[Rcpp-devel] Problem compling rinside_sample0 on OSX

Romain Francois romain at r-enthusiasts.com
Tue Apr 6 18:45:38 CEST 2010

Le 06/04/10 18:02, Dirk Eddelbuettel a écrit :
> On 6 April 2010 at 09:47, Dan Kulp wrote:
> | Success!   The snag was installing both packages from their source packages,
> | not the binary OSX ones.
> Fiddlesticks! I once knew that, but forgot, and it doesn't hit Romain on OS X
> as he builds from source anyway.

I'd guess that the RInside binary was built against Rcpp 0.7.10 which 
has some binary incompatibilities with Rcpp 0.7.11, or the other way 
around ..?.?

(Rcpp::Named as been promoted from a class to a function so that might 
be it. The interface has been retained, so I think this is why compiling 
both from source works. )

We try out best not to introduce binary incompatibilities like this one, 
but it is not an easy task to track them. thanks for reporting this.

I know that on windows, if package B that depends on package A and A is 
resubmitted, Uwe rebuilds both A and B for the windows binaries, not 
sure this also is the case for OSX binaries. (Beyond C incompatibilities 
other things can break such as S4 classes, etc ...)

> | I appreciate your guys time, too bad installation from CRAN didn't "just
> | work".  For me there is a lot of potential for this library.  I use R for
> Maybe "using source from CRAN" -- how/where do I document that better? Where
> would you have looked?

Maybe we can couple more tightly, e.g have RInside depend on Rcpp (>= 
0.7.11). Dunno.

Not sure this would have happened with dynamic linking, but this might 
be completely unrelated.

> | all my data analysis, but I have a set of c++ libraries used for generating
> | the data (my application is protein design), it would be great to fuse the
> | two.
> Yup. And given how feature-complete Rcpp has gotten it is quite a joy to use too.
> | One question, does RInside allow you to load installed packages (
> | library("RSvgDevice") ) ?
> "Should". See the other examples which load libraries.  Devices have their
> own pecularities so I don't know off-hand about this for SVG generation. In
> principle it should and if it doesn't in practice we can now count on you to
> debug it ;-)
> Good it works now.
> Dirk

Same analysis, it should work.

Romain Francois
Professional R Enthusiast
+33(0) 6 28 91 30 30
|- http://bit.ly/9aKDM9 : embed images in Rd documents
|- http://tr.im/OIXN : raster images and RImageJ
|- http://tr.im/OcQe : Rcpp 0.7.7

More information about the Rcpp-devel mailing list