[R-SIG-Mac] Problem loading a package libraries (Error in dyn.load) on mac environments

Dan Tenenbaum dtenenba at fhcrc.org
Mon Dec 3 21:18:56 CET 2012 

On Mon, Dec 3, 2012 at 12:15 PM, Simon Urbanek
<simon.urbanek at r-project.org> wrote:
>
> On Dec 3, 2012, at 2:58 PM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:
>
>> On Mon, Dec 3, 2012 at 11:55 AM, Davor Cubranic <cubranic at stat.ubc.ca> wrote:
>>> Is rTandem built with the same version of Rcpp that is now installed (0.10.1)? The API changed a fair bit over the past few weeks.
>>
>>
>> Yes, in fact I just built Rcpp from source immediately before building
>> rTANDEM. I get the same results after installing the Rcpp binary from
>> CRAN. In both cases, I build and then attempt to load rTANDEM with the
>> exact same version of Rcpp.
>>
>
> CRAN version of Rcpp has both x86_64 and i386 libraries - unlike yours. May I suggest someone with the necessary knowledge respond to my initial e-mail before someone gets killed running of the wrong cliff? ;)

Hi Simon, you're right. If Rcpp is removed and then installed from
CRAN (binary version), the installed package then has both x86_64 and
i386 libraries, and rTANDEM then installs on both architectures.

Thanks for your help, sorry for the noise.
Dan



>
> Cheers,
> Simon
>
>
>
>> Dan
>>
>>
>>>
>>> Davor
>>>
>>> On 2012-12-03, at 11:52 AM, Dan Tenenbaum wrote:
>>>
>>>> On Mon, Dec 3, 2012 at 11:13 AM, Davor Cubranic <cubranic at stat.ubc.ca> wrote:
>>>>> Looks like you might not have the 64-bit version of the Rcpp dynamic library installed on the first machine. On the second, check if you have the 32-bit version of both XML and expat's dynamic libraries.
>>>>
>>>>
>>>> Thanks. I'm working with Frederic on this.
>>>> On the first machine, both the 32 and 64 bit versions are Rcpp are
>>>> installed (I can library(Rcpp) from both R --arch i386 and R --arch
>>>> x86_64), and the installed Rcpp has these files present:
>>>>
>>>> pwd
>>>> /Library/Frameworks/R.framework/Versions/Current/Resources/library/Rcpp
>>>> PETTY:Rcpp biocbuild$ ls -lR libs
>>>> total 0
>>>> drwxr-xr-x  3 biocbuild  admin  102 Dec  3 11:40 i386
>>>> drwxr-xr-x  3 biocbuild  admin  102 Dec  3 11:40 x86_64
>>>>
>>>> libs/i386:
>>>> total 1216
>>>> -rwxr-xr-x  1 biocbuild  admin  619064 Dec  3 11:40 Rcpp.so
>>>>
>>>> libs/x86_64:
>>>> total 1280
>>>> -rwxr-xr-x  1 biocbuild  admin  651920 Dec  3 11:40 Rcpp.so
>>>> PETTY:Rcpp biocbuild$ ls -lR lib
>>>> total 0
>>>> drwxr-xr-x  4 biocbuild  admin  136 Dec  3 11:40 i386
>>>>
>>>> lib/i386:
>>>> total 11336
>>>> -rw-r--r--  1 biocbuild  admin  5177484 Dec  3 11:40 libRcpp.a
>>>> -rwxr-xr-x  1 biocbuild  admin   619064 Dec  3 11:40 libRcpp.dylib
>>>>
>>>> After building rTANDEM against this Rcpp, I then get the same error
>>>> Frederic reports in this thread when trying to library(rTANDEM) on the
>>>> x86_64 architecture. Here it is again with sessionInfo():
>>>>
>>>>> library(rTANDEM)
>>>> Loading required package: XML
>>>> Loading required package: Rcpp
>>>> Loading required package: data.table
>>>> data.table 1.8.6  For help type: help("data.table")
>>>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>>>> unable to load shared object
>>>> '/Library/Frameworks/R.framework/Versions/2.16/Resources/library/rTANDEM/libs/x86_64/rTANDEM.so':
>>>> dlopen(/Library/Frameworks/R.framework/Versions/2.16/Resources/library/rTANDEM/libs/x86_64/rTANDEM.so,
>>>> 6): Symbol not found: __ZTIN4Rcpp7RObjectE
>>>> Referenced from:
>>>> /Library/Frameworks/R.framework/Versions/2.16/Resources/library/rTANDEM/libs/x86_64/rTANDEM.so
>>>> Expected in: dynamic lookup
>>>>
>>>> Error: package or namespace load failed for 'rTANDEM'
>>>>> sessionInfo()
>>>> R Under development (unstable) (2012-10-31 r61060)
>>>> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>>>>
>>>> locale:
>>>> [1] C
>>>>
>>>> attached base packages:
>>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>>>
>>>> other attached packages:
>>>> [1] data.table_1.8.6 Rcpp_0.10.1      XML_3.95-0.1
>>>>>
>>>>
>>>> I'm not sure if Frederic wants the package source to be posted at this
>>>> point. I'll let him answer that.
>>>>
>>>> Thanks,
>>>> Dan
>>>>
>>>>
>>>>>
>>>>> Davor
>>>>>
>>>>>
>>>>> On 2012-12-03, at 8:34 AM, Frederic Fournier wrote:
>>>>>
>>>>>> Hello everyone,
>>>>>>
>>>>>> I'm working on a small R package for proteomics analysis, and I am facing perplexing problems related to the 'dyn.load' function when loading the package on mac environments.
>>>>>> The package relies on two external libraries, expat and pthread, and on three cran packages: XML, Rcpp and data.table. I have no trouble building, checking or installing the package. But when I load the package on mac testing environments, I face various problems related to this error message: "Error in dyn.load(file, DLLpath=DLLpath, ...): ..."
>>>>>> On one machine, the package loads without problem on i386 R, but fails to load on x64 R; on another machine, things are reversed and the package loads on x64, but fails to load on i386.
>>>>>>
>>>>>> Here is the output from the machinge where the package loads on i386 but fails on x64 R:
>>>>>>
>>>>>>> library(rTANDEM)
>>>>>> Loading required package: XML
>>>>>> Loading required package: Rcpp
>>>>>> Loading required package: data.table
>>>>>> data.table 1.8.6  For help type: help("data.table")
>>>>>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>>>>>> unable to load shared object
>>>>>> '/Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM/libs/
>>>>>> x86_64/rTANDEM.so':
>>>>>> dlopen(/Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM
>>>>>> /libs/x86_64/rTANDEM.so, 6): Symbol not found: __ZTIN4Rcpp7RObjectE
>>>>>> Referenced from:
>>>>>> /Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM/libs/x
>>>>>> 86_64/rTANDEM.so
>>>>>> Expected in: dynamic lookup
>>>>>>
>>>>>> Error: package or namespace load failed for 'rTANDEM'
>>>>>>> sessionInfo()
>>>>>> R Under development (unstable) (2012-11-27 r61172)
>>>>>> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>>>>>>
>>>>>> locale:
>>>>>> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
>>>>>>
>>>>>> attached base packages:
>>>>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>>>>>
>>>>>> other attached packages:
>>>>>> [1] data.table_1.8.6    Rcpp_0.10.1         XML_3.9-4
>>>>>> [4] BiocInstaller_1.9.4
>>>>>>
>>>>>> loaded via a namespace (and not attached):
>>>>>> [1] tools_2.16.0
>>>>>>
>>>>>>
>>>>>> Here is the ouput from the machine where the package loads fine on x64 R but fails on i386:
>>>>>>
>>>>>>> library(rTANDEM)
>>>>>> Loading required package: XML
>>>>>> Loading required package: Rcpp
>>>>>> Loading required package: data.table
>>>>>> data.table 1.8.6  For help type: help("data.table")
>>>>>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>>>>>> unable to load shared object
>>>>>> '/Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM/libs/
>>>>>> i386/rTANDEM.so':
>>>>>> dlopen(/Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM
>>>>>> /libs/i386/rTANDEM.so, 6): Symbol not found: _XML_GetErrorCode
>>>>>> Referenced from:
>>>>>> /Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM/libs/i
>>>>>> 386/rTANDEM.so
>>>>>> Expected in: flat namespace
>>>>>> in
>>>>>> /Library/Frameworks/R.framework/Versions/2.16new/Resources/library/rTANDEM/libs/i
>>>>>> 386/rTANDEM.so
>>>>>> Error: package or namespace load failed for 'rTANDEM'
>>>>>>
>>>>>>
>>>>>> I am left clueless by this. As anyone faced a similar issue?
>>>>>>
>>>>>> Thank you very much for your help!
>>>>>>
>>>>>> Frederic
>>>>>>
>>>>>> _______________________________________________
>>>>>> R-SIG-Mac mailing list
>>>>>> R-SIG-Mac at r-project.org
>>>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>>>>
>>>>> _______________________________________________
>>>>> R-SIG-Mac mailing list
>>>>> R-SIG-Mac at r-project.org
>>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>>
>>
>> _______________________________________________
>> R-SIG-Mac mailing list
>> R-SIG-Mac at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>
>>
>

More information about the R-SIG-Mac mailing list