[R] package installation from source
Duncan Murdoch
murdoch at stats.uwo.ca
Mon Oct 19 12:24:47 CEST 2009
Petr PIKAL wrote:
> Hi
>
> Duncan Murdoch <murdoch at stats.uwo.ca> napsal dne 16.10.2009 17:01:59:
>
>
>> On 10/16/2009 9:31 AM, Petr PIKAL wrote:
>>
>>> Dear all
>>>
>>> I noticed from NEWS 2.11.0,dev
>>> SIGNIFICANT USER-VISIBLE CHANGES
>>>
>>> o Packages must have been installed under R 2.10.0 or later, as
>>> the current help system is the only one now supported.
>>>
>>> So I tried to follow instructions in manual, Duncan Murdoch
>>>
> presentation
>
>>> and help pages to prepare and accomplish installation of a set of
>>> functions I use. However in R 2.11.0dev and too in R 2.10.0dev I was
>>>
> not
>
>>> able to finish it.
>>>
>>>
>
> < snip>
>
>
>>>> install.packages("D:/temp/fun", repos=NULL, type="source")
>>>>
>>> sh nenˇ n zvem vnitýnˇho ani vnŘjçˇho pýˇkazu,
>>> spustiteln‚ho programu nebo d vkov‚ho souboru.
>>> Warning message:
>>> In install.packages("D:/temp/fun", repos = NULL, type = "source") :
>>> installation of package 'D:/temp/fun' had non-zero exit status
>>>
>>> which basically tells that sh is not a name of command, exe or bat
>>>
> file.
>
>> There should be a sh.exe in d:\programy\Rtools\bin if that's where you
>> installed the Rtools. Are you sure you have the path set the way you
>>
>
> Yes there is it
>
>
>> list.files("D:/programy/Rtools/bin")
>>
> [1] "basename.exe" "cat.exe" "cmp.exe"
>
> < snip>
> 40] "sh.exe" "sort.exe" "tar.exe"
> [43] "texindex.exe" "touch.exe" "tr.exe"
> [46] "uniq.exe" "unzip.exe" "wc.exe"
> [49] "zip.exe"
>
>
>
>> think you do? Remember that PATH is an environment variable, and
>> environment variable settings are local: so setting the PATH in a CMD
>> window has no effect on R or any other command window. You need to use
>>
>
> Hm. It is a problem. I do not have admin rights. Therefore I set PATH in
> the command window and used the same window to issue R CMD INSTALL fun,
> but it gave me the same error message as above.
>
That should have worked. Just to confirm, try running R from that shell
(using either Rgui or R), and then print Sys.getenv("PATH") and make
sure it shows what you expected.
>
>> the Control Panel to set the default system path, and then restart any
>> process that needs to see it.
>>
>
> As I can not change system wide PATH I tried to set user PATH. This
> resulted in modified PATH value but the required values for installation
> was at the end and I got the same error.
>
Yes, I believe Windows always puts the user PATH after the system one.
> As I am not sure if our IT guys change the system path for me
> (permanently) I would like to know if there is some workaround like few
> years before with make, make install etc. Maybe more complicated but
> without need to fiddle with PATH value.
>
>
I think you'll need to fiddle with the PATH, but maybe not at the system
level. I would do what you did (setting it in the command window). You
just need to figure out what went wrong.
Duncan Murdoch
> Thank you and best regards
> Petr
>
>
>> Duncan Murdoch
>>
>>
>>> The same I get when I try to run
>>>
>>> D:\temp R CMD INSTALL fun
>>>
>>> Please can you give me any suggestion for making installation work. I
>>> remember when going from 1.9 to 2.0 versions it was also necessary to
>>> install my bunch of functions, however at that time I need to use
>>> something like make, make install or so and gather all necessary
>>>
> programs
>
>>> myself.
>>>
>>> Best regards
>>> Petr
>>>
>>>
>>>
>>>
>>>
>>> ______________________________________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide
>>>
> http://www.R-project.org/posting-guide.html
>
>>> and provide commented, minimal, self-contained, reproducible code.
>>>
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
More information about the R-help
mailing list