[Phylobase-devl] [R-sig-phylo] Phylobase 0.4 Forge (installing issues in my old G4 powerBook)

Fri Mar 13 14:11:46 CET 2009

Hi,
	Take this with a grain of salt, because I've never done any  
development in R, but...

	I should be possible to compile and install NCL into a location on  
your system's LD_LIBRARY_PATH ( the DYLD_LIBRARY_PATH on Mac), such  
that phylobase picks up NCL from that location.  (note that  
LD_LIBRARY_PATH and DYLD_LIBRARY_PATH are environmental variables that  
can be set in your environment if you install in "non-standard"  
locations, if you install to /usr or /usr/local you usually do not  
have to set any variables).

	You'd have to make sure that you install a version of NCL that has  
all of the functions that phylobase uses, but that should not be hard  
to do (I'd pull the NCL out of the phylobase distribution, if I were  
trying this).

	If NCL itself (outside of phylobase) won't build on your PPC, then I  
should be able to help you with that.

	Perhaps a more experienced R user can chime in on whether installing  
to a system location is going to cause problem for an R package.

All the best,
Mark

On Mar 12, 2009, at 11:03 PM, Brian O'Meara wrote:

> This error seems to have come up with another package (gee), and the
> solution is setting an environmental variable. See thread at https://stat.ethz.ch/pipermail/r-sig-mac/2008-October/005450.html
>> . Basically do Sys.setenv("MACOSX_DEPLOYMENT_TARGET"="10.4") so that
> it targets 10.4 rather than 10.1.
>
> Brian
>
>
> On Mar 12, 2009, at 10:41 PM, Raul Sedano wrote:
>
>> Thanks Brian,
>> While your 4 steps did work in a Mac OSX 10.5, it did not work on my
>> old PPC 10.4.11. This is what I got from R (newest version) in my
>> old mac:
>>
>> After a long process of installation 10.4.11. systems, this is just
>> the last part in red
>>
>> ld: flag: -undefined dynamic_lookup can't be used with
>> MACOSX_DEPLOYMENT_TARGET environment variable set to: 10.1
>> /usr/bin/libtool: internal link edit command failed
>> make: *** [phylobase.so] Error 1
>> chmod: ** Removing '/Library/Frameworks/R.framework/Versions/2.8/
>> Resources/library/phylobase'
>> /Library/Frameworks/R.framework/Versions/2.8/Resources/library/
>> phylobase/libs/ppc/*: No such file or directory
>> ERROR: compilation failed for package 'phylobase'
>>
>> The downloaded packages are in
>> 	/private/tmp/RtmpcBvRaE/downloaded_packages
>> Updating HTML index of packages in '.Library'
>> Warning message:
>> In install.packages("phylobase", repos = "http://R-Forge.R-
>> project.org",  :
>>  installation of package 'phylobase' had non-zero exit status
>>
>> However, I was able to follow most of the Brian's four steps to
>> remove phylobase and then re-install it but source compilation
>> failed. Can you guys tell what is wrong with compilation? is this a
>> issue just with OSX 10.4.11?
>>
>> Raul Sedano
>>
>>
>> On Thu, Mar 12, 2009 at 3:13 PM, Brian O'Meara
>> <bcomeara at nescent.org> wrote:
>> Thanks, Peter and Roland, for  your emails. Installing from source
>> with install.packages("phylobase",repos="http://R-Forge.R-
>> project.org", type='source') worked. I think my problem earlier was
>> that while I was using "R CMD REMOVE phylobase", I didn't restart R
>> between attempted installs or unload and then load phylobase, so the
>> version being used was never updated in memory. Peter, the reason I
>> suspected a 10.4/10.5 was that my particular error, based on
>> googling for it, occurs with code compiled on 10.5 without
>> explicitly also being compiled for 10.4 (i.e., see <http://snipurl.com/crosscompile
>>> ).
>>
>> So, Raul, if you 1) Quit R, 2) Run "R CMD REMOVE phylobase" in
>> Terminal, 3) type "install.packages("phylobase",repos="http://R-Forge.R-project.org
>> ", type='source')" in R, 4) type "library(phylobase)", does
>> everything work?
>>
>> Brian
>>
>>
>>
>> On Mar 12, 2009, at 5:54 PM, Peter Cowan wrote:
>>
>> On Mar 12, 2009, at 1:00 PM, Brian O'Meara wrote:
>>
>> Well, using 10.4.11 on an intel mac, with developer tools installed,
>> running R 2.8.1, and got the error when trying to load phylobase
>> (install appeared to work, just error on load):
>>
>>> install.packages("phylobase",repos="http://R-Forge.R-project.org")
>> trying URL 'http://R-Forge.R-project.org/bin/macosx/universal/contrib/2.8/phylobase_0.4.tgz'
>> Content type 'application/x-gzip' length 663910 bytes (648 Kb)
>> opened URL
>> ==================================================
>> downloaded 648 Kb
>>
>>
>> The downloaded packages are in
>>       /tmp/RtmpFOnUD2/downloaded_packages
>>
>>> library(phylobase)
>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>> unable to load shared library '/Library/Frameworks/R.framework/
>> Resources/library/phylobase/libs/i386/phylobase.so':
>> dlopen(/Library/Frameworks/R.framework/Resources/library/phylobase/
>> libs/i386/phylobase.so, 6): Symbol not found:
>> Referenced from: /Library/Frameworks/R.framework/Resources/library/
>> phylobase/libs/i386/phylobase.so
>> Expected in: /usr/lib/libSystem.B.dylib
>>
>> Error in library(phylobase) : .First.lib failed for 'phylobase'
>>
>> So, it's not an issue of Raul not having developer tools (though I
>> get a slightly different error, I'm assuming it is the same root
>> cause). My error comes from the compiled version being made just for
>> Mac OS 10.5, not cross-compiled for 10.4. My error persists when
>> installing from source, though (after removing phylobase first),
>> both from the 0.4 release and the trunk.
>>
>> How can you tell it is due to the 10.4/10.5 difference.  I'm no
>> expert, but I would have guessed that they wouldn't differ when it
>> comes to building R packages.
>>
>> Brian when you run
>>
>> install.packages("phylobase",repos="http://R-Forge.R-project.org",
>> type='source')
>>
>> Do you get any errors?  Or is it when loading the package?
>>
>> Is this email helpful here:
>>
>> <http://lists.r-forge.r-project.org/pipermail/phylobase-devl/2008-July/000192.html
>>>
>>
>> We've run through this before with phylobase, I think the fix
>> involved a makefile, but I wonder how it got broken again.
>>
>> This might add impetus to update NCL, I think that the building
>> process might have been improved.
>>
>> peter
>>
>>
>> Brian
>>
>>
>> On Mar 12, 2009, at 12:32 AM, Peter Cowan wrote:
>>
>> Raul,
>>
>> I'm CCing r-sig-phylo and the phylobase-dev list as other people may
>> be having the same problem.
>>
>> On Mar 11, 2009, at 8:10 PM, Raul Sedano wrote:
>>
>> Hi Peter, thanks for your comments, I already tried your suggestion.
>> This is
>> what I got when using install.packages("phylobase", repos="
>> http://R-Forge.R-project.org")
>> Error in library(phylobase) : .First.lib failed for 'phylobase'
>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>> unable to load shared library
>> '/Library/Frameworks/R.framework/Versions/2.8/Resources/library/
>> phylobase/libs/ppc/phylobase.so':
>> dlopen(/Library/Frameworks/R.framework/Versions/2.8/Resources/
>> library/phylobase/libs/ppc/phylobase.so,
>> 6): Symbol not found: __ZN8NxsBlock16GetImpliedBlocksEv
>> Referenced from:
>> /Library/Frameworks/R.framework/Versions/2.8/Resources/library/
>> phylobase/libs/ppc/phylobase.so
>> Expected in: dynamic lookup
>>
>> This should have worked.  It works for me, however I'm using 10.5 on
>> an intel computer.  Guessing from the error message you're using a
>> PPC computer.
>>
>> Any ideas to solve this? I though the problem was related to the
>> lack of
>> compilers. I have tried multiple compilers including those you  
>> pointed
>> through the website.
>>
>> Well we know appear to have two problems, one is that the build from
>> R-forge doesn't work on your system (does any one on the list have a
>> PPC with Mac os X 10.4, to confirm this).  The other problem is that
>> it's not compiling either.
>>
>> Are you still getting the same errors even with the developer tools
>> installed?
>>
>> Peter
>>
>>
>> Raul Sedano
>>
>>
>>
>> On Wed, Mar 11, 2009 at 6:02 PM, Peter Cowan <pdc at berkeley.edu>  
>> wrote:
>>
>> Raul,
>>
>> Yes, you need to have the full developer tools package installed for
>> compiling to work.  It's installs a bunch of programs necessary to
>> build
>> from source.  Out of curiosity, is there a reason why you're trying  
>> to
>> install from source?
>>
>> You should be able to install the package using the following line
>> in R:
>>
>> install.packages("phylobase", repos="http://R-Forge.R-project.org")
>>
>> Peter
>>
>>
>>
>> On Mar 11, 2009, at 5:10 PM, Raul Sedano wrote:
>>
>> Hey Peter, I got a different output from the terminal. Do you have  
>> any
>> suggestion to correct this in my system? This might be the reason
>> for my
>> current problem installing phylobase.Raul:~ Milvago$ Make
>> -bash: Make: command not found
>> Raul:~ Milvago$
>>
>> Should I use installers from Apple Developer Tools package to solve
>> this
>> problem. Yes I did install gcc3.3 (but also tried gcc4.0) software
>> from
>> the
>> Xcode Tools /Applications/Installers in my PowerBook.
>>
>> Thanks in advance,
>>
>> Raul Sedano
>>
>> On Wed, Mar 11, 2009 at 4:33 PM, Peter Cowan <pdc at berkeley.edu>  
>> wrote:
>>
>> Raul,
>>
>> When you say you installed the compilers in the R-FAQ, which ones do
>> you
>> mean?  Did you install the tools individually or as the Apple
>> Developer
>> Tools package?  For the mac the best guide to compliers for R stuff,
>> is
>> here:
>>
>> <http://r.research.att.com/tools/>
>>
>> When you open the terminal and run the command "make" what do you  
>> get?
>>
>> I see:
>>
>> Macintosh-7:~ peter$ make
>> make: *** No targets specified and no makefile found.  Stop.
>> Macintosh-7:~ peter$
>>
>> Peter
>>
>>
>> On Mar 11, 2009, at 4:22 PM, Raul Sedano wrote:
>>
>> Hey,
>>
>> First, I want to thanks Phylobase developers for such flexible
>> application,
>> I look forward to squeeze phylobase capabilities. However I need some
>> advice
>> to install Phylobase 0.4 (source) in my old G4/10.4/mac. I have
>> installed compilers as indicated in R-FAQ but still does not work.  
>> The
>> same
>> error show up  in multiple newer PowerBook machines. I got the
>> following
>> R
>> "response" suggesting the lack of a Make command. I will appreciate
>> any
>> comments in such basic topic.
>>
>> Raul Sedano
>>
>>
>>
>> trying URL '
>> http://r-forge.r-project.org/src/contrib/phylobase_0.4.tar.gz
>> '
>> Content type 'application/x-gzip' length 771914 bytes (753 Kb)
>> opened URL
>> ==================================================
>> downloaded 753 Kb
>>
>> WARNING: ignoring environment value of R_HOME
>> * Installing *source* package 'phylobase' ...
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> /Library/Frameworks/R.framework/Resources/bin/config: line 1: make:
>> command
>> not found
>> checking for g++... no
>> checking for c++... no
>> checking for gpp... no
>> checking for aCC... no
>> checking for CC... no
>> checking for cxx... no
>> checking for cc++... no
>> checking for cl... no
>> checking for FCC... no
>> checking for KCC... no
>> checking for RCC... no
>> checking for xlC_r... no
>> checking for xlC... no
>> checking for C++ compiler default output file name... configure:
>> error:
>> C++
>> compiler cannot create executables
>> See `config.log' for more details.
>> ERROR: configuration failed for package 'phylobase'
>> ** Removing
>>
>> '/Library/Frameworks/R.framework/Versions/2.8/Resources/library/
>> phylobase'
>> ** Restoring previous
>>
>> '/Library/Frameworks/R.framework/Versions/2.8/Resources/library/
>> phylobase'
>>
>> The downloaded packages are in
>> /private/tmp/RtmpMpRHQe/downloaded_packages
>>
>>
>>
>>   [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-phylo mailing list
>> R-sig-phylo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> R-sig-phylo mailing list
>> R-sig-phylo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
>>
>>
>> ________________________________
>> Brian O'Meara
>> NESCent
>> Durham, NC
>> http://www.brianomeara.info
>> ________________________________
>>
>>
>>
>>
>>
>>
>> ________________________________
>> Brian O'Meara
>> NESCent
>> Durham, NC
>> http://www.brianomeara.info
>> ________________________________
>>
>>
>>
>>
>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-phylo mailing list
> R-sig-phylo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo

Mark Holder

mtholder at ku.edu
http://www.people.ku.edu/~mtholder/

==============================================
Department of Ecology and Evolutionary Biology
University of Kansas
6031 Haworth Hall
1200 Sunnyside Avenue
Lawrence, Kansas 66045

lab phone:  785.864.5789

fax (shared): 785.864.5860
==============================================