On 30 August 2010 at 07:25, Dirk Eddelbuettel wrote: | Having discussed this some more with Romain who has no strong views on the | matter, I am inclined to the remove the patch that added the ability for | compilation with MSVC as I cannnot come up with answers to these questions: | | i) Why would it make sense to have this? Honestly, what is it good for to | compile Rcpp _in isolation_ when one cannot load the resulting object | code in R ? | | ii) The new files lack proper headers, credits, copyrights. Our code plays | by the rules in terms of credits, copyrights and licensing. I see no | reason to lower our standards and risk getting into trouble when as | per i) there is no reason or upside anyway. | | iii) I fear it attracts the wrong crowd of Windows users with little | knowledge about R, and little C/C++ understanding outside their | cherished IDE. If people want something for Visual Whatever, I just | learned from Bryan Lewis the other day that he is working on an | Rserve-on-Windows improvement -- see http://illposed.net/rserve.html | and in particular the last paragraph. That is a better route as it may | actually work with dotWhatever etc. | | But before I remove the patch I would like to hear from potential users | (ideally: others than just the patch submitters) as to why this would be a | bad idea.
As nobody came forward with a well-reasoned argument as to may we should have a patch that enables a compiler nobody can use, I have now removed it. For what it is worth, the contributed files remain in the SVN repo in a new subdirectory deprecated/msvc/ Those who care deeply enough can alwaus patch source locally. The affected SVN revisions are 2065 to 2067. Cheers, Dirk -- 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
