[R-SIG-Mac] Package installation: setting default library
Simon Urbanek
simon.urbanek at r-project.org
Tue Jan 6 23:58:31 CET 2009
On Jan 6, 2009, at 5:20 PM, Carl Witthoft wrote:
> Ok, here is the response I get after R freshly booted via icon:
>
>
> Rgames: .libPaths()
> [1] "/Users/carl/Library/R/2.8/library"
> [2] "/Library/Frameworks/R.framework/Versions/2.8/Resources/library"
>
> -----
> then I selected Install local package directory, with System
> framework as target location
> ---
>
> Warning in install.packages("/Users/carl/Downloads/gtools", , NULL,
> type = "source") :
> argument 'lib' is missing: using '/Users/carl/Library/R/2.8/library'
> WARNING: ignoring environment value of R_HOME
> * Installing *binary* package 'gtools' ...
> * DONE (gtools)
> Rgames:
>
> So, it sort of looks like the PackageInstaller GUI window isn't
> filling in the right library path when I click the Install At System
> Level button.
>
Yes, thanks, that's what it looks like. In fact all three of the last
choices do the same thing which is incorrect in several ways. This is
a historical relic from times where we didn't have native
file.choose(). At any rate I think I have fixed that in the SVN, so
please try the tonight's build tomorrow.
Thanks,
Simon
> Or, is there some way to get .libPaths to list the paths in a
> different order or something like that?
>
>
> Simon Urbanek wrote:
>
>>> Using the R.app GUI, I selected "Install at System Level" and get
>>> an error message something like
>>>
>>> " 'lib' variable not found, using default {my local Library/
>>> Frameworks directory}..."
>>>
>> Can you be, please, more precise? "something like" doesn't really
>> help ...
>> Thanks,
>> S
>>> The default value of libPaths was
>>>
>>> [1] "/Users/carl/Library/R/2.8/library"
>>> [2] "/Library/Frameworks/R.framework/Versions/2.8/Resources/library"
>>>
>>> If I ran the command
>>>
>>> R> .libPaths(.libPaths()[2])
>>>
>>> {thus removing the reference to my local tree}
>>>
>>> then the GUI properly installed to the system library tree.
>>>
>>> So, is there a better, or simpler, way to make this happen?
>>>
>>> thanks
>>> Carl
>>>
>>> _______________________________________________
>>> R-SIG-Mac mailing list
>>> R-SIG-Mac at stat.math.ethz.ch
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>>
>>>
>
>
More information about the R-SIG-Mac
mailing list