Max pin length should be determined by token info data

[simo5]

Wow, nice find!
Although I wonder if we should have such a tight limit on PINs?
Right now there is a variable holing MAX_PIN_LENGHT but in truth the largest length is defined by the token's info ulMaxPinLen ... so it seems this 32 byte limitation is quite artificial and may unduly restrict access to some tokens ...

Originally posted by @simo5 in #654 (comment)

[numinit]

Good point. I've noticed that some tokens say something like 128 in practice. (Of course, whether they truly respect that is up to the token.)