Skip to content

‘package ‘XML’ is not installed for 'arch=x86_64'’

16 messages · Davor Cubranic, Simon Urbanek, John Maindonald +5 more

Original
Ricardo.Julio.Rodriguez.Fernandez at sergas.es
# 
Hi,

I'm trying to confirm I understand how packages are installed and manage in R to be able to better interpret error/warning messages.

In a Mac OS X installation of R I get this error:
Loading required package: XML
Failed with error:  ?package ?XML? is not installed for 'arch=x86_64'?

But the graphical interface R Package Installer says that version 3.9-4 is installed.

Please, where am I wrong? Thanks!
R version 2.15.1 (2012-06-22)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

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

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

other attached packages:
[1] foreign_0.8-50

loaded via a namespace (and not attached):
[1] tools_2.15.1
--
Ricardo Rodr?guez
Research Management and Promotion Technician
Technical Secretariat
Health Research Institute of Santiago de Compostela (IDIS)
http://www.idisantiago.es

Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm
Davor Cubranic
# 
Maybe it is installed only for the 32-bit arch?

Davor
On 2012-09-27, at 3:25 PM, <Ricardo.Julio.Rodriguez.Fernandez at sergas.es> <Ricardo.Julio.Rodriguez.Fernandez at sergas.es> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Simon Urbanek
    

    

      
      #
    

  


  

      
On Sep 27, 2012, at 6:25 PM, <Ricardo.Julio.Rodriguez.Fernandez at sergas.es> wrote:

        

            

      
      
      
    

        
It is installed, but only for 32-bit and you are trying to load it in 64-bit R.

            

      
      
      
    

        
Use 32-bit R or if you want to use XML in 64-bit R, you have to compile it for 64-bit. AFAICS XML currently doesn't pass checks on any platform, so there are no binaries and you have to compile it from sources. If you want both 32-bit and 64-bit you'll have to use --merge-multiarch INSTALL option or use the two-step process described in the Mac FAQ.

Cheers,
Simon

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      John Maindonald
    

    

      
      #
    

  


  

      
That is odd. I can do:

            

      
      
      
    

        
R version 2.15.1 (2012-06-22)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

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

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

other attached packages:
[1] XML_3.9-4

loaded via a namespace (and not attached):
[1] tools_2.15.1

Also under help(package='XML'), go to the Description file, I see:
  Date/Publication: 2012-01-31 07:37:18
  Built: R 2.15.0; universal-apple-darwin9.8.0; 2012-03-31 00:12:17 UTC; unix
  Archs: i386, ppc, x86_64

I've not compiled it.  But maybe there is a risk that if I try to do anything with
it, it will fail?   Is the difference that I have a 2.15.0 build?

John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.
http://www.maths.anu.edu.au/~johnm

        
On 28/09/2012, at 10:24 AM, Simon Urbanek <simon.urbanek at r-project.org> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Dan Tenenbaum
    

    

      
      #
    

  


  

      
On Thu, Sep 27, 2012 at 6:32 PM, John Maindonald

        
<john.maindonald at anu.edu.au> wrote:

            

      
      
      
    

        
There is a Mac binary for XML on CRAN:
http://cran.r-project.org/web/packages/XML/index.html

Dan

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Ricardo.Julio.Rodriguez.Fernandez at sergas.es
    

    

      
      #
    

  


  

      
Hi!

        

      
      
    

        

            

      
      
      
    

        
I've installed it twice from R Console, once launching it with R and once with R64, by using the graphical R Package Installer and picking up CRAN binaries as repository.

It seems that some thing failed following this way or I've done something the wrong way. Downloading and installing the binary available at CRAN as Dan Tenenbaum suggested did the trick. Now I get XML installed both in 32 and 64 bits consoles.

            

      
      
      
    

        
R version 2.15.1 (2012-06-22)
Platform: i386-apple-darwin9.8.0/i386 (32-bit)

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

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

other attached packages:
[1] XML_3.9-4      foreign_0.8-50

            

      
      
      
    

        
R version 2.15.1 (2012-06-22)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

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

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

other attached packages:
[1] XML_3.9-4      foreign_0.8-50

            

      
      
      
    

        
Please, must we expect to get it installed for both architectures by installing it from the graphical interface? Where am I lost?

Thanks for helping me to better understand R framework!

Ricardo

            

      
      
      
    

        

      
      
    

        
Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm

  
  


  


      

    

    

      
      

        


  

        

      Ricardo.Julio.Rodriguez.Fernandez at sergas.es
    

    

      
      #
    

  


  

      
Hi!

        

      
      
    

        

            

      
      
      
    

        
Following other replies to this thread this is what I get here after installing from the command line without switches the XML binary available at CRAN...

Package: XML
Version: 3.9-4
ReleaseDate: 2012/01/30
...
Repository: CRAN
Date/Publication: 2012-01-31 07:37:18
Built: R 2.15.0; universal-apple-darwin9.8.0; 2012-03-31 00:12:17 UTC; unix
Archs: i386, ppc, x86_64

So, it seems that i386 and x86_64 are available by installing that binary. Please, what am I misunderstanding here?

Thanks for your help!

Ricardo

            

      
      
      
    

        
Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm

  
  


  


      

    

    

      
      

        


  

        

      Brian Ripley
    

    

      
      #
    

  


  

      
With a package such as XML available from multiple repositories, you 
have to be very precise what you are talking about.

On CRAN, there is a source package for version 3.9-4.  That can be 
compiled for multiple architectures on a Mac, and there are such 
binaries on CRAN.

It only works in a UTF-8 locale, and hence there has needed to be a 
modified version for Windows, on CRANextras.

However, the home for XML is www.omegahat.org.  There are two different 
versions there, and neither work on Linux.  (3.93-0 in the Omegahat 
repository and http://www.omegahat.org/RSXML/XML_3.94-0.tar.gz.)

The CRAN team have been trying for weeks to sort this out, but the XML 
maintainer is barely responsive.

With a CRAN build of R for MacOS, you get binary packages by default, 
either from the command-line or from R.app.   You get those of the 
achitecture(s) which were built: this is normally both ix64 and x64_86 
(and perhaps ppc) but for a few tens of packages the person in charge 
(for CRAN, Simon Urbanek) needs to set up exceptions.  As packages do 
change, sometimes the need to add an exception pops up unexpectedly.

        
On 28/09/2012 07:05, Ricardo.Julio.Rodriguez.Fernandez at sergas.es wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Ricardo.Julio.Rodriguez.Fernandez at sergas.es
    

    

      
      #
    

  


  

      
Thanks Brian, thanks everybody. I think I can easily understand the difficulties you are speaking about: R is a huge project running in a massive amount of different, and from time to time, fancy environments. I is by no means easy to keep all packages smoothly running in all and every installation around the World! For me, maintaining CRAN is a big responsibility and of extreme difficulty!

But I'm not able to understand yet what is the difference between using the R Package Installer (using CRAN repository as package repository), running in this case R Mac OS X 10.5.8  with this R installation...

R version 2.15.1 (2012-06-22) -- "Roasted Marshmallows"
Copyright (C) 2012 The R Foundation for Statistical Computing
ISBN 3-900051-07-0
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

And downloading XML_3.9-4.tgz from http://cran.r-project.org/web/packages/XML/index.html and installing it by submitting sudo R CMD INSTALL XML_3.93-0.tgz.

The process using R Package Installer (CRAN binaries as repositoy) doesn't lead to a functional XML packages and downloading and installing it does lead to running XML packages both in i386 and x86_64 architectures.

Sorry for asking once again, but, please, where am I lost with this issue?

Thanks!

        

      
      
    

        

        
On 28/09/2012 07:05, Ricardo.Julio.Rodriguez.Fernandez at sergas.es wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        
--
Brian D. Ripley,                  ripley at 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

_______________________________________________
R-SIG-Mac mailing list
R-SIG-Mac at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-mac

Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm

  
  


  


      

    

    

      
      

        


  

        

      John Maindonald
    

    

      
      #
    

  


  

      
Others may be able to give a more authoritative answer.  I doubt that 
the method of installation was per se the difference.  I installed using 
the Package installer, on April 6 2012.  The publication date 
(2012-01-31 07:37:18) was the same as for the package that you 
installed by first downloading.  You may have been unlucky that 
something went wrong when you installed using the Package Installer.   
Or, just possibly, a different file was temporarily on CRAN when you 
used the package installer.

John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.
http://www.maths.anu.edu.au/~johnm

        
On 29/09/2012, at 10:13 AM, Ricardo.Julio.Rodriguez.Fernandez at sergas.es wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Simon Urbanek
    

    

      
      #
    

  


  

      
On Sep 28, 2012, at 9:05 PM, John Maindonald wrote:

        

            

      
      
      
    

        
It think it's actually more simple than that - my guess is that  Ricardo has (apparently inadvertently) installed the *source* package and not the binary which is why he got only one architecture. The Package Installer allows the user to do either depending what the user selects.

Cheers,
Simon

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Ricardo.Julio.Rodriguez.Fernandez at sergas.es
    

    

      
      #
    

  


  

      
I bet Simon is right. Two main reasons:

1. I was following a blog post on how to set up FastRWeb on Mac OS X. This post (http://www.r-bloggers.com/setting-up-fastrweb-on-mac-os-x/) speaks about installing from sources and, even though "I think" I move to binaries prior to posting here on this issue, I can't be sure about that.
2. My understanding about multiple sub-architectures is more than poor. So, I could be sure I used the correct terms and concepts when analyzing the issue I was facing here.

I always try to stick to "official" sources of information when studying how to use a given R piece of software, but Google and other search engines put tones of other documents a couple of clicks away and is by no means easy to decide which one deserves to be followed. I think many of them can be useful when read with enough background as to be able to separate noise and information.

Please, allow me to hijack this thread to put in another, for me, related question I'm trying to work out on the list. To the best of my understanding and by following the documentation when I install R binary for Mac OS X both i386 and x86_64 are available. In fact I can easily choose among them by launching R Console by click on R or R64 "icons". But I'm not able to find the switch I must use to get the same thing when running R from the command line: I always get x86_64. I'm not able to solve this question by following R for Mac OS X FAQ or other sources of information available from CRAN.

Sorry for the noise I introduced in this list. Thanks to everybody!

Ricardo

        

      
      
    

        
Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm

  
  


  


      

    

    

      
      

        


  

        

      R. Michael Weylandt
    

    

      
      #
    

  


  

      
On Sat, Sep 29, 2012 at 10:58 AM,

        
<Ricardo.Julio.Rodriguez.Fernandez at sergas.es> wrote:

            

      
      
      
    

        
For me, R32 and R64 do the job. They're simlinked to

R* -> /Library/Frameworks/R.framework/Resources/bin/R*

where * is 32 or 64. I don't remember setting those up, so they
probably happen automatically if you use the CRAN installer.

Cheers,
Michael

  
  


  


      

    

    

      
      

        


  

        

      Berend Hasselman
    

    

      
      #
    

  


  

      
On 29-09-2012, at 11:58, Ricardo.Julio.Rodriguez.Fernandez at sergas.es wrote:

        

            

      
      
      
    

        
For 32-bit R in Terminal

R32
R --arch=i386

For 64bit R in Terminal

R
R64
R --arch=x86_64

Berend

  
  


  


      

    

    

      
      

        


  

        

      Brian Ripley
    

    

      
      #
    

  


  

      
On 29/09/2012 11:11, R. Michael Weylandt wrote:

            

      
      
      
    

        
True (but you may have other scripts of that name earlier in your path: 
I do from before this was done by the installer).

But the documented way is R --arch i386 (or x86_64).

Which is the default sub-architecture depends on the version of OS X 
which was used for installation: from memory this is i386 for Leopard 
and x86_64 since.

Please note that we do need to know the OS X version for questions like 
this: what the installer does depends on the version.

  
    

      
      
    

  


  


      

    

    
1 day later

    

      
      

        


  

        

      Ricardo.Julio.Rodriguez.Fernandez at sergas.es
    

    

      
      #
    

  


  

      
Bared, Michel, Simon, all,

            

      
      
      
    

        
In this OS X 10.5.8 box running R version 2.15.1 (2012-06-22), these options launch R with 32bits architecture:

R
R --arch i386
R --arch=i386

**There is not a R32 symlink here.

These ones launches R with 64bits architecture:

R64
R --arch x86_64
R --arch=x86_64

            

      
      
      
    

        
Scripts are devil here! This thread has its origin in an issue while running a R script used to launch FastRWeb. I'm not able to find now where I read that it only works in a 32 bits environment. As far as the script launching it, sudo /var/FastRWeb/code/start, runs in 64bits mode, my guess was that the problems to put it up and running where related with this issue.

The point is that I have a fully operative RFastWeb installation running locally in my MacBook Pro running Mac OS X 10.5.8. Simply great! Absolutely great! Thanks all for your help.

But I'm still struggling to understand why /var/FastRWeb/code/start launches in 64bits mode. Something in the related scripts, but I'm not able to find it.

            

      
      
      
    

        
I can confirm this is correct for Leopard (10.5.x); I'll check the default for Snow Leopard and beyond when regain access to some now not available remote boxes. Thanks!

            

      
      
      
    

        
Indeed! I'm currently moving back and forth to/from two Mac systems running OS X 10.5.8 and 10.6.8. **All the information posted in this thread concerns only 10.5.8 box**. In fact I'm facing some behavior in the 10.6.8 that doesn't fit what I'm seeing in the 10.5.8 one.

--
Brian D. Ripley,                  ripley at 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

Nota: A informaci?n contida nesta mensaxe e os seus posibles documentos adxuntos ? privada e confidencial e est? dirixida ?nicamente ? seu destinatario/a. Se vostede non ? o/a destinatario/a orixinal desta mensaxe, por favor elim?nea. A distribuci?n ou copia desta mensaxe non est? autorizada.

Nota: La informaci?n contenida en este mensaje y sus posibles documentos adjuntos es privada y confidencial y est? dirigida ?nicamente a su destinatario/a. Si usted no es el/la destinatario/a original de este mensaje, por favor elim?nelo. La distribuci?n o copia de este mensaje no est? autorizada.

See more languages: http://www.sergas.es/aviso_confidencialidad.htm