[Phylobase-devl] Rcpp and OS X compiliation

[Hilmar Lapp]

So what does file or otool say when you run it on the archive? Does it  
confirm that it is the expected and desired architecture?

	-hilmar

On Dec 15, 2009, at 4:37 PM, Dirk Eddelbuettel wrote:

>
> On 15 December 2009 at 15:54, Hilmar Lapp wrote:
> | On Dec 15, 2009, at 3:26 PM, Brian O'Meara wrote:
> |
> | > PKG_LIBS=/Library/Frameworks/R.framework/Resources/library/Rcpp/ 
> lib/
> | > i386/libRcpp.a: No such file or directory"
> | > [...]
> | > even though /Library/Frameworks/R.framework/Resources/library/ 
> Rcpp/
> | > lib/i386/libRcpp.a clearly exists on my system:
> |
> |
> | Is it a given that libRcpp.a is of the right architecture, and has
> | been compiled with a compatible compiler version?
> |
> | I've sometimes got these things when it turned out after some  
> digging
> | that that was not the case (even though the library itself was  
> there).
>
> Good hunch.  As Brian and I check by looking at his 'R CMD INSTALL  
> pkg' logs
> for Rcpp, it does build three times (i386, ppc, x86_64) but each  
> time we do
>
> userLibrary: 	$(USERLIB) $(USERLIBST)
> 		- at if test ! -e $(USERDIR)$(R_ARCH); then mkdir -p $(USERDIR)$ 
> (R_ARCH); fi
> 		cp $(USERLIB) $(USERDIR)$(R_ARCH)
> 		cp Rcpp*.h $(USERDIR)$(R_ARCH)
> 		cp $(USERLIBST) $(USERDIR)$(R_ARCH)
> 		rm $(USERLIB) $(USERLIBST)
>
> USERLIB is the dynamic library, USERLIBST is the static library.  
> This should
> rotate over the R_ARCH vales and copy appropriately.
>
> If someone has good OS X and wants to help Brian and myself...  I am  
> at wit's
> end (but I no close to nada about OS X).
>
> Dirk
>
>
> -- 
> Three out of two people have difficulties with fractions.

-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:- hlapp at duke dot edu :
===========================================================