Interesting. I don't think it has anything to do with inline or not (see
below). If in doubt, do use wrap() as the 'it dropped the dim' reveals a
fallback to vector defaults.
When I use your last solution (the function) in inline, it gets the same
result:
edd@max:~$ r /tmp/soren.r
Loading required package: methods
$X
[,1] [,2]
[1,] 1 3
[2,] 2 4
$Xinv
[,1] [,2]
[1,] -2 1.5
[2,] 1 -0.5
$X
[,1] [,2]
[1,] 1 3
[2,] 2 4
$Xinv
[,1] [,2]
[1,] -2 1.5
[2,] 1 -0.5
edd@max:~$
So I doubt we can blame inline. Code below.
Dirk
library(inline)
code <- '
arma::mat X = as<arma::mat>(X_);
arma::mat Xinv = inv( X );
return Rcpp::List::create(Rcpp::Named("X", X),Rcpp::Named("Xinv", Xinv));'
sol <- cxxfunction(signature(X_ = "matrix"), code, plugin = "RcppArmadillo",
verbose=F)
M <- matrix(c(1,2,3,4),nr=2)
print(sol(M))
inc <- '
RcppExport SEXP sol2 ( SEXP X_ ){
arma::mat X = as<arma::mat>(X_);
arma::mat Xinv = inv( X );
return Rcpp::List::create(Rcpp::Named("X", wrap(X)),Rcpp::Named("Xinv",
wrap(Xinv)));
}
'
sol3 <- cxxfunction(signature(X_ = "matrix"), plugin = "RcppArmadillo",
inc=inc, body='return sol2(X_);')
print(sol3(M))
--
Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com
_______________________________________________
Rcpp-devel mailing list
Rcpp-devel@lists.r-forge.r-project.org
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
