R devel - Dec 2009 - different behaviour of NAs under valgrind (PR#14171)

Full_Name: Martin Schlather
Version: 2.10.0
OS: linux
Submission from: (NULL) (91.3.220.231)


 
Bug summary:
    some functions behave differently for NAs when
    valgrind is used in R, e.g. sum and prod


Bug demonstration:
---  without  valgrind ----
>  sum(c(NA,1)[1])
[1] NA

---  with  valgrind ----
> sum(c(NA,1)[1])
[1] NaN


R call with valgrind:
R -d "valgrind --tool=memcheck --memcheck:leak-check=yes --num-callers=20
"

using
valgrind-3.3.0


platform = i686-pc-linux-gnu
 arch = i686
 os = linux-gnu
 system = i686, linux-gnu
 status  major = 2
 minor = 10.0
 year = 2009
 month = 10
 day = 26
 svn rev = 50208
 language = R
 version.string = R version 2.10.0 (2009-10-26)

Locale:
LC_CTYPE=en_GB.UTF-8;LC_NUMERIC=C;LC_TIME=en_GB.UTF-8;LC_COLLATE=en_GB.UTF-8;LC_MONETARY=C;LC_MESSAGES=en_GB.UTF-8;LC_PAPER=en_GB.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_GB.UTF-8;LC_IDENTIFICATION=C

Search Path:
 .GlobalEnv, package:stats, package:graphics, package:grDevices, package:utils,
package:datasets, package:methods, Autoloads, package:base

On Tue, 29 Dec 2009, schlather at math.uni-goettingen.de wrote:
> Full_Name: Martin Schlather
> Version: 2.10.0
Not current ....
> OS: linux
Not really specific enough to be helpful: these things depend on the 
compiler and libc.
> Submission from: (NULL) (91.3.220.231)
>
> Bug summary:
>    some functions behave differently for NAs when
>    valgrind is used in R, e.g. sum and prod
Why is that something to be reported to R-bugs?

For me (with the current valgrind 3.5.0 on i686 Fedora 12) exactly the 
same compiled code gives different results when run natively and under 
valgrind.  That looks very like a bug in valgrind's runtime, and it is 
incomprehensible as to why you assign blame to R.

>
> Bug demonstration:
> ---  without  valgrind ----
>>  sum(c(NA,1)[1])
> [1] NA
>
> ---  with  valgrind ----
>> sum(c(NA,1)[1])
> [1] NaN
>
>
> R call with valgrind:
> R -d "valgrind --tool=memcheck --memcheck:leak-check=yes
--num-callers=20 "
>
> using
> valgrind-3.3.0
>
>
> platform = i686-pc-linux-gnu
> arch = i686
> os = linux-gnu
> system = i686, linux-gnu
> status > major = 2
> minor = 10.0
> year = 2009
> month = 10
> day = 26
> svn rev = 50208
> language = R
> version.string = R version 2.10.0 (2009-10-26)
>
> Locale:
>
LC_CTYPE=en_GB.UTF-8;LC_NUMERIC=C;LC_TIME=en_GB.UTF-8;LC_COLLATE=en_GB.UTF-8;LC_MONETARY=C;LC_MESSAGES=en_GB.UTF-8;LC_PAPER=en_GB.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_GB.UTF-8;LC_IDENTIFICATION=C
>
> Search Path:
> .GlobalEnv, package:stats, package:graphics, package:grDevices,
package:utils,
> package:datasets, package:methods, Autoloads, package:base
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
-- 
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