[R-SIG-Mac] R and Valgrind hangs on Snow Leopard

[Prof Brian Ripley]

Erika tracked this down to a known issue with valgrind and (C-level) 
system calls.  See https://bugs.kde.org/show_bug.cgi?id=301281

I've added a workaround in R-patched/R-devel: set the environment 
variable R_OSX_VALGRIND (to anything), preferably before starting the 
session but I guess it would also work from the R session.  This 
circumvents the system call used to remove the session temporary 
directory at the end of the session.



On 13/08/2012 13:14, Prof Brian Ripley wrote:
> On Mon, 13 Aug 2012, Cule, Erika wrote:
>
>> (Reposted in plain text, with apologies)
>>
>> I am having the problem described in this post:
>>
>> https://stat.ethz.ch/pipermail/r-sig-mac/2011-November/008848.html
>>
>> In summary, running Valgrind with R hangs.
>>
>> As in the linked case above, if a file test.R contains the single
>> command print(seq(10))
>>
>> Then running the command
>>
>> R -d "valgrind --tool=memcheck --leak-check=full" --vanilla < test.R
>>
>> Produces the following output and hangs
>>
>> ==78868== Memcheck, a memory error detector
>> ==78868== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.
>> ==78868== Using Valgrind-3.9.0.SVN and LibVEX; rerun with -h for
>> copyright info
>> ==78868== Command:
>> /Library/Frameworks/R.framework/Resources/bin/exec/x86_64/R --vanilla
>> ==78868==
>> ==78868== Warning: ignored attempt to set SIGUSR2 handler in sigaction();
>> ==78868==          the SIGUSR2 signal is used internally by Valgrind
>>
>> R version 2.15.1 (2012-06-22) -- "Roasted Marshmallows"
>> Copyright (C) 2012 The R Foundation for Statistical Computing
>> ISBN 3-900051-07-0
>> Platform: x86_64-apple-darwin10.8.0/x86_64 (64-bit)
>>
>> R is free software and comes with ABSOLUTELY NO WARRANTY.
>> You are welcome to redistribute it under certain conditions.
>> Type 'license()' or 'licence()' for distribution details.
>>
>>  Natural language support but running in an English locale
>>
>> R is a collaborative project with many contributors.
>> Type 'contributors()' for more information and
>> 'citation()' on how to cite R or R packages in publications.
>>
>> Type 'demo()' for some demos, 'help()' for on-line help, or
>> 'help.start()' for an HTML browser interface to help.
>> Type 'q()' to quit R.
>>
>>>
>>>
>>>
>>> print(seq(10))
>> [1]  1  2  3  4  5  6  7  8  9 10
>>>
>>>
>>>
>>>
>>>
>>>
>>
>> I have followed the advice previously posted by Simon, used the SVN
>> build of valgrind and recompiled R from source with
>> --with-valgrind-instrumentation=2
>>
>> I am running Mac OS X 10.6.8, valgrind-3.9.0.SVN, R version 2.15.1 and
>> Platform: x86_64-apple-darwin10.8.0/x86_64 (64-bit).
>>
>> As was the case for the previous poster, a command such as
>>
>> valgrind ls
>>
>> works as expected.
>>
>> I noted that Simon said he had no problem using valgrind from SVN on
>> both Lion and Snow Leopard. Does anyone have any suggestion of what I
>> might be doing differently, or what I could do to fix this problem?
>
>
> Not multiple post ... this is a second copy.
>
> It does not work for me on Lion (I've never tried Snow Leopard).  As
> valgrind does not yet (even in the recent 3.8.0) really support Mountain
> Lion and we will be upgraded shortly (and BTW, Snow Leopard is now
> end-of-life for security patches), I've pretty much given up on valgrind
> on OS X.
>


-- 
Brian D. Ripley,                  ripley at 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