Distribute aclocal.m4, install-sh, config.guess, and config.sub
to applications at build time instead of having multiple identical
copies committed in the repository.

* rickard/libatomic_ops_improvements/OTP-12302:
  Implement support for double word atomics using libatomic_ops
  Improve usage of libatomic_ops for word size atomics

Turns out that isfinite emits a function call and not
an instruction in gcc, this makes estone float arith
about 50-75% slower. finite emits the instruction so
we use that instead.

Use AO_fetch_compare_and_swap*() when present

OS X Mavericks builds result in a number of warnings about finite() being
deprecated, like these:

beam/erl_arith.c:451:7: warning: 'finite' is deprecated: first deprecated in OS X 10.9 [-Wdeprecated-declarations]
                    ERTS_FP_ERROR(p, f1.fd, goto badarith);
                    ^
sys/unix/erl_unix_sys.h:319:33: note: expanded from macro 'ERTS_FP_ERROR'
                                        ^
sys/unix/erl_unix_sys.h:244:51: note: expanded from macro '__ERTS_FP_ERROR'
                                                  ^
/usr/include/math.h:718:12: note: 'finite' has been explicitly marked deprecated here
extern int finite(double) __OSX_AVAILABLE_BUT_DEPRECATED(__MAC_10_0, __MAC_10_9, __IPHONE_NA, __IPHONE_NA);

Add checks to use isfinite() instead of finite() where available. Verified
on OS X Mavericks 10.9.5 and Ubuntu 12.04.