Fixing memory corruption issues
when parsing brainfuck code Previous commit introduced an assertion that can actually fail. In the hannoi example, the assertion failed. Luckily this was no malicious code that exploited the wrong computation in pre-assertion times. Simple reproduction of erroneous parsing: +--+ // does nothing but qunatity goes to zero with second - +--- // negative quantity. possible memory corruption exploit
Bug introduced in 5b33047
Checking that a brainfuck program only accesses it's tape memory. Fixes arbitrary heap memory corruption exploit.