[Rcpp-devel] RNGScope

[Rodney Sparapani]

This is my first post so please be gentle ;o)  I have been through
all of the docs that I can find (but I haven't looked at the source
code directly).  Here is the behavior that I see:

 > require(inline)
Loading required package: inline
 > source("helper.R")
 > source("corrMH.R")
 >
 > set.seed(66)
 >
 > for(i in 1:3) print(corrMH(0.2, 0.5, 100, 0))
[1] -0.07071726
[1] -8.085111
[1] -4.004675
 >
 > corrMH.Rcpp <- cxxfunction(
+  signature(a="numeric", b="numeric", N="numeric", x="numeric"),
+  body=cat.file("corrMH.Rcpp.cpp"), plugin="Rcpp")
 >
 > set.seed(66)
 >
 > for(i in 1:3) print(corrMH.Rcpp(0.2, 0.5, 100, 0))
[1] -0.07071726
[1] -0.07071726
[1] -0.07071726

Notice that the Rcpp version of the function returns the same value
each time (only 3 shown, but this is true no matter how many times
that you do it).  I guess this is because we follow the RNGScope state
of the RNG, but we don't update it.  However, updating is really needed
in this case.  I hope I am making this clear.  Please let me know if
you have any ideas.  Thanks

Attachments seem to be allowed so I'm including my code...

Red Hat Enterprise Linux Server release 6.2 (Santiago)
g++ (GCC) 4.4.6 20110731 (Red Hat 4.4.6-3)
R version 2.14.2 (2012-02-29)

 > installed.packages()["Rcpp", ]
                      Package                      LibPath
                       "Rcpp" "/opt/local/lib64/R/library"
                      Version                     Priority
                     "0.9.10"                           NA
                      Depends                      Imports
     "R (>= 2.12.0), methods"                           NA
                    LinkingTo                     Suggests
                           NA  "RUnit, inline, rbenchmark"
                     Enhances                      OS_type
                           NA                           NA
                      License                        Built
                 "GPL (>= 2)"                     "2.14.2"
-- 
Rodney Sparapani, PhD  Center for Patient Care and Outcomes Research
Sr. Biostatistician               http://www.mcw.edu/pcor
4 wheels good, 2 wheels better!   Medical College of Wisconsin (MCW)
WWLD?:  What Would Lombardi Do?   Milwaukee, WI, USA
-------------- next part --------------
A non-text attachment was scrubbed...
Name: corrMH.Rcpp.cpp
Type: text/x-c++src
Size: 372 bytes
Desc: not available
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20120814/980323f4/attachment.cpp>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: corrMH.R
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20120814/980323f4/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: helper.R
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20120814/980323f4/attachment-0001.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: corrMH.Rcpp.test.R
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20120814/980323f4/attachment-0002.ksh>