On 26 November 2012 at 22:29, Theodore Lytras wrote: | Στις Δευ 26 Νοε 2012, ο/η Dirk Eddelbuettel έγραψε: | > I had this working years ago at a prior workplace (and would not have | > published / pushed it otherwise). | > | > I think I did it the other way around: rebuild Qt with the mingw from | > Rtools. Not entirely sure. | | Then I will try that (after uninstalling everything Qt-related), and post the | results.
Hang in there. It would be nice to get this going, and document it. It is, as so many things, frightfully tedious in Windows but I think it can be done. | Might I conceivably need to compile R for windows from source? | That would be dreadful, but I would do it if necessary. That is easier and quicker than to rebuild Qt... But no, you can trust the binary from CRAN with respect to the Rtools toolchain. But you do have to have a compatible Qt. And when I did this, I think they did not yet offer the "all-in" option but only source (Qt 4.1 or 4.2 ?). So I had to anyway. It may of course also work with the Qt binary you got. Sometimes it is just link order etc pp. But hard to tell. Good luck, and keep trying. Dirk | Do you recall anything like that?? | | > Or you could | > forget it and switch to Rserve and tcp/ip connections to avoid the linker | > issue. | | Oh bummer, that would really not be an option for me.... | | Thanks, | | Theodore -- 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
