Fix process_continue and check_code_block

[tompng]

Rename and refactor process_continue and check_code_block

def process_continue(tokens)
  # doing something like maybe_should_continue?
  # check if last token is BEG (half part of `should_continue?`)
end

def check_code_block(code)
  # doing something like code_syntax_is_not_complete_or_maybe_should_continue?
  # check code is syntax_valid recoverable_error unrecoverable_error
  # check if the last token is DOT and checks other duplicated condition with syntax check and open_tokens check (other half part of `should_continue?`)
end

↓

def should_continue?(tokens) = true or false
def check_code_syntax(code) = :valid or :recoverable_error or :unrecoverable_error or :other_error

def code_terminated?(code, tokens, opens)
  # calculate by should_continue?() and check_code_syntax()
end

This will fix prompt bug introduced in #500 or #515. (because process_continue was incomplete, half of the necessary implementation was in check_code_block)

# actual
irb(main):001:1> if 1
irb(main):002:0> end.

# expected
irb(main):001:1> if 1
irb(main):002:0* end.

Other fix

Previous implementation of process_continue only uses last two tokens.
It ignores only the last token newline tokens
I change it to ignore consecutive trailing comments, spaces, newlines.
It will fix this old bug.

actual
irb(main):001:0* 1; # comment
irb(main):002:0* 
irb(main):003:0* 

expected
irb(main):001:0> 1; # comment
=> 1
irb(main):002:0> 

[st0012]

LGTM 👍 Just need to fix some conflicts