prlw1 at cam.ac.uk
Wed Feb 20 18:46:54 CET 2013
I was compiling R on a system which admittedly was in the odd situation
that it declared a function in math.h, but didn't have the associated
implementation in libm, and the R 2.15.1 compilation failed.
It turns out that R_CHECK_FUNCS is rather less robust than the standard
AC_CHECK_FUNCS which actually attempts to link. On my temporarily broken
system, AC_CHECK_FUNCS would not have been fooled, and R would have
Thoughts on making R_CHECK_FUNCS simply an alias for AC_CHECK_FUNCS?
More information about the R-devel