[R] Still have problems with tcltk in R 64 bit

Uwe Ligges ligges at statistik.tu-dortmund.de
Wed Jun 15 15:32:03 CEST 2011


Well, the R code in package tcltk for startup under Windows is, as you 
could have found out yourself easily:

.onLoad <- function(lib, pkg)
{
     packageStartupMessage("Loading Tcl/Tk interface ...",
                           domain = "R-tcltk", appendLF = FALSE)
     if(!nzchar(tclbin <- Sys.getenv("MY_TCLTK"))) {
         tclbin <- file.path(R.home(), "Tcl",
                             if(.Machine$sizeof.pointer == 8) "bin64" 
else "bin")
         if(!file.exists(tclbin))
             stop("Tcl/Tk support files were not installed", call.=FALSE)
         if(.Machine$sizeof.pointer == 8) {
             lib64 <- gsub("\\", "/", file.path(R.home(), "Tcl", "lib64"),
                           fixed=TRUE)
             Sys.setenv(TCLLIBPATH = lib64)
         }
     }
     library.dynam("tcltk", pkg, lib, DLLpath = tclbin)
     .C("tcltk_start", PACKAGE="tcltk")
     addTclPath(system.file("exec", package = "tcltk"))
     packageStartupMessage(" ", "done", domain = "R-tcltk")
     invisible()
}


This tells us that if you do not have "MY_TCLTK" defined on startup of 
R, you probably forgot to select the tcltk files for 64 bit from the 
installer when installing your version of R.

Uwe Ligges




On 15.06.2011 14:28, Arnaud Mosnier wrote:
> I agree that this is a really outdated source but I did not find the
> way to tell R using correctly the tcl version included (at least for
> the 64 bit version).
> If I remove the environment variables, things work for R 32 bit (it
> uses the tcl version included), but it does not work in R 64 bit.
>
> Where are the configuration files used to define the path to each tcl version ?
>
> Arnaud
>
> 2011/6/14 Uwe Ligges<ligges at statistik.tu-dortmund.de>:
>>
>>
>> On 14.06.2011 22:01, Arnaud Mosnier wrote:
>>>
>>> I achieve to make tcltk work on R 64 installing Active tcltk8.5 64bit
>>> version then setting windows environment variables as in
>>> http://www.sciviews.org/_rgui/tcltk/InstallRTclTk.html.
>>
>> Don't read outdated sources but the manuals.
>>
>> The R binary distribution comes with tcltk under Windows (in ${R_HOME}/tcl)
>> for both 32-bit and 64-bit and will user a different tcl if you set
>> environment variables.
>>
>> Hence the easiest thing is just to tell R not to use your otehrwise set
>> environment variabes and use its own tcl version.
>>
>> Uwe Ligges
>>
>>
>>
>>> But now, it uses only this 64 bit version and thus do not work anymore
>>> in R 32 bit !
>>>
>>> In my case, it solves my problem as I will probably use only R 64bit
>>> but I do not like to end with an half solution.
>>>
>>> Arnaud
>>>
>>> 2011/6/14 Peter Langfelder<peter.langfelder at gmail.com>:
>>>>
>>>> On Tue, Jun 14, 2011 at 12:47 PM, Adrienne Wootten<amwootte at ncsu.edu>
>>>>   wrote:
>>>>>
>>>>> Taking a quick look for it, it seems that they have replaced it with
>>>>> tcltk2.  I just did the installation with the same version in windows
>>>>> and it auto loaded the tcltk package and I never installed that
>>>>> package to begin with.  I would try it with tcltk2 and see if you get
>>>>> the package to install appropriately.  I'm not sure why tcltk isn't on
>>>>> CRAN anymore, it makes no sense not to have both tcltk and tcltk2, but
>>>>> here's hoping this helps you out.
>>>>>
>>>>> A
>>>>
>>>> FWIF, tcltk still exists but is now part of the standard R
>>>> distribution ("core packages" is the term I think?) and as such is
>>>> installed automatically when you install R. Therefore it is not
>>>> available from CRAN.
>>>>
>>>> Peter
>>>>
>>>
>>> ______________________________________________
>>> 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