[R-SIG-Mac] mac-builder

[Prof Brian Ripley]

On Tue, 10 Mar 2009, Sundar Dorai-Raj wrote:

> Thanks for the replies from both Simon and Brian.
>
> Brian, I have one clarification question about your comment below. You
> say,  "Mac binaries are tied to 2.x.* for one x...." Does this mean
> that if I build a Mac binary using R-2.8.*, then someone using 2.7.*
> would not be able to install and/or load it?

Correct (if it contains compiled code).  E.g. I have

tystie% otool -L tree/libs/tree.so
ree/libs/i386/tree.so:
 	tree.so (compatibility version 0.0.0, current version 0.0.0)
 	/Library/Frameworks/R.framework/Versions/2.8/Resources/lib/libR.dylib 
(compatibility version 2.8.0, current version 2.8.0)

and so it needs at least 2.8.0.

There are lots of other differences: only the Leopard builds have the 
64-bit architectures, for example.

I've tended to just install my own packages from sources: that way 
always works for me.


>
> Thanks,
>
> --sundar
>
> On Mon, Mar 9, 2009 at 12:13 PM, Prof Brian Ripley
> <ripley at stats.ox.ac.uk> wrote:
>> On Mon, 9 Mar 2009, Sundar Dorai-Raj wrote:
>>
>>> Hi, all,
>>>
>>> Uwe Ligges has a nice tool for creating R binaries for Windows.
>>>
>>> http://win-builder.r-project.org/
>>>
>>> Is there an equivalent for Mac to create the tgz file?
>>
>> It is currently tricky: packages built on MacOS by you or I under 2.8.1 are
>> likely to depend on things in /usr/local, and there are OS differences
>> (Tiger vs Leopard) and the Mac binaries are tied to 2.x.* for one x (unlike
>> the Windows ones).  Simon tells me he has now fixed some of the issues, so
>> hopefully this will be possible in future.
>>
>> It is I am told possible to use R-forge to do this (but I have two Macs
>> amongst my menagerie, so I don't need to find out).
>>
>> But winbuilder is a great initiative from Uwe (with some help from his
>> friends), so all credit to him.  Stefano's original Mac package building
>> service was based on Uwe's work too (same rider applies).
>>
>> Brian
>

-- 
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