You seem to be assuming that the C++ code is correct and the compiler is
wrong. It seems very much more plausible that the reverse is true.
Gcc3 was notoriously loose in its standards conformance in C++ and
Fortran. Gcc4 is better, but g++4 still allows at least 10 packages on
CRAN to compile
>>
I am trying to compile rseries from Whit Armstrong and a colleague of mine found
a problem with using GCC4
I get the following error when compiling rseries
g++-sjlj -Ic:/R/include -O2 -Wall -c Rutils.cpp -o Rutils.o
Rutils.cpp: In function 'double* getColPointer(SEXPREC*, int)':
Rutils
