[R-SIG-Mac] [R] 64-bit OSX binary for 2.9.2

David Winsemius dwinsemius at comcast.net
Tue Sep 15 01:08:33 CEST 2009


On Sep 14, 2009, at 6:34 PM, Loren Engrav wrote:

> I too would vote for a 1 2 3 set of install instructions, maybe for  
> 32bit
> and 64bit if they differ

If you install Urbanek's, you get both, no choice. If you install  
CRAN's, R you get only 32 bit.

When run from the command line, the default is 32 bit. With  
(Urbanek's)  R --arch=x86_64 , you get 64 bit R running. Or you can  
use the 32 bit or 64 bit GUI's, just not both at the same time.  
Obviously the 64 bit GUI will _not_ make the CRAN R into a 64 bit  
version.

>
> I did the 123 below and
>
> 1-the were many messages during install like " x
> Library/Frameworks/R.framework/Versions/2.9/Resources/share/texmf/:  
> Can't
> update time for
> Library/Frameworks/R.framework/Versions/2.9/Resources/share/texmf:  
> Operation
> not permitted, but
>
> 2-2.9.2 seems to function OK after simple testing;

I think this means ... success.

> is that warning insignificant?

I have learned to ignore the ...." chmod failed ... operation not  
permitted" warnings. Happens all the time when building packages from  
source and seems to have no ill consequences. Not sure about the  
"Can't update time errors".

>
> 3-R.app available at http://r.research.att.com/ is 5426 but 5455 32  
> and 64
> bit available at
> http://r.research.att.com/R-GUI-5455-2.9-leopard-Leopard.dmg as per  
> 8/6/09
> r-sig-mac discussion so I put in 5455; but why 5426 here and 5455  
> there?

Can't help you there. I have the 2.9.0 GUI [R.app GUI 1.29 (5463)  
x86_64-apple-darwin9.7.0] running with the R 2.9.1 Patched and am not  
having noticeable difficulties. My guess is that the "right" way to do  
this would be to use 5426 with R 2.9.2

Note to web page maintainer. The labels on the R-GUI's must be wrong,  
they say "R 2.1.x". Is that supposed to say R 2.9.x and 2.10.x?

>
> Thank you
>
> Loren Engrav, MD
> U Washington
>
>
>> From: David Winsemius <dwinsemius at comcast.net>
>> Date: Mon, 14 Sep 2009 13:00:52 -0400
>> To: Steve Lianoglou <mailinglist.honeypot at gmail.com>
>> Cc: r-help <r-help at stat.math.ethz.ch>, ivo welch  
>> <ivo_welch at brown.edu>,
>> "r-sig-mac at stat.math.ethz.ch" <r-sig-mac at stat.math.ethz.ch>
>> Subject: Re: [R-SIG-Mac] [R] 64-bit OSX binary for 2.9.2
>>
>>
>> On Sep 14, 2009, at 11:59 AM, Steve Lianoglou wrote:
>>
>>> Hi,
>>>
>>> On Sep 14, 2009, at 11:44 AM, ivo welch wrote:
>>>
>>>> dear R wizards:  I am looking for a binary package distribution of
>>>> R 2.9.2
>>>> for OSX .  Looking at http://r.research.att.com/ , there seems to
>>>> be only a
>>>> binary for 2.9.0 .  is the 2.9.2 version binary package available
>>>> somewhere?   (at this point, would it make sense to elevate the 64-
>>>> bit
>>>> version to a "standard recommended" rather than just a "boutique"
>>>> version?)
>>>
>>> There isn't a *.pkg / *.dmg install for it, but it's there. In this
>>> section:
>>>
>>> Universal nightly builds for Mac OS X (10.4 Tiger and 10.5 Leopard)
>>>
>>> Third row in the table:
>>>
>>> http://r.research.att.com/R-2.9-branch-leopard-universal.tar.gz
>>>
>>> You have to inflate it at the root of your HD. The text right below
>>> the d/l table tells you how to do that:
>>>
>>> """Alternatively you can use the tar-ball (*.tar.gz) which must be
>>> unpacked in the root (e.g. tar fvxz R*.tar.gz -C /), but doesn't
>>> contain the GUI (see below for a separate download)."""
>>
>> The problem may arise for  *NIX-newbies ( a group in which I still
>> claim membership)  that they are unable to operationalize that
>> message. For one thing they may not even know how to open a Terminal
>> window. For another they may not know that the "/" at the end of that
>> command will take care of "unpacking at the root", so they will not
>> need to navigate away from the place that they find themselves when
>> Terminal opens up a bash session.  The steps broken down would be
>> something along these lines:
>>
>> First : Download the nightly build, and select "Save to file" rather
>> than opening with Stuffit, and then it by default should end up in  
>> the
>> <user>/Downloads/ folder.
>>
>> 1. a-c) Go to the Utilities folder of the Applications folder and
>> double-click on Terminal.app
>> 2. a) Go back to the browser and select the string:  tar fvxz
>> R*.tar.gz -C /
>>
>> (I would solicit some input here from more competent UNIXers  
>> regarding
>> whether one would ever need to prefix this command with sudo.)
>>
>> 2. b) cmd-C to put it in the clipboard
>> 2. c) click on the Terminal window and cmd-V to paste-in that  
>> command.
>> 2 d) use the arrow and backspace keys (since mouse-clicking does not
>> change the cursor position in Terminal sessions) to delete the
>> "R*.tar.gz" on the command line.
>>
>> 3 a) Navigate with the Finder to the Downloads/ folder and locate the
>> R<build>. Click-drag that file to the Terminal window and
>> "unclick" (or is it lift-click) when the pointer is at the place on
>> the command line where the "R*.tar.gz" used to be. A full path  
>> version
>> of the file name will get inserted (an no quotes are needed, unlike
>> what one might expect after working in R regularly.)
>>
>> On my machine, one now sees this on the command line:
>> tar fvxz /Users/davidwinsemius/Downloads/R-2.9-branch-leopard-
>> universal.tar.gz  -C /
>>
>> 4) Hit <return>
>>
>> Sit back and watch the show.
>>
>> You can the separately download the appropriate R-GUI and put in in
>> the /Applications/ folder.
>>
>>>
>>> You'll have to install the R.app GUI separately if you want to use
>>> the 2.9.2 install, otherwise I think the 2.9.0 *.pkg installer
>>> includes it, though I'm not sure.
>>
>> It does, In fact, it has both the 32 bit GUI and the 64 bit GUI. What
>> I did was install that package _first_ which gave me 2.9.0 and then
>> performed the steps above. You really would not want to do it in a
>> reversed order, since you would be replacing R 2.9.2 with R 2.9.0.
>>
>> And of course this message _should_ have gone to r-sig-mac at stat.math.ethz.ch
>>  and any follow up should trim the r-help address.
>>
>> -- 

David Winsemius, MD
Heritage Laboratories
West Hartford, CT



More information about the R-SIG-Mac mailing list