Our current intention is to no longer include Rversion.h from R.h as from R 1.6.0 (which is months away). Package maintainers when revising code might like to check if they actually use Rversion.h, and if so include it explicitly. I did a quick grep through the CRAN packages, and in all but one case (mda) the test is for R < 1.2.0, so it may be simplest just to tidy up and add `Depends: R (>= 1.2.0)' to the DESCRIPTION file. -- Brian D. Ripley, ripley@stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
ripley@stats.ox.ac.uk writes:> Our current intention is to no longer include Rversion.h from R.h as from > R 1.6.0 (which is months away).For the inquisitive minds: The reason we want to do this lies in makefile dependencies: Every file that includes R.h currently ends up depending on the date stamp which changes automatically every 24 hours. So even a simple bug fix involves substantial rebuilding activity, which is annoying for the developers, and also carries some risk of compiler warning getting missed. -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
>>>>> Peter Dalgaard BSA writes:> ripley@stats.ox.ac.uk writes: >> Our current intention is to no longer include Rversion.h from R.h as from >> R 1.6.0 (which is months away).> For the inquisitive minds: The reason we want to do this lies in > makefile dependencies: Every file that includes R.h currently ends up > depending on the date stamp which changes automatically every 24 > hours. So even a simple bug fix involves substantial rebuilding > activity, which is annoying for the developers, and also carries some > risk of compiler warning getting missed.Note that this implies that packages which unconditionally use R_VERSION or R_Version (note that R-exts in facts recommends #if defined(R_VERSION) && R_VERSION >= R_Version(0, 99, 0) will fail to install under the current r-devel. These packages seems to be Matrix RArcInfo RODBC XML chron gafit mda odesolve (two of which I happen to maintain myself ...) as well as the Omegahat packages CORBA RGnumeric RSPerl RSPython SJava SNetscape SXalan Slcc Sxslt It would be great if we could have updates of these packages in the near future, either via following the recommended procedure from R-exts or, perhaps preferably, removing the pre-1.2.0 compatibility code and adding Depends: R (>= 1.2.0) in the package DESCRIPTION file. Thanks -k -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._