#430 Minor signed integer overflow in DBOPL's Chip::Setup

[NY00123]

Crash reproduced with a 32-bit Linux build using the following command (not with a 64-bit build):

../configure CXXFLAGS="-g -ftrapv"

Also reproduced on Windows while using a bit different revision of DBOPL (basically a C port) in another application.

Location of crash from gdb (in __mulvsi3):

#4  0x0817cb18 in DBOPL::Chip::Setup (this=0xa8f0b6c, rate=49716)
    at ../../../src/hardware/dbopl.cpp:1277
1277     guessAdd = ((guessAdd * mul) >> 12);
(gdb) print guessAdd
$1 = 524286
(gdb) print mul
$2 = 4159

A patch is attached, casting guessAdd to a 64-bit integer in two locations. This seems to fix the issue, and may also explain why isn't this reproduced with a 64-bit build.