From jcarew@stat.wisc.edu Mon Feb 3 02:50:07 2003 From: jcarew@stat.wisc.edu (John Carew) Date: Sun, 2 Feb 2003 20:50:07 -0600 (CST) Subject: [R-SIG-Mac]Date conversion problems Message-ID: Hello, I have a data set with dates stored as a factor with levels coded as, for example "01/03/1996" I tried the following: onset <- strptime(as.character(cf$dx.dt), format="%m%d%Y") where cf$dx.dt are the dates in the data frame cf. For output, I get NA for every entry. If it matters, I'm running the fink version of R 1.6.1 on a powerbook G4 with OS 10.2.2. I have little experience dealing with dates in R. Is there a reference that discusses the general issues? Any help would be much appreciated. Thanks, John -- John Carew Department of Statistics University of Wisconsin--Madison http://www.stat.wisc.edu/~jcarew From macq@llnl.gov Mon Feb 3 15:10:21 2003 From: macq@llnl.gov (Don MacQueen) Date: Mon, 3 Feb 2003 07:10:21 -0800 Subject: [R-SIG-Mac]Date conversion problems In-Reply-To: References: Message-ID: Try format="%m/%d/%Y" The help pages are pretty good, and the examples in ?strptime show something quite similar to your case. Vol. 1/2 of R News (available at the R website) has the article that introduced the POSIXt classes. -Don >Hello, >I have a data set with dates stored as a factor with levels coded as, for >example "01/03/1996" > >I tried the following: > >onset <- strptime(as.character(cf$dx.dt), format="%m%d%Y") > >where cf$dx.dt are the dates in the data frame cf. For output, I get NA >for every entry. If it matters, I'm running the fink version of R 1.6.1 >on a powerbook G4 with OS 10.2.2. > >I have little experience dealing with dates in R. Is there a reference >that discusses the general issues? > >Any help would be much appreciated. >Thanks, >John > >-- >John Carew >Department of Statistics >University of Wisconsin--Madison >http://www.stat.wisc.edu/~jcarew > >_______________________________________________ >R-SIG-Mac mailing list >R-SIG-Mac@stat.math.ethz.ch >http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac -- -------------------------------------- Don MacQueen Environmental Protection Department Lawrence Livermore National Laboratory Livermore, CA, USA -------------------------------------- From tlumley@u.washington.edu Mon Feb 3 15:14:16 2003 From: tlumley@u.washington.edu (Thomas Lumley) Date: Mon, 3 Feb 2003 07:14:16 -0800 (PST) Subject: [R-SIG-Mac]Date conversion problems In-Reply-To: Message-ID: On Sun, 2 Feb 2003, John Carew wrote: > Hello, > I have a data set with dates stored as a factor with levels coded as, for > example "01/03/1996" > > I tried the following: > > onset <- strptime(as.character(cf$dx.dt), format="%m%d%Y") > If the dates look like "01/03/1996" then the format is "%m/%d/%Y" -thomas From joe@bio.umass.edu Tue Feb 4 14:07:14 2003 From: joe@bio.umass.edu (Joseph Kunkel) Date: Tue, 4 Feb 2003 09:07:14 -0500 Subject: [R-SIG-Mac]Memory leaks associated with R on Mac OsX Message-ID: I am disturbed by the phenomenon that after a large amount of R work on my Mac I become incommunicado via my Email client. Email sometimes does not connect to a functional and sometimes its window will not appear when prompted. Restarting the Mac fixes communication. It is only after I have done considerable R programming. I have a Ti80 with 512Mb memory running OsX 10.2.3 and having installed the Jan de Leeuw Jan 22 packages. It might be associated with R script failures (i.e. my errors during script development) particularly, although I have not examined this systematically. Have any other Mac R users experienced this phenomenon. Are there memory leak issues associated with R or R script failures? Joe ----------------------------------------- Joseph G. Kunkel, Professor Biology Department University of Massachusetts Amherst Amherst MA 01003 http://www.bio.umass.edu/biology/kunkel/ From deleeuw@stat.ucla.edu Tue Feb 4 19:29:00 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Tue, 4 Feb 2003 11:29:00 -0800 Subject: [R-SIG-Mac]Memory leaks associated with R on Mac OsX In-Reply-To: Message-ID: Run R with the Process Viewer application open, and watch the memory and CPU usage (and the number of R processes). --- Jan On Tuesday, February 4, 2003, at 06:07 AM, Joseph Kunkel wrote: > I am disturbed by the phenomenon that after a large amount of R work > on my Mac I become incommunicado via my Email client. Email sometimes > does not connect to a functional and sometimes its window will not > appear when prompted. Restarting the Mac fixes communication. It is > only after I have done considerable R programming. I have a Ti80 with > 512Mb memory running OsX 10.2.3 and having installed the Jan de Leeuw > Jan 22 packages. > > It might be associated with R script failures (i.e. my errors during > script development) particularly, although I have not examined this > systematically. Have any other Mac R users experienced this > phenomenon. Are there memory leak issues associated with R or R > script failures? > > Joe > ----------------------------------------- > Joseph G. Kunkel, Professor > Biology Department > University of Massachusetts Amherst > Amherst MA 01003 > http://www.bio.umass.edu/biology/kunkel/ > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > From rpeng@stat.ucla.edu Wed Feb 5 04:49:17 2003 From: rpeng@stat.ucla.edu (Roger Peng) Date: Tue, 4 Feb 2003 20:49:17 -0800 (PST) Subject: [R-SIG-Mac]Memory leaks associated with R on Mac OsX In-Reply-To: Message-ID: What version of R are you using? I believe there was a memory leak discovered in R 1.6.0. -roger _______________________________ UCLA Department of Statistics rpeng@stat.ucla.edu http://www.stat.ucla.edu/~rpeng On Tue, 4 Feb 2003, Joseph Kunkel wrote: > I am disturbed by the phenomenon that after a large amount of R work on > my Mac I become incommunicado via my Email client. Email sometimes > does not connect to a functional and sometimes its window will not > appear when prompted. Restarting the Mac fixes communication. It is > only after I have done considerable R programming. I have a Ti80 with > 512Mb memory running OsX 10.2.3 and having installed the Jan de Leeuw > Jan 22 packages. > > It might be associated with R script failures (i.e. my errors during > script development) particularly, although I have not examined this > systematically. Have any other Mac R users experienced this > phenomenon. Are there memory leak issues associated with R or R script > failures? > > Joe > ----------------------------------------- > Joseph G. Kunkel, Professor > Biology Department > University of Massachusetts Amherst > Amherst MA 01003 > http://www.bio.umass.edu/biology/kunkel/ > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > From macq@llnl.gov Fri Feb 7 15:41:05 2003 From: macq@llnl.gov (Don MacQueen) Date: Fri, 7 Feb 2003 07:41:05 -0800 Subject: [R-SIG-Mac]ROracle and OS X, yes Message-ID: Yesterday, with some help from David James, I managed to install the ROracle package. OS X 10.2.3. The server is Oracle 8i running on a Sun Solaris box. The client has Oracle 9i (files only, no database created, no Oracle processes running). In a single test with a very simple query, it works. > version _ platform powerpc-apple-darwin6.3 arch powerpc os darwin6.3 system powerpc, darwin6.3 status major 1 minor 6.2 year 2003 month 01 day 10 language R This version of R was installed from Jan's gifi, not from CRAN source code. There were a couple of strange things having to do with the Oracle distribution that I had to work around. -Don -- -------------------------------------- Don MacQueen Environmental Protection Department Lawrence Livermore National Laboratory Livermore, CA, USA -------------------------------------- From jago@mclink.it Fri Feb 7 18:12:02 2003 From: jago@mclink.it (Stefano Iacus) Date: Fri, 7 Feb 2003 19:12:02 +0100 (CET) Subject: [R-SIG-Mac]ROracle and OS X, yes Message-ID: <1.0.2.200302071911.92923@mclink.it> Don, it would be nice to have such details. stefano > > Yesterday, with some help from David James, I managed to install > the > ROracle package. OS X 10.2.3. > > The server is Oracle 8i running on a Sun Solaris box. > The client has Oracle 9i (files only, no database created, no > Oracle > processes running). > > In a single test with a very simple query, it works. > > > version > _ > platform powerpc-apple-darwin6.3 > arch powerpc > os darwin6.3 > system powerpc, darwin6.3 > status > major 1 > minor 6.2 > year 2003 > month 01 > day 10 > language R > > This version of R was installed from Jan's gifi, not from CRAN > source code. > > There were a couple of strange things having to do with the Oracle > distribution that I had to work around. > > -Don > > -- > -------------------------------------- > Don MacQueen > Environmental Protection Department > Lawrence Livermore National Laboratory > Livermore, CA, USA > -------------------------------------- > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac From macq@llnl.gov Fri Feb 7 20:06:49 2003 From: macq@llnl.gov (Don MacQueen) Date: Fri, 7 Feb 2003 12:06:49 -0800 Subject: [R-SIG-Mac]ROracle and OS X, more details In-Reply-To: <1.0.2.200302071911.92923@mclink.it> References: <1.0.2.200302071911.92923@mclink.it> Message-ID: Sefano, here is a better description of what I had to do to get ROracle to install. David, I posted a note on r-sig-mac that I had done this, and Stefano asked for more information. So here it is. -Don ------- ROracle ------ In the Makefile.in in ROracle/src, change PARSE = FULL to PARSE = PARTIAL This is a directive to proc, the Oracle ProC/C++ precompiler Thanks to David James, ROracle's maintainer (and one of the authors) for this information. ------- Oracle ------- I downloaded the "Oracle9i Database Developer's Release for Apple Mac OS X Version 10.2 (Jaguar)" from Oracle's "Oracle Technology Network." It's a little hard to find, have to look under "Oracle9i JDeveloper Download". Install according to instructions from Oracle. This puts it in /Users/oracle/9iR2/orahome. Create appropriate tnsnames.ora and listener.ora files in $ORACLE_HOME/network/admin. The first work around is this: cp $ORACLE_HOME/rdbms/lib/libclntst9.a $ORACLE_HOME/lib cd $ORACLE_HOME/lib ranlib -s libclntst9.a (look in the file $ORACLE_HOME/precomp/lib/env_precomp.mk for references to libclntst9.a) The absence of libclntst9.a in $ORACLE_HOME/lib may represent an error in the Oracle distribution, or it may represent a lack of understanding on my part. The second work around is truly puzzling. It may be due to something I did without realizing the side effect, so don't assume this work around will be necessary. I installed Oracle here: [662]% echo $ORACLE_HOME /Users/oracle/9iR2/orahome but for some unknown reason Oracle's proc behaved as if ORACLE_HOME is here: /Users/oracle/Source/9iR2 I worked around this by first creating the directory /Users/oracle/Source and then making a soft link: [663]% cd /Users/oracle/Source [664]% ls -l total 8 drwxr-xr-x 3 macq ora 102 Feb 6 16:16 ./ drwxrwxr-x 19 oracle ora 646 Feb 6 16:11 ../ lrwxr-xr-x 1 macq ora 15 Feb 6 16:16 9iR2@ -> ../9iR2/orahome ----- Environment variables ----- Then, before attempting to install ROracle, set some environment variables. Here is the script I use (csh syntax). The value of ORACLE_SID is probably wrong, but I haven't seen any problems from it yet. ----------------------------- # This is your ORACLE_HOME setenv S /Users/oracle/9iR2/orahome # Root of All Oracle installation. setenv ORACLE_HOME $S # Same as ORACLE_HOME setenv SRCHOME $S # Add Oracle binaries to path. setenv PATH $S/bin:$PATH # Oracle SID setenv ORACLE_SID ab # setenv CDEBUG "-g -O0" # Required to find dynamic libraries. if ( $?DYLD_LIBRARY_PATH ) then setenv DYLD_LIBRARY_PATH $S/lib:$S/rdbms/lib:$DYLD_LIBRARY_PATH else setenv DYLD_LIBRARY_PATH $S/lib:$S/rdbms/lib endif # Required for the database application. if ($?CLASSPATH) then setenv CLASSPATH $S/DBCreate/oradev/classes/orapts.jar:$CLASSPATH else setenv CLASSPATH $S/DBCreate/oradev/classes/orapts.jar endif # Required for JDBC connection setenv CLASSPATH $S/jdbc/lib/classes12.zip:$CLASSPATH -- -------------------------------------- Don MacQueen Environmental Protection Department Lawrence Livermore National Laboratory Livermore, CA, USA -------------------------------------- From deleeuw@stat.ucla.edu Tue Feb 11 22:57:20 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Tue, 11 Feb 2003 14:57:20 -0800 Subject: [R-SIG-Mac]Updates on gifi Message-ID: <2CBE7184-3E14-11D7-9F9D-000393860F3C@stat.ucla.edu> http://gifi.stat.ucla.edu/pub R-1.6.2 is updated to use the newest CRAN packages and the newest fink libraries. It is actually R-patched of 02-10-03. I am not distributing separate fink libraries any more, you have to build your own, and get Apple's X11 (this uses version 0.2). There are now 295 R package binaries included. MacPython-2.3a is updated from CVS, with updated versions of PyObjc, Numeric, Numarray, wxPython, and now also with PyOpenGL and PYRO. Ruby-1.8.0 is updated from CVS, and it contains the newest version of RubyCocoa. From deleeuw@stat.ucla.edu Sat Feb 15 00:34:57 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Fri, 14 Feb 2003 16:34:57 -0800 Subject: [R-SIG-Mac]macpython upgrade Message-ID: <4F8107F0-407D-11D7-AA9B-000393860F3C@stat.ucla.edu> MacPython on http://gifi.stat.ucla.edu/pub has been upgraded to include pyMPI, PYRO, PyOpenGL, PyOrbit, and PyGTK. This makes it more comparable with R and its many packages, and because Rpy is included you can really mix and match now. Next week I will post a new version of R, which fixes a slight incompatibility with wxPython (currently you cannot load Rpy into pycrust or pyshell or PythonCard), and an even bigger version of Python with PythonCard included (as well as OmniORBpy, and a fixed version of numarray). From deleeuw@stat.ucla.edu Sat Feb 15 04:48:12 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Fri, 14 Feb 2003 20:48:12 -0800 Subject: [R-SIG-Mac]Re: Darwin/X11 R missing dylib ? In-Reply-To: Message-ID: Here's the problem. On my machine [cabledoc111:/usr/lib] deleeuw% fink list | grep curses i ncurses 5.2-9 Binaries for the full-screen ascii drawing library and [cabledoc111:/usr/lib] deleeuw% otool -L /usr/local/lib/R/bin/R.bin /usr/local/lib/R/bin/R.bin: /System/Library/Frameworks/Carbon.framework/Versions/A/Carbon (compatibility version 2.0.0, current version 122.0.0) /sw/lib/libreadline.4.dylib (compatibility version 4.2.0, current version 4.3.0) /sw/lib/libdl.0.dylib (compatibility version 0.0.0, current version 0.0.0) /sw/lib/libncurses.5.dylib (compatibility version 6.0.0, current version 6.2.0) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 63.0.0) /sw/lib/libpcre.0.dylib (compatibility version 1.0.0, current version 1.1.0) /sw/lib/libbz2.1.dylib (compatibility version 1.0.1, current version 1.0.2) /usr/lib/libz.1.1.3.dylib (compatibility version 1.0.0, current version 1.1.3) On Friday, Feb 14, 2003, at 20:38 US/Pacific, William Duckworth wrote: > Okay, I reinstalled Apple's X11 0.2.1 and the SDK. I then > updated/installed all the fink items you listed. (I don't have the OSX > Developers' Tools installed, so I did it via FinkCommander. Is that > okay or do I need to install the developer tools from the Jaguar CD?) > > This is what I get now: > > [wmd:~] wmd% R > dyld: /usr/local/lib/R/bin/R.bin version mismatch for library: > /usr/lib/libncurses.5.dylib (compatibility version of user: 6.0.0 > greater than library's version: 5.0.0) > Trace/BPT trap > [wmd:~] wmd% > > Sounds like I need a newer version of the ncurses library (newer than > what fink installed)---am I correct? Fink installed ncurses 5.2-8. It > shows no "unstable" version number. > > Can you help me here or have I installed something improperly? > > Thanks for your time and help... > >> first install apple's x11 and its sdk >> >> start with "fink install dlcompat readline libjpeg libpng libtiff >> bzip2" >> >> then check again >> >> On Friday, February 14, 2003, at 02:38 PM, William Duckworth wrote: >> >>> I downloaded and installed the Darwin version of R from your site >>> and followed your draft instructions (but you don't provide the >>> dynamic lib download file anymore). >>> >>> This is what I get: >>> >>> [wmd:~] wmd% R >>> dyld: /usr/local/lib/R/bin/R.bin can't open library: >>> /sw/lib/libdl.0.dylib (No such file or directory, errno = 2) >>> Trace/BPT trap >>> [wmd:~] wmd% >>> >>> >>> I have the latest fink installed but I do not know what libraries I >>> need to install via fink. I'm running 10.2.3 of OSX and I have >>> Apple's X11 0.2.1 installed. >>> >>> Any assistance you can give will be greatly appreciated. Thanks. >>> >>> >>> -- >>> =================================================================== >>> Dr. William Duckworth II email: wmd@iastate.edu >>> 326 Snedecor Hall phone: 515-294-7766 >>> Iowa State University fax: 515-294-4040 >>> Statistical Laboratory and SPRING 2003 OFFICE HOURS: >>> Department of Statistics **** none **** >>> Ames, IA 50011-1210 http://www.public.iastate.edu/~wmd/ >>> =================================================================== > > > -- > =================================================================== > Dr. William Duckworth II email: wmd@iastate.edu > 326 Snedecor Hall phone: 515-294-7766 > Iowa State University fax: 515-294-4040 > Statistical Laboratory and SPRING 2003 OFFICE HOURS: > Department of Statistics **** none **** > Ames, IA 50011-1210 http://www.public.iastate.edu/~wmd/ > =================================================================== > > === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From john.maindonald@anu.edu.au Sun Feb 16 01:00:05 2003 From: john.maindonald@anu.edu.au (John Maindonald) Date: Sun, 16 Feb 2003 12:00:05 +1100 Subject: [R-SIG-Mac]Re: Apple's X11 In-Reply-To: <2CBE7184-3E14-11D7-9F9D-000393860F3C@stat.ucla.edu> Message-ID: As others may hit the same issues, I am posting a few comments on points re Apple's X11, which is indeed a great piece of software, even as currently advertised in beta mode. I am using the public beta (0.2.1) that became available on Feb 12, with SDK installed also. In part, I am posting this in the hope that others will have light to shed on these setup issues. (1) Fink users may find that, before installing Apple's X11, it may pay to read an up-to-date version of Section 3.2 of the document at http://fink.sourceforge.net/doc/x11/inst-xfree86.php Currently, this advises users to delete existing xfree86 packages before proceeding. I am not sure how necessary this is; Apple's X11 will work even if this is not done, but some fink-based software may not? (2) There are useful hints in http://developer.apple.com/qa/qa2001/qa1232.html Anyone who uses a beta version is a developer, right? Well, this was a point that dawned on me only after a fair amount of searching in less appropriate places for help. (3) Apple's X11 ignores /etc/csh.login Source /sw/bin/init.sh (not /sw/bin/init.csh, even though you get, finally, a tcsh shell). I used my .tcshrc for this purpose. (probably .cshrc is also ok) [Under the fink XFree86 implementation of X11, /etc/csh.login was a good place to source /sw/bin/init.sh, and a good place (on a single user machine) in which to set paths. Setting them in a .cshrc or .tcshrc under had the slightly disastrous effect of duplicating relevant bits of the path every time that a new xterm window was opened!] (4) [Actually, when I sourced /sw/bin/init.sh, I had an error: Bad : modifier in $ (/). I backed up /sw/bin/init.sh and copied /sw/bin/init.csh to /sw/bin/init.csh, and all was well. It may just have been that my /sw/bin/init.sh was corrupt.] (5) For setting up a .xinitrc file, copy the version that it at /etc/X11R6/xinit/xinitrc to .xinitrc into the home directory and modify that. At all events, I found that, on my system, this is what I had to do. Such an .xinitrc can be used to start additional xterms, or to invoke a different window manager. [Document qa1232.html noted above says /etc/X11/xinit/xinitrc, but that is surely a mistake.] (6) qa1232.html describes how to activate the selection proxying features (copy & paste) of quartz-wm, while using another window manager. I have yet to experiment with this. [I rather like icewm, because it is simple, and gives up to 4 desktops.] (7) After attending to these other matters, getting emacs, ess and R set up is a breeze! John Maindonald email: john.maindonald@anu.edu.au phone : +61 2 (6125)3473 fax : +61 2(6125)5549 Centre for Bioinformation Science, Room 1194, John Dedman Mathematical Sciences Building (Building 27) Australian National University, Canberra ACT 0200. From deleeuw@stat.ucla.edu Sun Feb 16 02:02:02 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Sat, 15 Feb 2003 18:02:02 -0800 Subject: [R-SIG-Mac]Re: Apple's X11 In-Reply-To: Message-ID: On Saturday, Feb 15, 2003, at 17:00 US/Pacific, John Maindonald wrote: > > (1) Fink users may find that, before installing Apple's X11, it may pay > to read an up-to-date version of Section 3.2 of the document at > http://fink.sourceforge.net/doc/x11/inst-xfree86.php > Currently, this advises users to delete existing xfree86 packages > before proceeding. I am not sure how necessary this is; Apple's > X11 will work even if this is not done, but some fink-based > software may not? If you use fink and Apple's X11 together, then fink will get confused, because you may have installed something using Apple's version and fink has no way of knowing it's library etc has been replaced. So you install fink's system-xfree86 and remove fink's xfree86. > > (2) There are useful hints in > > http://developer.apple.com/qa/qa2001/qa1232.html > > Anyone who uses a beta version is a developer, right? > Well, this was a point that dawned on me only after a fair > amount of searching in less appropriate places for help. Anyone who compiles anything is a developer, because you need the developer tools, and you can only get these by becoming a (free) member of the ADC. > > (3) Apple's X11 ignores /etc/csh.login > Source /sw/bin/init.sh (not /sw/bin/init.csh, even though you get, > finally, a tcsh shell). I used my .tcshrc for this purpose. > (probably .cshrc is also ok) On OS X csh is just tcsh (same as sh is just bash) > [Under the fink XFree86 implementation of X11, /etc/csh.login > was a good place to source /sw/bin/init.sh, and a good place > (on a single user machine) in which to set paths. Setting them > in a .cshrc or .tcshrc under had the slightly disastrous effect of > duplicating relevant bits of the path every time that a new xterm > window was opened!] I don't see that happening. And I do source /sw/bin/init.sh in .tchsrc. > > (4) [Actually, when I sourced /sw/bin/init.sh, I had an error: > Bad : modifier in $ (/). > I backed up /sw/bin/init.sh and copied /sw/bin/init.csh to > /sw/bin/init.csh, and all was well. It may just have been that my > /sw/bin/init.sh was corrupt.] Did you source init.sh (a bash shell script) from tcshrc ? > > (5) For setting up a .xinitrc file, copy the version that it at > /etc/X11R6/xinit/xinitrc to .xinitrc into the home directory and > modify that. At all events, I found that, on my system, this is > what I had to do. Such an .xinitrc can be used to start additional > xterms, or to invoke a different window manager. > [Document qa1232.html noted above says /etc/X11/xinit/xinitrc, > but that is surely a mistake.] No, that is a link added in 0.2. > > (6) qa1232.html describes how to activate the selection > proxying features (copy & paste) of quartz-wm, while using another > window manager. I have yet to experiment with this. [I rather > like icewm, because it is simple, and gives up to 4 desktops.] > > (7) After attending to these other matters, getting emacs, ess and > R set up is a breeze! Indeed. > > John Maindonald email: john.maindonald@anu.edu.au > phone : +61 2 (6125)3473 fax : +61 2(6125)5549 > Centre for Bioinformation Science, Room 1194, > John Dedman Mathematical Sciences Building (Building 27) > Australian National University, Canberra ACT 0200. > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > > === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From deleeuw@stat.ucla.edu Sun Feb 16 02:24:51 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Sat, 15 Feb 2003 18:24:51 -0800 Subject: Fwd: [R-SIG-Mac]Re: Apple's X11 Message-ID: Note that the fink document is about version 0.1, not about 0.2, Gnome/KDE hints, for instance, have been fixed. But the instructions on replacing fink's X11 by Apple's are very useful, and I recommend everybody to actually do this. For our student labs we now run R in an xterm, using a script and iHook (http://rsug.itd.umich.edu/software/ihook/) to start things up. Looks like a Cocoa application allright. Begin forwarded message: > On Saturday, Feb 15, 2003, at 17:00 US/Pacific, John Maindonald wrote: >> >> (1) Fink users may find that, before installing Apple's X11, it may >> pay >> to read an up-to-date version of Section 3.2 of the document at >> http://fink.sourceforge.net/doc/x11/inst-xfree86.php >> === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From u9801539@leonard.anu.edu.au Sun Feb 16 05:23:26 2003 From: u9801539@leonard.anu.edu.au (John Maindonald) Date: Sun, 16 Feb 2003 16:23:26 +1100 (EST) Subject: [R-SIG-Mac]Re: Apple's X11 Message-ID: <200302160523.h1G5NQl05444@leonard.anu.edu.au> Hello Jan - Probably I should have contacted you earlier. I did however learn a great deal from my slow bungling exploration. > On Saturday, Feb 15, 2003, at 17:00 US/Pacific, John Maindonald wrote: > > > > (1) Fink users may find that, before installing Apple's X11, it may pay > > to read an up-to-date version of Section 3.2 of the document at > > http://fink.sourceforge.net/doc/x11/inst-xfree86.php > > Currently, this advises users to delete existing xfree86 packages > > before proceeding. I am not sure how necessary this is; Apple's > > X11 will work even if this is not done, but some fink-based > > software may not? > > If you use fink and Apple's X11 together, then fink will get confused, > because you > may have installed something using Apple's version and fink has no way > of knowing > it's library etc has been replaced. So you install fink's > system-xfree86 and remove > fink's xfree86. I figured that was probably what was happening. First time round, I did though install Apple's X11, unthinkingly, without checking the fink documentation. > > (2) There are useful hints in > > > > http://developer.apple.com/qa/qa2001/qa1232.html > > > > Anyone who uses a beta version is a developer, right? > > Well, this was a point that dawned on me only after a fair > > amount of searching in less appropriate places for help. > > Anyone who compiles anything is a developer, because > you need the developer tools, and you can only get these > by becoming a (free) member of the ADC. So that is what I signed up to! > > (3) Apple's X11 ignores /etc/csh.login > > Source /sw/bin/init.sh (not /sw/bin/init.csh, even though you get, > > finally, a tcsh shell). I used my .tcshrc for this purpose. > > (probably .cshrc is also ok) > > On OS X csh is just tcsh (same as sh is just bash) > > > [Under the fink XFree86 implementation of X11, /etc/csh.login > > was a good place to source /sw/bin/init.sh, and a good place > > (on a single user machine) in which to set paths. Setting them > > in a .cshrc or .tcshrc under had the slightly disastrous effect of > > duplicating relevant bits of the path every time that a new xterm > > window was opened!] > > I don't see that happening. And I do source /sw/bin/init.sh in > .tchsrc. At all events, this is not the behaviour under Apple's X11. > > (4) [Actually, when I sourced /sw/bin/init.sh, I had an error: > > Bad : modifier in $ (/). > > I backed up /sw/bin/init.sh and copied /sw/bin/init.csh to > > /sw/bin/init.csh, and all was well. It may just have been that my > > /sw/bin/init.sh was corrupt.] > > Did you source init.sh (a bash shell script) from tcshrc ? Yes, I did, after earlier trying putting it into .xinitrc. Sourcing the file as init.csh, whether from .xinitrc or from .tcshrc, had no effect on the path. Is there another, more appropriate, place for this? It was a question of where to go to next when I found that following the Fink document and putting . /sw/bin/init.sh as the third line in my .xinitrc did not do the trick. > > (5) For setting up a .xinitrc file, copy the version that it at > > /etc/X11R6/xinit/xinitrc to .xinitrc into the home directory and > > modify that. At all events, I found that, on my system, this is > > what I had to do. Such an .xinitrc can be used to start additional > > xterms, or to invoke a different window manager. > > [Document qa1232.html noted above says /etc/X11/xinit/xinitrc, > > but that is surely a mistake.] > > No, that is a link added in 0.2. I see now that I do have that link in my installation. Earlier, I had been unable to find it, perhaps because my .xinitrc was incomplete? > > (6) qa1232.html describes how to activate the selection > > proxying features (copy & paste) of quartz-wm, while using another > > window manager. I have yet to experiment with this. [I rather > > like icewm, because it is simple, and gives up to 4 desktops.] > > > > (7) After attending to these other matters, getting emacs, ess and > > R set up is a breeze! > > Indeed. > > > > > John Maindonald email: john.maindonald@anu.edu.au > > phone : +61 2 (6125)3473 fax : +61 2(6125)5549 > > Centre for Bioinformation Science, Room 1194, > > John Dedman Mathematical Sciences Building (Building 27) > > Australian National University, Canberra ACT 0200. > > > > _______________________________________________ > > R-SIG-Mac mailing list > > R-SIG-Mac@stat.math.ethz.ch > > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > > > > > === > Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; > Editor: Journal of Multivariate Analysis, Journal of Statistical > Software > US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 > phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu > homepage: http://gifi.stat.ucla.edu > > ------------------------------------------------------------------------ > ------------------------- > No matter where you go, there you are. --- Buckaroo Banzai > http://gifi.stat.ucla.edu/sounds/nomatter.au > > ------------------------------------------------------------------------ > ------------------------- John Maindonald email : john.maindonald@anu.edu.au Centre for Bioinformation Science, phone : (6125)3473 c/o MSI, fax : (6125)5549 John Dedman Mathematical Sciences Building (Building 27) Australian National University Canberra ACT 0200 Australia From deleeuw@stat.ucla.edu Tue Feb 18 21:56:04 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Tue, 18 Feb 2003 13:56:04 -0800 Subject: [R-SIG-Mac]upgraded R-1.6.2 on gifi Message-ID: http://gifi.stat.ucla.edu/pub has a new version of R-patched with some CRAN package updates and with a little hack that allows you to run Rpy and wxPython together in Python (which all of you were dying to do, I guess). From daver969@yahoo.com Wed Feb 19 06:50:48 2003 From: daver969@yahoo.com (David Richmond) Date: Tue, 18 Feb 2003 22:50:48 -0800 Subject: [R-SIG-Mac]Installing Packages Message-ID: <7A325DD5-43D6-11D7-ADE6-000393DBCEC2@yahoo.com> Hi, I am trying to install a package I created myself using the R CMD INSTALL command. I have successfully installed other packages using this command (in .tar.gz format). My own package is not gzipped, but when I "sudo R CMD INSTALL /my/path/mypackage" it seems to install perfectly well, and indeed when I run R (I'm using OS X 10.2.4, btw) and do "library()" it lists my package along with all the others. I can even "library(help=daver)" to see the description and index files for my package. BUT, when I "library(daver)" to load the package, I get the following error: Error: package daver was built for NA In addition: Warning message: package daver was built under R version 02-12-27 and it doesn't see any of my functions. I believe that my package has all the required elements, description, index, title, and an R folder with the source code in it. Why does my package not work, have I missed something? Dave R. From jago@mclink.it Wed Feb 19 08:37:44 2003 From: jago@mclink.it (Stefano Iacus) Date: Wed, 19 Feb 2003 09:37:44 +0100 Subject: [R-SIG-Mac]Installing Packages In-Reply-To: <7A325DD5-43D6-11D7-ADE6-000393DBCEC2@yahoo.com> Message-ID: <6A56FB94-43E5-11D7-9562-003065CC4CB8@mclink.it> Please, look into "Description" file to see if the Built: field is present. stefano On Mercoledì, feb 19, 2003, at 07:50 Europe/Rome, David Richmond wrote: > Hi, > I am trying to install a package I created myself using the R CMD > INSTALL command. I have successfully installed other packages using > this command (in .tar.gz format). My own package is not gzipped, but > when I "sudo R CMD INSTALL /my/path/mypackage" it seems to install > perfectly well, and indeed when I run R (I'm using OS X 10.2.4, btw) > and do "library()" it lists my package along with all the others. I > can even "library(help=daver)" to see the description and index files > for my package. BUT, when I "library(daver)" to load the package, I > get the following error: > > Error: package daver was built for NA > In addition: Warning message: > package daver was built under R version 02-12-27 > > and it doesn't see any of my functions. > > I believe that my package has all the required elements, description, > index, title, and an R folder with the source code in it. Why does my > package not work, have I missed something? > > Dave R. > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > From daver969@yahoo.com Wed Feb 19 15:44:32 2003 From: daver969@yahoo.com (David Richmond) Date: Wed, 19 Feb 2003 07:44:32 -0800 Subject: [R-SIG-Mac]Installing Packages In-Reply-To: <6A56FB94-43E5-11D7-9562-003065CC4CB8@mclink.it> Message-ID: <09D41BBE-4421-11D7-ADE6-000393DBCEC2@yahoo.com> I have Built: 2002-12-27 I was under the impression that the "built" field requires a date, but perhaps it is asking for the version of R under which the package was compiled? Dave On Wednesday, February 19, 2003, at 12:37 AM, Stefano Iacus wrote: > Please, look into "Description" file to see if the Built: field is > present. > > stefano > > On Mercoledì, feb 19, 2003, at 07:50 Europe/Rome, David Richmond wrote: > >> Hi, >> I am trying to install a package I created myself using the R CMD >> INSTALL command. I have successfully installed other packages using >> this command (in .tar.gz format). My own package is not gzipped, but >> when I "sudo R CMD INSTALL /my/path/mypackage" it seems to install >> perfectly well, and indeed when I run R (I'm using OS X 10.2.4, btw) >> and do "library()" it lists my package along with all the others. I >> can even "library(help=daver)" to see the description and index files >> for my package. BUT, when I "library(daver)" to load the package, I >> get the following error: >> >> Error: package daver was built for NA >> In addition: Warning message: >> package daver was built under R version 02-12-27 >> >> and it doesn't see any of my functions. >> >> I believe that my package has all the required elements, description, >> index, title, and an R folder with the source code in it. Why does my >> package not work, have I missed something? >> >> Dave R. >> >> _______________________________________________ >> R-SIG-Mac mailing list >> R-SIG-Mac@stat.math.ethz.ch >> http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac >> > From deleeuw@stat.ucla.edu Thu Feb 20 19:25:39 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Thu, 20 Feb 2003 11:25:39 -0800 Subject: [R-SIG-Mac]Python, R, Rpy, wxPython updates on gifi Message-ID: <183EF188-4509-11D7-B2A3-000393860F3C@stat.ucla.edu> --Apple-Mail-2-889946395 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII; format=flowed http://gifi.stat.ucla.edu/pub The update of the macpython-2.3a archive now includes /Applications/wxPythonOSX-2.4.0.2, with an application bundle for pyShell. It also include wxWindowsMac. The framework Python build contains Numeric, numarray, Rpy (0.3.1), wxPython, pyObjc, pygtk. One nice thing about this is that the following sequence actually works (both from pyShell and from PythonIDE) >>> from rpy import * >>> x=r.seq(100) >>> y=r.rnorm(100) >>> r.quartz() >>> r.plot(x,y) It produces a scatterplot in the quartz device window that actually responds to events, i.e. can be brought to the foreground and so on. And you can run >>> r.demo("graphics") Another nice thing, of course, is that you can do wxWindows and Cocoa Programming and R from the same pyShell or PythonIDE window, and every other thing Python. One thing you should NOT do is to try pygtk or the R tcltk package, because they still need X11. My next step is to incorporate PythonCard and build a version of libR.dylib that uses tkAqua (Python already uses tkAqua). I'll remove pygtk, until it can be compiled with the native Carbon gtk. By the way, the IDLE in /Applications/MacPython-2.3a currently does not work. The combined application you have is, of course, a compromise. It has quite a few things (LAPACK, tcl/tk, basic functionality) both on the R and the Python side, which is wasteful. The Carbon stuff in Python seems sort of slow. And Rpy is no speed daemon either. --Apple-Mail-2-889946395 Content-Transfer-Encoding: 7bit Content-Type: text/enriched; charset=US-ASCII http://gifi.stat.ucla.edu/pub The update of the macpython-2.3a archive now includes /Applications/Lucida GrandewxPythonOSX-2.4.0.2, with an application bundle for pyShell. It also include wxWindowsMac. The framework Python build contains Numeric, numarray, Rpy (0.3.1), wxPython, pyObjc, pygtk. One nice thing about this is that the following sequence actually works (both from pyShell and from PythonIDE) >>> from rpy import * >>> x=r.seq(100) >>> y=r.rnorm(100) >>> r.quartz() >>> r.plot(x,y) It produces a scatterplot in the quartz device window that actually responds to events, i.e. can be brought to the foreground and so on. And you can run >>> r.demo("graphics") Another nice thing, of course, is that you can do wxWindows and Cocoa Programming and R from the same pyShell or PythonIDE window, and every other thing Python. One thing you should NOT do is to try pygtk or the R tcltk package, because they still need X11. My next step is to incorporate PythonCard and build a version of libR.dylib that uses tkAqua (Python already uses tkAqua). I'll remove pygtk, until it can be compiled with the native Carbon gtk. By the way, the IDLE in /Applications/MacPython-2.3a currently does not work. The combined application you have is, of course, a compromise. It has quite a few things (LAPACK, tcl/tk, basic functionality) both on the R and the Python side, which is wasteful. The Carbon stuff in Python seems sort of slow. And Rpy is no speed daemon either. --Apple-Mail-2-889946395-- From deleeuw@stat.ucla.edu Thu Feb 20 21:01:42 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Thu, 20 Feb 2003 13:01:42 -0800 Subject: [R-SIG-Mac]how to configure R-devel Message-ID: <83446799-4516-11D7-B2A3-000393860F3C@stat.ucla.edu> If you don't want X11 and you don't want ATLAS configure --with-tcl-config=/Library/Frameworks/Tcl.framework/tclConfig.sh --with-tk-config=/Library/Frameworks/Tk.framework/tkConfig.sh --without-x --enable-R-shlib --with-blas='-framework vecLib' --with-lapack --with-aqua TCLTK_LIBS='-framework Tcl -framework Tk' TCLTK_CPPFLAGS='-framework Tcl -framework Tk' From deleeuw@stat.ucla.edu Sat Feb 22 17:52:17 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Sat, 22 Feb 2003 09:52:17 -0800 Subject: [R-SIG-Mac]dynamic loader in darwin Message-ID: <620A58D6-468E-11D7-8274-000393BB6D36@stat.ucla.edu> One way to check which dynamic libraries you need if you download an R binary is to run "otool -L" on all executables. i.e. on R.bin and all foo.so bundles. Another way to find out which libraries dyld is actually loading at run time is to do setenv DYLD_PRINT_LIBRARIES 1 and then start R. Then dyld will tell you which libraries it is loading. === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From david_strong@cox.net Tue Feb 25 15:52:45 2003 From: david_strong@cox.net (David Strong) Date: Tue, 25 Feb 2003 10:52:45 -0500 Subject: [R-SIG-Mac]win.graph Message-ID: <2E32EFD1-48D9-11D7-9FA5-0030657C49C2@cox.net> Hello, I am trying to adapt code for use in R 1.6 on a Mac in OS 10.2. The program uses 'win.graph' to set plot size. What is the Mac equivalent? I have searched where I can but no luck. Any help would be appreciated. Here is a snip of code where this is implemented.: if (Gr.Avg) { if (Gr.Orient == 1) win.graph(width=8.5, height=11) else win.graph(width=11, height=8.5) par(mfrow=c(Gr.Rows, Gr.Cols), omi=c(.5,.5,.5,.5)) X.Values <- Curves[[1]][,1] X.Range <- c(min(X.Values), max(X.Values)) Y.Values <- rep(0,length(X.Values)) for (i in 1:length(Curve.Info[,1])) Y.Values <- Y.Values + Curves[[i]][,2] Y.Values <- Y.Values / length(Curve.Info[,1]) Y.Range <- c(min(Y.Values), max(Y.Values)) P.Avg <- 1/((Y.Values[length(Y.Values)]/Y.Values[1]) + 1) if (P.Avg < 0) P.Avg <- 0 # correct for possibilities of negative values at extremes if (P.Avg > 1) P.Avg <- 1 plot(X.Values, Y.Values, type = "p", pch = 16, ylim = Y.Range, xlim = X.Range, ylab = "Mean Difference", xlab = "Input") lines(X.Values, Y.Values) # solid line for raw data only mtext(paste("MAMBAC AVG CURVE, ",Data.Label),outer=T) } return(P.Avg) From Jean.Thioulouse@biomserv.univ-lyon1.fr Wed Feb 26 12:34:01 2003 From: Jean.Thioulouse@biomserv.univ-lyon1.fr (Jean Thioulouse) Date: Wed, 26 Feb 2003 13:34:01 +0100 Subject: [R-SIG-Mac]compiling r-devel on Darwin Message-ID: Hello, I am trying to compile the latest r-devel under MacOS 10.2.4, and I get the following error : gcc -dynamiclib -install_name /usr/local/lib/R/bin/libRlapack.dylib -L/sw/lib -L/usr/local/lib -o libRlapack.dylib dlapack0.lo dlapack1.lo dlapack2.lo dlapack3.lo cmplx.lo -lf77blas -latlas -L/sw/lib -L/usr/local/lib -L/sw/lib/gcc-lib/powerpc-apple-darwin6.2/3.1 -L/sw/lib/gcc-lib/powerpc-apple-darwin6.2/3.1/../../.. -lfrtbegin -lg2c -lSystem ld: common symbols not allowed with MH_DYLIB output format /usr/bin/libtool: internal link edit command failed /sw/lib/libg2c.a(err.o) definition of common _f__cblank (size 4) /sw/lib/libg2c.a(fmt.o) definition of common _f__cnt (size 40) /sw/lib/libg2c.a(fmt.o) definition of common f(char, *) (size 4) /sw/lib/libg2c.a(err.o) definition of common f(char, long *, short __restrict) (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__cursor (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__doed (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__doend (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__doned (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__dorevert (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__elist (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__fmtbuf (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__fmtlen (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__init (size 4) /sw/lib/libg2c.a(fmt.o) definition of common _f__nonl (size 4) /sw/lib/libg2c.a(fmt.o) definition of common _f__parenlvl (size 4) /sw/lib/libg2c.a(fmt.o) definition of common f(char *) (size 4) /sw/lib/libg2c.a(fmt.o) definition of common _f__ret (size 40) /sw/lib/libg2c.a(fmt.o) definition of common _f__revloc (size 4) /sw/lib/libg2c.a(fmt.o) definition of common f(long double, *) (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__scale (size 4) /sw/lib/libg2c.a(fmt.o) definition of common _f__workdone (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__units (size 4800) /sw/lib/libg2c.a(err.o) definition of common f(char, float) (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__curunit (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__donewrec (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__external (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__formatted (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__hiwater (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__putn (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__reading (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__recpos (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__sequential (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__getn (size 4) /sw/lib/libg2c.a(err.o) definition of common _f__icptr (size 4) /sw/lib/libg2c.a(err.o) definition of common f(short, void, int, char) (size 4) make[4]: *** [libRlapack.dylib] Error 1 make[3]: *** [R] Error 2 make[2]: *** [R] Error 1 make[1]: *** [R] Error 1 make: *** [R] Error 1 I have no problem compiling 1.6.2. What should I do ? Thanks, Jean -- Jean Thioulouse - Equipe "Ecologie Statistique" - UMR CNRS 5558 Universite Lyon 1, Bat. Mendel, 69622 Villeurbanne Cedex, France Fax: (33) 4 78 89 27 19 Tel: (33) 4 72 43 27 56 http://pbil.univ-lyon1.fr/JTHome.html From deleeuw@stat.ucla.edu Wed Feb 26 16:46:32 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Wed, 26 Feb 2003 08:46:32 -0800 Subject: [R-SIG-Mac]compiling r-devel on Darwin In-Reply-To: Message-ID: I don't have this. What this says is that the modules in libg2c.a have not been compiled with -fno-common., which indicates there is something wrong with your g77 from fink. Make sure you have the latest OS 10.2 version. In any case use the g77 from the unstable tree, and try "fink rebuild g77" On Wednesday, Feb 26, 2003, at 04:34 US/Pacific, Jean Thioulouse wrote: > Hello, > > I am trying to compile the latest r-devel under MacOS 10.2.4, and > I get the following error : > > gcc -dynamiclib -install_name /usr/local/lib/R/bin/libRlapack.dylib > -L/sw/lib -L/usr/local/lib -o libRlapack.dylib dlapack0.lo dlapack1.lo > dlapack2.lo dlapack3.lo cmplx.lo -lf77blas -latlas -L/sw/lib > -L/usr/local/lib -L/sw/lib/gcc-lib/powerpc-apple-darwin6.2/3.1 > -L/sw/lib/gcc-lib/powerpc-apple-darwin6.2/3.1/../../.. -lfrtbegin > -lg2c -lSystem > ld: common symbols not allowed with MH_DYLIB output format > /usr/bin/libtool: internal link edit command failed > /sw/lib/libg2c.a(err.o) definition of common _f__cblank (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common _f__cnt (size 40) > /sw/lib/libg2c.a(fmt.o) definition of common f(char, *) (size 4) > /sw/lib/libg2c.a(err.o) definition of common f(char, long *, short > __restrict) (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__cursor (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__doed (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__doend (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__doned (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__dorevert (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__elist (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__fmtbuf (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__fmtlen (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__init (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common _f__nonl (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common _f__parenlvl (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common f(char *) (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common _f__ret (size 40) > /sw/lib/libg2c.a(fmt.o) definition of common _f__revloc (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common f(long double, *) (size > 4) > /sw/lib/libg2c.a(err.o) definition of common _f__scale (size 4) > /sw/lib/libg2c.a(fmt.o) definition of common _f__workdone (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__units (size 4800) > /sw/lib/libg2c.a(err.o) definition of common f(char, float) (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__curunit (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__donewrec (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__external (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__formatted (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__hiwater (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__putn (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__reading (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__recpos (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__sequential (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__getn (size 4) > /sw/lib/libg2c.a(err.o) definition of common _f__icptr (size 4) > /sw/lib/libg2c.a(err.o) definition of common f(short, void, int, char) > (size 4) > make[4]: *** [libRlapack.dylib] Error 1 > make[3]: *** [R] Error 2 > make[2]: *** [R] Error 1 > make[1]: *** [R] Error 1 > make: *** [R] Error 1 > > I have no problem compiling 1.6.2. What should I do ? > > Thanks, > Jean > > -- > Jean Thioulouse - Equipe "Ecologie Statistique" - UMR CNRS 5558 > Universite Lyon 1, Bat. Mendel, 69622 Villeurbanne Cedex, France > Fax: (33) 4 78 89 27 19 Tel: (33) 4 72 43 27 56 > http://pbil.univ-lyon1.fr/JTHome.html > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > > === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From eglen@pcg.wustl.edu Wed Feb 26 23:31:05 2003 From: eglen@pcg.wustl.edu (Stephen Eglen) Date: Wed, 26 Feb 2003 17:31:05 -0600 Subject: [R-SIG-Mac]Compiling R libraries on MAC OS X Message-ID: <15965.20025.273557.622742@mosaics.wustl.edu> Hi, can anyone help me with the following problem? [Using R-1.6.2 patched on Mac OS X 10.2.4, Developer tools from ~August?] I have made four personal R libaries on Linux, and I'm trying to get them to work on Mac OS X. All of them have C code that is called from R using .C(). Three out of the four work fine, but when I try to compile one of them (a Voronoi library, based on Steve Fortune's code) ld complains. % R CMD check sjevor * checking for working latex ... OK * using log directory '/Users/stephen/langs/R/sjevor.Rcheck' * checking for file 'sjevor/DESCRIPTION' ... OK * checking if this is a source package ... OK * Installing *source* package 'sjevor' ... ** libs gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c edgelist.c -o edgelist.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c geometry.c -o geometry.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c heap.c -o heap.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c main.c -o main.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c memory.c -o memory.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c output.c -o output.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c sjevor.c -o sjevor.o gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include -I/usr/local/include -g -fno-common -g -O2 -c voronoi.c -o voronoi.o gcc -bundle -flat_namespace -undefined suppress -L/sw/lib -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o main.o memory.o output.o sjevor.o voronoi.o ld: multiple definitions of symbol _ELhash edgelist.o definition of _ELhash in section (__DATA,__common) geometry.o definition of _ELhash in section (__DATA,__common) ld: multiple definitions of symbol _ELhashsize edgelist.o definition of _ELhashsize in section (__DATA,__common) geometry.o definition of _ELhashsize in section (__DATA,__common) ld: multiple definitions of symbol _ELleftend ... ld produces many similar complaints, before the build fails. Has anyone else come across this problem? The package builds fine on my linux box. Thanks for any pointers, Stephen From deleeuw@stat.ucla.edu Wed Feb 26 23:54:59 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Wed, 26 Feb 2003 15:54:59 -0800 Subject: [R-SIG-Mac]Compiling R libraries on MAC OS X In-Reply-To: <15965.20025.273557.622742@mosaics.wustl.edu> Message-ID: If you use flat namespaces you cannot have multiple definitions of symbols, because the linked application has no way of knowing which definition you meant. There are a number of ways to deal with this. You can use two-level namespaces (which give warnings for multiple definitions) > gcc -bundle -bundle-loader /usr/local/lib/R/bin/R.bin -undefined error > -L/sw/lib -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o > main.o memory.o output.o sjevor.o voronoi.o You can change the C code such that all definitions but one use extern (this is best). And you can force the link by adding the -m flag, which also transforms errors to warnings, i.e. > gcc -bundle -flat_namespace -undefined suppress -Xlinker -m -L/sw/lib > -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o main.o > memory.o output.o sjevor.o voronoi.o On Wednesday, Feb 26, 2003, at 15:31 US/Pacific, Stephen Eglen wrote: > > Hi, > can anyone help me with the following problem? > > [Using R-1.6.2 patched on Mac OS X 10.2.4, Developer tools from > ~August?] > > I have made four personal R libaries on Linux, and I'm trying to get > them to work on Mac OS X. All of them have C code that is called from > R using .C(). Three out of the four work fine, but when I try to > compile one of them (a Voronoi library, based on Steve Fortune's code) > ld complains. > > % R CMD check sjevor > > * checking for working latex ... OK > * using log directory '/Users/stephen/langs/R/sjevor.Rcheck' > * checking for file 'sjevor/DESCRIPTION' ... OK > * checking if this is a source package ... OK > * Installing *source* package 'sjevor' ... > ** libs > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c edgelist.c -o > edgelist.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c geometry.c -o > geometry.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c heap.c -o heap.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c main.c -o main.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c memory.c -o memory.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c output.c -o output.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c sjevor.c -o sjevor.o > gcc -no-cpp-precomp -I/usr/local/lib/R/include -I/sw/include > -I/usr/local/include -g -fno-common -g -O2 -c voronoi.c -o voronoi.o > gcc -bundle -flat_namespace -undefined suppress -L/sw/lib > -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o main.o > memory.o output.o sjevor.o voronoi.o > ld: multiple definitions of symbol _ELhash > edgelist.o definition of _ELhash in section (__DATA,__common) > geometry.o definition of _ELhash in section (__DATA,__common) > ld: multiple definitions of symbol _ELhashsize > edgelist.o definition of _ELhashsize in section (__DATA,__common) > geometry.o definition of _ELhashsize in section (__DATA,__common) > ld: multiple definitions of symbol _ELleftend > ... > > ld produces many similar complaints, before the build fails. > > Has anyone else come across this problem? The package builds fine on > my linux box. > > Thanks for any pointers, > Stephen > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > > === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From eglen@pcg.wustl.edu Thu Feb 27 01:29:58 2003 From: eglen@pcg.wustl.edu (Stephen Eglen) Date: Wed, 26 Feb 2003 19:29:58 -0600 Subject: [R-SIG-Mac]Compiling R libraries on MAC OS X In-Reply-To: References: <15965.20025.273557.622742@mosaics.wustl.edu> Message-ID: <15965.27158.339360.908337@mosaics.wustl.edu> Jan de Leeuw writes: > If you use flat namespaces you cannot have multiple definitions of > symbols, > because the linked application has no way of knowing which definition > you meant. There are a number of ways to deal with this. You can use > two-level namespaces (which give warnings for multiple definitions) > > > gcc -bundle -bundle-loader /usr/local/lib/R/bin/R.bin -undefined error > > -L/sw/lib -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o > > main.o memory.o output.o sjevor.o voronoi.o > > You can change the C code such that all definitions but one use extern > (this is best). And you can force the link by adding the > -m flag, which also transforms errors to warnings, i.e. > > > gcc -bundle -flat_namespace -undefined suppress -Xlinker -m -L/sw/lib > > -L/usr/local/lib -o sjevor.so edgelist.o geometry.o heap.o main.o > > memory.o output.o sjevor.o voronoi.o Thanks very much Jan for your help. I have updated the C code so that the global variables are defined once, and then "extern"ed in other files that need them. This works very well. Stephen From deleeuw@stat.ucla.edu Thu Feb 27 02:19:42 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Wed, 26 Feb 2003 18:19:42 -0800 Subject: [R-SIG-Mac]nice combo on os x Message-ID: You need (a) a recent OS X framework build of python, which also installs IDLE (b) a recent OS X framework build of AquaTk (i.e. native Tcl/Tk) (c) a version of R compiled with AquaTk and without X11 (d) a recent version of Rpy Now you can open IDLE and say >> from rpy import * >> r.quartz() >> r.library("tcltk") >> r.demo("tkdensity") This all works. You get the Aqua Tk widgets and the plotting in the quartz device. Windows can be brought to the foreground, minimized to the dock, and widgets can be operated with the mouse. No X11 is required at all. === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ ------------------------- From rick@cs.colorado.edu Fri Feb 28 04:29:00 2003 From: rick@cs.colorado.edu (Rick Osborne) Date: Thu, 27 Feb 2003 21:29:00 -0700 (MST) Subject: [R-SIG-Mac]Newbie Question Message-ID: Hello, I recently installed R on my new PowerBook G4 moving from the Windows world. In that world, I was a fairly competent user. However there are two things about the Mac implementation that I can not figure out. Stefano has already tried to answer one of my questions. However, it did not work or I did not understand. Here are the questions: 1) When I try to change the directory and read in data using the load("filename.Rda") command, I keep getting the following error message: >load("filename.Rda") Error: bad restore file magic number (file may be corrupted)-- no data loaded. All this is is a simple text file. I have tried changing the directory every which way, changing the extension etc but I can not read data in and I keep getting that error message. What do I need to do? This is very frustrating for such a simple operation. 2) The next question is how does one install a package that is not on the CRAN website? I downloaded the Simple package but do not know how to get it working. Dragging it in to the R folder does not seem to do it. Thanks very much and thanks for the port to Mac OS X. Rick Osborne From jago@mclink.it Fri Feb 28 07:30:18 2003 From: jago@mclink.it (Stefano Iacus) Date: Fri, 28 Feb 2003 08:30:18 +0100 Subject: [R-SIG-Mac]Newbie Question In-Reply-To: Message-ID: <7CF7E8BC-4AEE-11D7-BEF7-003065CC4CB8@mclink.it> On Venerdì, feb 28, 2003, at 05:29 Europe/Rome, Rick Osborne wrote: > Hello, > > I recently installed R on my new PowerBook G4 moving from the Windows > world. In that world, I was a fairly competent user. However there are > two > things about the Mac implementation that I can not figure out. Stefano > has > already tried to answer one of my questions. However, it did not work > or I > did not understand. Here are the questions: > > 1) When I try to change the directory and read in data using the > load("filename.Rda") command, I keep getting the following error > message: > >> load("filename.Rda") > > Error: bad restore file magic number (file may be corrupted)-- no data > loaded. Rick, try to read the help files. If it is a simple txt file, then you should use read.table instead. There are a couple of good intros written by R-core members like: * An introduction to R * R Data Import/Export you can find both inside the doc folder of you installation of R or (better) on the www.r-project.org site in the "Manuals" section > > All this is is a simple text file. I have tried changing the directory > every which way, changing the extension etc but I can not read data in > and > I keep getting that error message. > > What do I need to do? This is very frustrating for such a simple > operation. > > 2) The next question is how does one install a package that is not on > the > CRAN website? I downloaded the Simple package but do not know how to > get > it working. Dragging it in to the R folder does not seem to do it. > If packages does not contain C/Fortran code, then you can use any prebuilt version of it (either Windows or Unix) You cannot use packages as they are beacusa they need to be built. If you want to build packages on your own, have a look inside the R sources in the file INSTALL of the src/macintosh directory. > Thanks very much and thanks for the port to Mac OS X. > The "true" MacOSX version of R is the Darwin/R and not the Carbon one you are using. The Carbon version of R will be soon abandoned and fully replaced by the Darwin one. stefano > Rick Osborne > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > http://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac > From deleeuw@stat.ucla.edu Fri Feb 28 19:39:06 2003 From: deleeuw@stat.ucla.edu (Jan de Leeuw) Date: Fri, 28 Feb 2003 11:39:06 -0800 Subject: [R-SIG-Mac]R plus Python Message-ID: <4CAF1071-4B54-11D7-926F-000393BB6D36@stat.ucla.edu> Later today the file python_plus_R.tar.gz will appear on http://gifi.stat.ucla.edu/pub. It contains R-1.6.2 (patched) with binaries for 200-300 packages. It also includes the fink dynamic libs needed for running R.bin (these are installed in /sw/lib). Second, it contains a Python framework build, that comes with PyObjc, PyOpenGL, Numeric, numarray, PyPostgres, PYRO, PythonCardPrototype, wxPython, and, most importantly, Rpy. Thus you can program in Cocoa, OpenGL, wxWindows, and R from this python. Third, it has the Tcl and Tk native framework builds (used both in tkinter and R). In /Applications it installs a MacPython-2.3 and a wxPythonOSX-2.4.0.2 folder, which contain various front ends. For instance, there is IDLE (useful for tkinter programming) and PyShell (useful for wxWindows programming). In /usr/local/bin you find the python, pythonw, pyshell, pycrust, codeEditor, and resourceEditor scripts (the first two are from the Python distro, the next two from wxPython, the last two from PythonCard). Of course you can still run R from the command line (Terminal.app) if you want to. If you open IDLE, or PyCrust, or PythonIDE, you can say >>>import * from rpy >>>r.quartz() >>>r.demo("graphics") and things should roll along. No X11 server required ! In IDLE you can even say, in addition to the stuff above, >>>r.library("tcltk") >>>r.demo("tkcanvas") Still no X11 required. If you start python from the command line (using the pythonw script), you can give the same sequence of commands, but then you will use the X11 graphics device (and the X11 version of tk). Plenty of choices. Of course the nice thing is that you can combine R programming with Cocoa programming or wxWindows programming, and you can write python scripts that use python versions of R, numarray, and OpenGL commands. You can even run >>>r.update_packages() from the Python GUI's. === Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; Editor: Journal of Multivariate Analysis, Journal of Statistical Software US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://gifi.stat.ucla.edu ------------------------------------------------------------------------ ------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.au ------------------------------------------------------------------------ -------------------------