Hi Daniel,
Unfortunately, you cannot use // [[Rcpp::export]] to automagically export
template functions.
The general alternative approach is to define a template implementation,
and then dynamically dispatch to that implementation, e.g.
template
void doStuffImpl() { ... }
with dispatch a
Hello,
I'm trying to build a function that calculates (for example) the log
of all elements of a container. I want this container to be a vector,
a Rcpp::numericVector , or perhaps a arama::colvec , so I'm trying
with templates ,and it compiles without the // [[Rcpp::export]] part,
but when I put
If I have data like
list(list('a', 1, T), list('b', 2, F), ...)
(list of lists with same types)
and I want to convert to
list(c('a', 'b'), c(1, 2), c(T, F))
(list of vectors)
where the types are not known in advance (its trivial if you know the
sequence of types in advance).
Anyone know how
The problem is solved, as you explained I had to update the ldconf cache.
I guess that I didn't get any issues with the direct compilation of the example
because I was building an executable file and not a shared library.
Thanks a lot for taking the time to explain the key aspects.
- Mail
On 1 May 2015 at 19:03, Marie-Pierre ETIENNE wrote:
| So I tried to apply your suggestion and I still have the same error
|
| Error in dyn.load(file, DLLpath = DLLpath, ...) : unable to load shared
object
'/home/metienne/R/x86_64-pc-linux-gnu-library/3.1/testCMaes/libs/testCMaes.so':
libcmae
Oups I put my answer to your comments in the body of the mail and therefore it
is almost invisible.
So I tried to apply your suggestion and I still have the same error
Error in dyn.load(file, DLLpath = DLLpath, ...) : unable to load shared object
'/home/metienne/R/x86_64-pc-linux-gnu-librar
Thanks Dirk for your advice
- Mail original -
De: "Dirk Eddelbuettel"
À: "Marie-Pierre ETIENNE"
Cc: [email protected]
Envoyé: Vendredi 1 Mai 2015 17:33:11
Objet: Re: [Rcpp-devel] Linking Rcpp and libcmaes within a package
On 1 May 2015 at 16:12, Marie-Pierre
On 1 May 2015 at 16:12, Marie-Pierre ETIENNE wrote:
| CXX_STD = CXX11
| PKG_CPPFLAGS=-I/usr/local/include/libcmaes -I../inst/include
-I/usr/local/include/eigen3/
If you add LinkingTo: RcppEigen then you don't need the explicit include
for Eigen3, and are more portable. Of course, you would
Dear Rcpp and R developpers,
I have recently used Rcpp to build a package and found it very well documented
and easy as I don't need to go very deep in Writing R extensions mechanisms.
My code uses an optimisation algorithm named cmaes, and this use was first made
through a cmaes R package. As
