Is the random library used by the standard client in any way deterministic, or can one count on it being truly random?
For example, often random libraries are initialized with seed value of the current time, like in C++:
srand(time(NULL));
But if someone was to iterate over reasonable ranges of time when a client started, eventually one would use the same seed value, and thus be able to generate the same addresses.
Is the random library used by the standard client vulnerable to such an attack, or does it also use some other variables that are unpredictable (say, temperature of the processor down to such decimal places it becomes an unpredictable noise)?