Full_Name: Jeremy Fox Version: 1.5.0 OS: Solaris 8, 64 bit Submission from: (NULL) (171.64.232.63) I am compiling R under Solaris and the compilation stops at this point. The problem only happens when I link in readline. The R binary is actually compiled and readlne works. The main problem with the compiled binary (other than that all the supporting files were not correctly installed) is that it prints out the error Warning message: There is no package called `ctest' in: library(package, char = TRUE, logical TRUE, warn.conflicts = warn.conflicts, the first time I try to do something. It generally prints out only once, but may be the cause of the problem with the break in the make process I am worried about. I am using 64 bit Solaris, with the 64 bit with Sun Forte flags as advocated in Appendix B. 7.2 of the Installation Guide. I compiled readline with the same flags. For definiteness, they are CC="cc -xarch=v9" CFLAGS="-xO5 -xlibmil -dalign" F77="f95 -xarch=v9" FFLAGS="-xO5 -xlibmil -dalign" CXX=CC CXXFLAGS="-xO5 -xlibmil -dalign -xarch=v9" Without readline, the make command exectutes fine. Here is the actual bug mkdir -p -- ../../../../library/methods/libs make[5]: Leaving directory `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' make[4]: Leaving directory `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' make[4]: Entering directory `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods' dumping R code in package 'methods' Warning message: There is no package called `ctest' in: library(package, char = TRUE, logical TRUE, warn.conflicts = warn.conflicts, Error in file == "" : comparison (1) is possible only for atomic types Execution halted make[4]: *** [../../../library/methods/R/all.rda] Error 1 The file all.rda is not created. This may be because of this unimportant error message about ctest that somehow involves the readline package. I am using Readline 4.2a. Thanks for any help, Jeremy -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
1) Was readline compiled under those same flags? Was it compiled under 64-bit mode at all? (The manual does warn about that.) 2) You should be able to get ctest built by (cd src/library/ctest; make) but that should have happened anyway. 3) The actual error seems to be in interpreting the methods code. That has nothing to do with readline (it is run in batch mode) so there must be another difference. It all points to problems with your readline library. Are you sure that there are not earlier error messages you have overlooked/ignored? On Sat, 15 Jun 2002 jerfox@stanford.edu wrote:> Full_Name: Jeremy Fox > Version: 1.5.0 > OS: Solaris 8, 64 bit > Submission from: (NULL) (171.64.232.63) > > > > I am compiling R under Solaris and the compilation stops at this point. The > problem only happens when I link in readline. The R binary is actually compiled > and readlne works. > The main problem with the compiled binary (other than that all the supporting > files were not > correctly installed) is that it prints out the error > > Warning message: > There is no package called `ctest' in: library(package, char = TRUE, logical > TRUE, warn.conflicts = warn.conflicts, > > the first time I try to do something. It generally prints out only once, but may > be the > cause of the problem with the break in the make process I am worried about. > > I am using 64 bit Solaris, with the 64 bit with Sun Forte flags as advocated in > Appendix > B. 7.2 of the Installation Guide. I compiled readline with the same flags. For > definiteness, they are > > CC="cc -xarch=v9" > CFLAGS="-xO5 -xlibmil -dalign" > F77="f95 -xarch=v9" > FFLAGS="-xO5 -xlibmil -dalign" > CXX=CC > CXXFLAGS="-xO5 -xlibmil -dalign -xarch=v9" > > Without readline, the make command exectutes fine. Here is the actual bug > > mkdir -p -- ../../../../library/methods/libs > make[5]: Leaving directory > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' > make[4]: Leaving directory > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' > make[4]: Entering directory > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods' > dumping R code in package 'methods' > Warning message: > There is no package called `ctest' in: library(package, char = TRUE, logical > TRUE, warn.conflicts = warn.conflicts, > Error in file == "" : comparison (1) is possible only for atomic types > Execution halted > make[4]: *** [../../../library/methods/R/all.rda] Error 1 > > > The file all.rda is not created. This may be because of this unimportant error > message > about ctest that somehow involves the readline package. I am using Readline > 4.2a.-- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Thanks for the quick reply. On Sat, Jun 15, 2002 at 11:12:55AM +0100, ripley@stats.ox.ac.uk wrote:> 1) Was readline compiled under those same flags? Was it compiled > under 64-bit mode at all? (The manual does warn about that.)Which manual? But anyway, I double checked and the version of readline that I compiled uses the exact same flags. The configure script also does not recognize readline when I do not link in the directory where I compiled readline, so it is not a case of some other 32 bit readline library being read in.> 2) You should be able to get ctest built by > > (cd src/library/ctest; make) > > but that should have happened anyway.It compiled fine but R does not stop its error message.> 3) The actual error seems to be in interpreting the methods code. > That has nothing to do with readline (it is run in batch mode) so > there must be another difference. It all points to problems with your > readline library.OK.> Are you sure that there are not earlier error messages you have > overlooked/ignored?Here are the two warnings before the fatal errors. I checked the line numbers and I would think these files are not causing the problems I am experiencing. I am certain the X libraries and whatever is being read in from /usr/local/include are not 64 bit, but X11() seems to work fine anyway. cc -xarch=v9 -I. -I../../src/include -I../../src/include -I/usr/local/include -DHAVE_CONFIG_H -xO5 -xlibmil -dalign -c graphics.c -o graphics.o "graphics.c", line 5358: warning: initializer does not fit or is out of range: -1 cc -xarch=v9 -I. -I../../../src/include -I../../../src/include -I/usr/pubsw/X11R6/include -I/usr/local/include -DHAVE_CONFIG_H -KPIC -xO5 -xlibmil -dalign -c devX11.c -o .libs/devX11.o "devX11.c", line 1081: warning: statement not reached "devX11.c", line 1095: warning: statement not reached "devX11.c", line 1407: warning: statement not reached Thanks, Jeremy> On Sat, 15 Jun 2002 jerfox@stanford.edu wrote: > > > Full_Name: Jeremy Fox > > Version: 1.5.0 > > OS: Solaris 8, 64 bit > > Submission from: (NULL) (171.64.232.63) > > > > > > > > I am compiling R under Solaris and the compilation stops at this point. The > > problem only happens when I link in readline. The R binary is actually compiled > > and readlne works. > > The main problem with the compiled binary (other than that all the supporting > > files were not > > correctly installed) is that it prints out the error > > > > Warning message: > > There is no package called `ctest' in: library(package, char = TRUE, logical > > TRUE, warn.conflicts = warn.conflicts, > > > > the first time I try to do something. It generally prints out only once, but may > > be the > > cause of the problem with the break in the make process I am worried about. > > > > I am using 64 bit Solaris, with the 64 bit with Sun Forte flags as advocated in > > Appendix > > B. 7.2 of the Installation Guide. I compiled readline with the same flags. For > > definiteness, they are > > > > CC="cc -xarch=v9" > > CFLAGS="-xO5 -xlibmil -dalign" > > F77="f95 -xarch=v9" > > FFLAGS="-xO5 -xlibmil -dalign" > > CXX=CC > > CXXFLAGS="-xO5 -xlibmil -dalign -xarch=v9" > > > > Without readline, the make command exectutes fine. Here is the actual bug > > > > mkdir -p -- ../../../../library/methods/libs > > make[5]: Leaving directory > > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' > > make[4]: Leaving directory > > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods/src' > > make[4]: Entering directory > > `/afs/ir.stanford.edu/users/w/o/wolak/jerfox/RProj/R-1.5.0/src/library/methods' > > dumping R code in package 'methods' > > Warning message: > > There is no package called `ctest' in: library(package, char = TRUE, logical > > TRUE, warn.conflicts = warn.conflicts, > > Error in file == "" : comparison (1) is possible only for atomic types > > Execution halted > > make[4]: *** [../../../library/methods/R/all.rda] Error 1 > > > > > > The file all.rda is not created. This may be because of this unimportant error > > message > > about ctest that somehow involves the readline package. I am using Readline > > 4.2a. > > > -- > 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-- ------------------------ Jeremy T. Fox jerfox@stanford.edu -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._