Re: [Rd] R_LIBS_USER bugs
On Wed, Feb 17, 2010 at 07:49:20AM -0800, Seth Falcon wrote: Hi, On 2/16/10 10:31 AM, Jens Elkner wrote: Having currently a big problem with R 2.10.1 vanilla (Solaris): As soon as the R_LIBS_USER env var gets bigger than 1023 chars R completely ignores it and uses the default: BTW: Sometimes even more than 1000 chars cause an 'overflow' ... I guess the first question is, why do need such a long list of library directories? E.g. for a clean/uncluttered build system. Distributed filesystem for several hundred students is another use case ... Sys.getenv('R_LIBS_USER'); R_LIBS_USER ${R_LIBS_USER-~/R/i386-pc-solaris2.11-library/2.10} I see the same thing with R-devel on OS X. I can set R_LIBS_USER from within R using Sys.setenv to a value longer than 1024 and retrieve it again. But if I have such a value in my shell, it gets overwritten. Strange and buggy IMHO ... Regards, jel. -- Otto-von-Guericke University http://www.cs.uni-magdeburg.de/ Department of Computer Science Geb. 29 R 027, Universitaetsplatz 2 39106 Magdeburg, Germany Tel: +49 391 67 12768 __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] R_LIBS_USER bugs
Hi, On 2/16/10 10:31 AM, Jens Elkner wrote: Having currently a big problem with R 2.10.1 vanilla (Solaris): As soon as the R_LIBS_USER env var gets bigger than 1023 chars R completely ignores it and uses the default: I guess the first question is, why do need such a long list of library directories? Sys.getenv('R_LIBS_USER'); R_LIBS_USER ${R_LIBS_USER-~/R/i386-pc-solaris2.11-library/2.10} I see the same thing with R-devel on OS X. I can set R_LIBS_USER from within R using Sys.setenv to a value longer than 1024 and retrieve it again. But if I have such a value in my shell, it gets overwritten. I'm not yet sure what is going on. + seth __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
[Rd] R_LIBS_USER bugs
Hi, Having currently a big problem with R 2.10.1 vanilla (Solaris): As soon as the R_LIBS_USER env var gets bigger than 1023 chars R completely ignores it and uses the default: Sys.getenv('R_LIBS_USER'); R_LIBS_USER ${R_LIBS_USER-~/R/i386-pc-solaris2.11-library/2.10} The strange thing is, if I set another envar to the same value, there seems to be no problem at all. E.g.: BLA=/develop/lnf/i386/R/LNFr-car/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-Formula/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-lmtest/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-sandwich/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-strucchange/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-zoo/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-dynlm/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-ineq/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-KernSmooth/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-np/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-plm/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-pscl/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-ROCR/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-sampleSelection/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-systemfit/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-tseries/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-urca/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-quantreg/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-m! vtnorm/reloc/R-2.10/library:/develop/lnf/i386/R/LNFr-leaps/reloc/R-2.10/library elkner.idev2 ~ echo $BLA | wc 1 11065 elkner.idev2 ~ R ... Sys.getenv('BLA') ... BLA /develop/lnf/i386/R/LNFr-car/reloc/R-2.10/library:...:/develop/lnf/i386/R/LNFr-leaps/reloc/R-2.10/library Any hints, what magic is going on here? Regards, jel. -- Otto-von-Guericke University http://www.cs.uni-magdeburg.de/ Department of Computer Science Geb. 29 R 027, Universitaetsplatz 2 39106 Magdeburg, Germany Tel: +49 391 67 12768 __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel