[Rd] Runtime error when run a RInside program compiled by intel c++ on windows
Dirk Eddelbuettel
edd at debian.org
Fri Dec 11 14:26:42 CET 2015
Eric,
This must now be your third or fourth email on this matter (counting the one
to me directly, and to be clear I much prefer this on list). I'll be brief:
On 11 December 2015 at 20:50, 李琥 wrote:
|
| I have intstalled R-3.2.2,Rcpp-0.12.2,RInside-0.2.13 on windows. I compiled the example1 of RInside in the
| example directory which create a RInside instance and prints "hello world". The compiler is intel c++. I included
| the RInside source files in the project and fixed the compile and link errors. But when run the executable file it
| prints some error message and quites.
| I debug into the code.The error messages come from the RInside::initialize() function. It gives following message after
| Rf_initEmbeddedR():
|
| Warning message:
| In local(( :bytecode version mismatch: using eval
| Error in objects(db.pos, all.name = TRUE) :
| 3 arguments passed to .Internal(ls) which requres 2
|
| And it gives following error after Rf_eval(Rf_lang2(suppressMessagesSymbol,Rf_lang2(requireSymbol, Rf_mkString("Rcpp"))),R_GlobalEnv)
|
| Error in .Internal(getOption(x)):
| there is no .Internal function 'getOption'
| Execution halted.
|
| Does anyone know how to fix this error? Thanks.
This is rather indicative of an unsuccessful build.
You may have managed to link this in some, but R clearly does not instantiate
right and does not start right. getOption() is part of the base package. If
that goes astray you have /a lot more/ problems to expect.
So if I were you, I'd reverse course and forget icc. R, Rcpp, RInside are all
tested against a well known set of compilers __and work with those__. So on
Windows I would stick with MinGW.
Lastly, as the Rcpp + RInside documentation says in several places: please
ask on the rcpp-devel list.
R-devel is not the correct forum for questions with particular packages.
Thanks, Dirk
--
http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org
More information about the R-devel
mailing list