail) Copyright 1987-1999, Larry Wall Binary build 522 provided by ActiveState Tool

Hi, I am trying to compile a package i am developing on my home computer where i am running Windows NT 4.0 service pack 6. Everything works fine on my other computer at work running windows 98. I am using the cygwin 1.1 bash shell and tolls with the MingW compiler, version 2.95.2-1 installed seperately and first in the path. I have active perl 6.16 as well. Here is the error I get when i run

Hello, I am trying to build an R package for Windows from an R package for Linux. There seems to be a "windres" problem somewhere. Here's what I get. C:\Temp>Rcmd build --binary VGAM > mm make[2]: windres: Command not found make[2]: *** [VGAM_res.o] Error 127 make[1]: *** [src/VGAM.dll] Error 2 make: *** [pkg-VGAM] Error 2 In more details, it goes like the following.

Prof Brian Ripley wrote [22.01.2001]: > The code is new in 1.2.1, and identical code in Rcmd clean works. It looks > like a Windows Perl bug. On my machines under NT4 and 98 this works > because -d foo seems to be always true! It is probably different on 2000. > > Try changing the line > > unless (-d $TMPDIR); > > to > unless (-e $tmpdir); >