This, hands down, has been the single worst stretch of decompilation so far.
Three extremely difficult functions that each still required inline assembly.
And no, this didn't even work out with any of the optimization features in
Borland C++ that aren't included in Turbo C++.
nmlgc