These are known bugs that you should know about in "other" software:

 - Warning : DOS/4GW v1.97 has no callbacks under DOS so the signal() init
   will give a warning message.  Why are there no callbacks available?
   DOS/4G v2.01a works.

 - Warning : Watcom C/C++ (v10.6 and v11.0) do not support "long double"!
   Try running \test\float.c.  It will say the sizeof(long double) is 8 while
   BCv5.0 gives me 10.  So be warned.  This means that %Lf will not work and
   you should use %lf with Watcom.  I suggest to never use long double in
   Watcom, or all together.

 - Warning : In 32bit C/C++ when you give a float to a function directly ie:
     printf("%f",5.678);
   The float will be a "double" so you must use %lf above.  I've tested this
   with BC and WC.  In 16bit C/C++ they use to be just "float"s.  You
   can use 'L' after the number to indicate it is a "long double", although
   it will do nothing in WC since it does not support "long double".  What
   a fucked world...

 - Warning : BCC32 and WCC386 both expand "float" to "double" when
   using varargs (ie: ...).  See \test\math.c for a good example.
