[R] R can't find gcc library that other programs can.

Dr. David Kirkby david.kirkby at onetel.net
Sat Jun 12 03:32:28 CEST 2010

R is used in the Sage project. R is building on Solaris 10 with SPARC 
processors. Until recently, I did not give it much more thought, as it appeard 
to build ok.

More recently someone noticed a test failure. It would appear a number of 
modules are not building (Matrix, class, mgcv, nnet, rpart, spatial, and 
survival) are all failing to build. But the R build process does not terminate. 
Is that intensional?

The way the build is failing is very odd indeed.

  ld.so.1: R: fatal: libgcc_s.so.1: open failed: No such file or directory

This is despite the location of libgcc_s.so is specified in LD_LIBRARY_PATH, and 
numerous other parts in Sage link to this library ok.

There's some information about the issue here


Does anyone have any suggestions? In particular, do you consider the failure of 
one or more of those modules sufficiently serious that the build of R should 
stop, or do you consider some failures like this not so serious, and so it is 
right for the build to continue? (The reason I want to know this, is that we 
want to test for this failure in Sage. Depending on the seriousness of this, we 
make break the build if such a module fails, or we may decide to continue, and 
test things later, where a test failure would not stop Sage working.


More information about the R-help mailing list