30 bits is already enough for that. some languages have beefed up tests already.

done, but the description's markdown code is starting to be really bloated...

Wouldn't be a 6 kyu challenge then

yes. Like... well, "infinite". ;)

There is/was no ruby-specific description. Anyway, I updated the description to be a bit more clear

{'be': {'t': {}, 'g': {}}}

Those would have the same result ("be" is a prefix of both "bet" and "beg", the initial "be" in the first example will simply have no effect/will already be there). If it helps: try going through a few examples in random order, i.e. building a tree from the same arguments, but with a different ordering of them. That way, different cases emerge and you'll see how to handle them (or whether they are so different in the end).

done

27ms is cool! c++?

It should be said that we are to accept only symmetric keypads (which implies n x n), otherwise it's up to user's interpetation, considering one of random examples I've got [1,1,0,1,1], which was expected to be a No, nothing stops someone from seeing it as:

1 1
 0
1 1

Fixed.

Approved

Agree.
v[i+2] will be out of bounds when i >= v.size() - 2;
for loop condition is i<v.size(); I wonder how this code even passed all the tests!???

Are you modifying the input?