clib2 Code

- Small indentation changes for the stub code.

- Rewrote the GetRexxVar()/SetRexxVar() functions (mostly) in 'C', using available
code as a reference. The 68k stubs currently only work for the 68k version of
the library, and a solution for the OS4 build still needs to be found. The code
is currently untested, but it should be complete.

- Added local inline function definitions for the two rexxsyslib.library V45
API functions, rather than omitting support for them in the 68k build.

- reverted the soft-float subdirectory changes as the problem was solved by
olaf in a different way so that libm.a never contains any hard float
opcodes and instead libm881 contains them. so the separate soft-float
directory are not required.

[tboeckel]: when compiling amiga_rexxvars.c for m68k/OS3 the GetRexxVarFromMsg()
and SetRexxVarFromMsg() will not be used and therefore not cause linkage errors
with our libamiga.a stubs anymore.

- Added "-msoft-float" to the non-FPU libm build rules, which ought to
resolve the issues with FPU code popping up in the mathieee#?.library
software floating point code.

- The printf() family now ignores argument specifications, as in "%2$d %1$d",
which are used for localization on some platforms. This is a non-standard
feature and the way clib2 treats it for now is just intended to avoid
trouble while performing the conversion.

- Temporarily deactivated the separate -msoft-float builds.

- [jlangner]: changed the GNUmakefile.68k to also build soft-float variants of all
our libraries. In addition the specs file now also respects the -msoft-float
option at link time and sets the default link directory accordingly.

- the log() and log10() functions always returned -inf() even if the arguments
was within the valid range for a log() function. Using DBL_EPSILON as the
threshold was not correct as EPSILON is 2.2204460492503131E-16 whereas
values like 1E-200 are still valid double values for a log().