Yes, that worked, thanks a lot! Looking forward to the 2.11.0 release.

> install.packages("~/Projects/R 
> Library/potsdam/eyetrackR/eyetrackR_0.13.tar.gz", repos = NULL, type = 
> "source")
* installing *source* package ‘eyetrackR’ ...
** R
** data
** demo
** preparing package for lazy loading
** help
*** installing help indices
** building package indices ...
** testing if installed package can be loaded
* DONE (eyetrackR)
> library(eyetrackR)

> sessionInfo()
R version 2.11.0 alpha (2010-03-26 r51420) 
x86_64-apple-darwin10.2.0 

locale:
[1] de_DE.UTF-8/de_DE.UTF-8/C/C/de_DE.UTF-8/de_DE.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base


btw, this is probably known, but ... the gfortran in /usr/bin (gcc version 
4.2.1 (Apple Inc. build 5646)) did not work (complaining about 
/usr/lib/libgfortran.dylib, missing required architecture x86_64), but a 
different gfortran located in /usr/local/bin did (gcc version 4.2.3, probably 
from Simon Urbanek's page)

On Mar 26, 2010, at 11:51 , Prof Brian Ripley wrote:

> So please try 2.11.0 alpha, as I believe this is already fixed (not least, 
> gzip is not called on that platform).  Also, I suspect setting the envir 
> variable TAR to 'internal' would work on 2.10.1 (but my Mac is at home).
> 
> Another thing that does not work with spaces in file names is command-line 
> completion of filenames (in R, I did not try R.app).  If spaces in filenames 
> work (other than on Windows) it is somewhat accidental.  Reports of instances 
> where they do not, with patches against current sources, would be considered 
> but as low priority.
> 
> BTW, 'install' works on case-insensitive file systems, but it is likely to 
> stop working at any time: as on Windows, 'R CMD' is being moved away from 
> Perl/sh scripts to direct use of R.
> 
> On Fri, 26 Mar 2010, jochen laubrock wrote:
> 
>> I am sorry, this was from 2.10.1 on Mac OS X 10.6, so it might be 
>> Mac-specific.
> 
> I did try it on a Mac, but I did not try 2.10.1, since the 'upgrade before 
> posting' request in the posting guide did apply, and we are busy enough with 
> 2.11.0 alpha.
> 
> The behavior is reproducible both from the shell and the GUI version of R 
> (sessionInfo output below). It does not seem to depend on the particular 
> package: it can be reproduced by downloading an archive of a package from 
> CRAN to a path containing spaces and calling
>> install.packages("path/with spaces/downloadedPackage.tar.gz", repos = NULL, 
>> type = "source"):
>> 
>>> print(dl <- download.packages("gam", destdir="~/Projects/R library/tmp"))
>> [...] downloaded 300 Kb
>>    [,1]  [,2]
>> [1,] "gam" "~/Projects/R library/tmp/gam_1.01.tgz"
>>> install.packages(dl[,2], repos = NULL, type = "source")
>> [...]
>> gzip: /Users/jochen/Projects/R.gz: No such file or directory
>> gzip: library/tmp/gam_1.01.tgz: No such file or directory
>> ERROR: cannot extract package from ‘/Users/jochen/Projects/R 
>> library/tmp/gam_1.01.tgz’
>> 
>> # /usr/bin/R
>>> sessionInfo()
>> R version 2.10.1 (2009-12-14)
>> x86_64-apple-darwin9.8.0
>> 
>> locale:
>> [1] de_DE.UTF-8/de_DE.UTF-8/C/C/de_DE.UTF-8/de_DE.UTF-8
>> 
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods   base
>> 
>> # R.app (RAqua)
>>> sessionInfo()
>> R version 2.10.1 (2009-12-14)
>> i386-apple-darwin9.8.0
>> 
>> locale:
>> [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
>> 
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods   base
>> 
>> 
>> 
>> On Mar 26, 2010, at 10:19 , Prof Brian Ripley wrote:
>> 
>>> You haven't told us your version of R (nor any of the other information 
>>> requested in the posting guide).  As far as I can see this works in 2.11.0 
>>> alpha.
> 
>> Does that mean that it has been re-designed to work with paths containing 
>> spaces on Unix-alikes?
> 
> No.  It is an unintentional side-effect of
> 
>    o   R CMD INSTALL now uses the internal untar() in package utils:
>        this ensures that all platforms can install bzip2- and
>        xz-compressed tarballs.
> 
>>> 
>>> On Fri, 26 Mar 2010, jochen laubrock wrote:
>>> 
>>>> There appears to be a quoting problem in the way R CMD install handles 
>>>> file names containing spaces, more specifically, in the way the argument 
>>>> is passed through to gzip.
>>>> 
>>>> The install.packages command
>>>> 
>>>> (from R)
>>>> install.packages("~/Projects/R library/bar/eyetrackR_0.13.tar.gz", repos = 
>>>> NULL, type = "source")
>>>> 
>>>> expands to
>>>> 
>>>> system("R_LIBS='/Users/foo/Library/R/2.10/library' 
>>>> "/Library/Frameworks/R.framework/Resources/bin/R CMD INSTALL -l 
>>>> '/Users/foo/Library/R/2.10/library'   '/Users/foo/Projects/R\ 
>>>> library/bar/eyetrackR_0.13.tar.gz'")
>>> 
>>> So that is fine.
>>> 
>>>> 
>>>> and gives the same error messages as the following commands from 
>>>> Terminal.app on Mac OS X
>>>> 
>>>> (from bash)
>>>> R CMD install /Users/foo/Projects/R\ library/bar/eyetrackR_0.13.tar.gz
>>>> R CMD install "~/Projects/R library/bar/eyetrackR_0.13.tar.gz"
>>>> R CMD install '~/Projects/R library/bar/eyetrackR_0.13.tar.gz'
>>> 
>>> the documented command is INSTALL.
>> sorry
>>> 
>>>> , namely (the error messages):
>>>> 
>>>> gzip: /Users/foo/Projects/R.gz: No such file or directory
>>>> gzip: library/bar/eyetrackR_0.13.tar.gz: No such file or directory
>>> 
>>> Yes, and INSTALL is not designed to work with paths with spaces in on 
>>> Unix-alikes.
>>> 
>>>> 
>>>> The following commands do work
>>>> 
>>>> (from R)
>>>> setwd("~/Projects/R library/bar/")
>>>> install.packages("eyetrackR_0.13.tar.gz", repos = NULL, type = "source")
>>>> 
>>>> (from bash)
>>>> cd ~/Projects/R\ library/bar/
>>>> R CMD install eyetrackR_0.13.tar.gz
>>>> 
>>>> Interestingly, if the file is unpacked on the command line (tar xzvf), 
>>>> then both R CMD install and install.packages work fine using the quoted 
>>>> path name syntax, i.e.,
>>>> 
>>>> R CMD install /Users/foo/Projects/R\ library/bar/eyetrackR
>>>> install.packages("~/Projects/R library/bar/eyetrackR", repos = NULL, type 
>>>> = "source")
>>>> 
>>>> Is this a known problem?
>>> 
>>> It is a known restriction.
>>> 
>>>> I searched the archives, but did not find a decisive answer (only
>>> some rather old posts suggesting not to use path names containing 
>>> spaces--unfortunately this is not an option in the managed Windows 
>>> environment I need to work in next week).
>>>> 
>>>> Sorry for the long post and thanks for your time,
>>>> Jochen
>>>> 
>>>> ----
>>>> Jochen Laubrock, Dept. of Psychology, University of Potsdam,
>>>> Karl-Liebknecht-Strasse 24-25, 14476 Potsdam, Germany
>>>> phone: +49-331-977-2346, fax: +49-331-977-2793
>>>> 
>>>> ______________________________________________
>>>> R-devel@r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>>> 
>>> 
>>> --
>>> Brian D. Ripley,                  rip...@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 272866 (PA)
>>> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>> 
>> 
> 
> -- 
> Brian D. Ripley,                  rip...@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 272866 (PA)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595

----
Jochen Laubrock, Dept. of Psychology, University of Potsdam,
Karl-Liebknecht-Strasse 24-25, 14476 Potsdam, Germany
phone: +49-331-977-2346, fax: +49-331-977-2793

______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Reply via email to