Known bugs in Mathomatic
                           ------------------------

Compiling with Clang or Xcode 4 and up, which uses LLVM:
     This is actually an optimization bug in the LLVM backend used by Clang
and the Apple Xcode IDE. When dealing with memmove(3), LLVM optimization
introduces bugs and program hangs. So when you are compiling Mathomatic with
Clang or the latest Xcode, please be sure and compile with no optimization
options or the "-O0" option (which disables all optimization), otherwise the
regression tests and entering things like y=32^.5 and then simplifying will
cause Mathomatic to hang forever. memmove(3) is used throughout Mathomatic.
So when compiling Mathomatic with Clang or Xcode, always disable
optimization, it will run all the tests in 1 second and is not noticeably
slower. Mathomatic will hang during "make test" if compiled with optimization
enabled when using LLVM as a backend.

17! bug: Fixed.
     "factor number 17!" and higher don't work because the maximum safely
representable integer for double precision floating point is 15 digits long.
This is not a bug.

Absolute values and complex number problems:
     Absolute values and the imaginary unit (i) do not always remain intact
after being manipulated with the standard rules of algebra. For example, 1/i
correctly simplifies to -i, changing the sign of the result. So the fraction
command and other possible operations may give wrong results due to differing
sign after manipulations. The simplify command should always work, but it is
not guaranteed for these.
     The solve command may have a problem too, but using the verify option
should check that the result of solving absolute value or complex number
equations is 100% correct. Don't forget though that the absolute value |x|
definition in Mathomatic is ((x^2)^.5), which is different from the standard
absolute value function in the complex realm, which would always return a
positive and real distance on a complex Cartesian graph. Mathomatic is taking
shortcuts with its definition and will not always return a real value result
if any of the absolute value arguments are imaginary. It works perfectly with
real value arguments.

Solving modulus equations:
     May result in garbage, due to incomplete modulus operator solving code.
I could just disable it, but I choose to research it at a later time, because
it does sometimes work nicely. Again, the "solve verify" option tells if you
should trust the results.