Skip to content

build fails to build help for nlme

5 messages · Brian Ripley, Peter Dalgaard, Seth Falcon

#
Hi all,

I'm trying to build R-1.8.1 from source on Linux and getting the
following error when the makefile gets to the step of building the help
for 'nlme':

<snip>
ranef.lme                         text    html    latex   example
reStruct                          text    html    latex   example
/home/sfalcon/sw/R-related/R-1.8.1/bin/INSTALL: line 1:  8133 Segmentation fault      ${R_CMD} perl "${R_HOME}/share/perl/build-help.pl" ${build_help_opts} "${pkg_dir}" "${lib}" "${R_PACKAGE_DIR}" "${pkg_name}"
ERROR: building help failed for package 'nlme'
** Removing '/home/sfalcon/sw/R-related/R-1.8.1/library/nlme'
make[2]: *** [nlme.ts] Error 1
make[2]: Leaving directory
`/home/sfalcon/sw/R-related/R-1.8.1/src/library/Recommended'
make[1]: *** [recommended-packages] Error 2
make[1]: Leaving directory
`/home/sfalcon/sw/R-related/R-1.8.1/src/library/Recommended'
make: *** [stamp-recommended] Error 2
</snip>

Has anyone else encountered this?

I've had no difficulties compiling previous versions of R on this
machine.


Thanks,

+ seth
#
I have once seen this when a disc became full.
On Thu, 29 Jan 2004, Seth Falcon wrote:

            

  
    
#
Thanks for the response.  Disk space was not the issue (over 90G avail).
However, today I tried a "make clean && make" and everything went fine.
Wish I had an explanation, but I'll settle for the clean build ;-)
On Thu, Jan 29, 2004 at 08:18:41PM +0000, Prof Brian Ripley wrote:
#
Prof Brian Ripley <ripley at stats.ox.ac.uk> writes:
Flaky RAM chips and overheating might do it too. Also, running the
system near memory full condition (a runaway Mozilla or Java process
perchance?). Is the crash point reproducible?

  
    
#
On Thu, Jan 29, 2004 at 10:43:22PM +0100, Peter Dalgaard wrote:
Per my previous reply, I waited a day and tried "make clean && make" and
got a clean build.

Yesterday, I tried repeating the make command (I didn't clean first),
and I did get the crash in what looked to be the same point when
building the help for the nlme package.

Given this, flaky RAM or high mem load seem more likely.  If I run into
it again I will be sure to make note of the current system load and
such.

Thanks,

+ seth