[rkward-devel] Trouble loading packages on the Mac

meik michalke meik.michalke at uni-duesseldorf.de
Thu Mar 14 13:30:07 UTC 2013


hi,

Am Mittwoch, 13. März 2013, 18:19:26 schrieb Thomas Friedrichsmeier:
> well, the workaround shows that the breakage is on a fairly shallow level.

indeed. it should affect only (but all) packages including precompiled 
FORTRAN, and probably C code.

> As inelegant as that is going to be, I think it might be a better idea to
> use the official CRAN Mac binary version of R for the bundle. Not sure how
> this could be made to work. Perhaps roughly like this:
> 
> - First build as usual using Macports
> - From update_bundle.sh, wipe the R port, install CRAN R binary, in its
> place. - Rebuild RKWard only, to make sure it is linked to the CRAN R
> binary, appropriately
> - Package everything

i just tried to remove the dependecies to ports of R, install the CRAN package 
and compile RKWard, but get the following error:

<snip>
:info:configure -- Using R at /Applications/R.app/Contents/MacOS/R
:info:configure 2013-03-14 13:17:06.168 R[89268:a03] *** CFMessagePort: 
bootstrap_register(): failed 1100 (0x44c) 'Permission denied', port = 0x3403, 
name = 'WakeUpProcessPort'
:info:configure See /usr/include/servers/bootstrap_defs.h for the error codes.
:info:configure -- R architecture is 
:info:configure -- Looking for R_HOME
:info:configure 2013-03-14 13:17:21.106 R[89276:a03] *** CFMessagePort: 
bootstrap_register(): failed 1100 (0x44c) 'Permission denied', port = 0x3403, 
name = 'WakeUpProcessPort'
:info:configure See /usr/include/servers/bootstrap_defs.h for the error codes.
:info:configure CMake Error at rkward/rbackend/FindR.cmake:32 (MESSAGE):
:info:configure   Could NOT determine R_HOME (probably you misspecified the 
location of R)
:info:configure Call Stack (most recent call first):
:info:configure   rkward/rbackend/CMakeLists.txt:1 (INCLUDE)
</snip>

after i appended
 -DR_HOME=/Library/Frameworks/R.framework/Versions/Current/Resources
to the cinfigure arguments, the same error messages appeared, but RKWard at 
least started to build. and then it broke... more interestingly, during 
configuration, an R console pops up for two or three times, and configure only 
goes on after you closed them (this co-occurs with the above errors).

> (Normal macports-based source installation would continue to work the way it
> does, although probably, there, I should revert the default to installing R
> packages from source, not binary?).

i'd suppose we do the following:
 - revert to install source packages by default
   (can you make this a configure argument? then it's easier to have different 
   variants in one portfile)
 - build the bundle using the R-framework port
 - strip static libraries from the bundle

this will already reduce the bundle size significantly for the upcoming 
release, and we can experiment with other approaches after that.


viele grüße :: m.eik

-- 
  dipl. psych. meik michalke
  abt. f"ur diagnostik und differentielle psychologie
  institut f"ur experimentelle psychologie
  heinrich-heine-universit"at d"usseldorf
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20130314/4264c928/attachment.sig>


More information about the Rkward-devel mailing list