fix #100 to correctly mark parent checkbox

[mstojanovic]

Code is commented since the math is a bit tricky and not intuitive. And
vimscript doesn't like floating point calculations

[dkarter]

Thanks for the PR @mstojanovic!

Would be good to add a snapshot test to spec/checkboxes_spec.rb that would confirm what would happen at each step, something along these lines:

  it 'toggle a checkbox step wise for 6 bullets' do
    filename = "#{SecureRandom.hex(6)}.txt"
    write_file(filename, <<-TEXT)
      # Hello there
      - [ ] top bullet
        - [ ] first bullet
        - [ ] second bullet
        - [ ] third bullet
        - [ ] fourth bullet
        - [ ] fifth bullet
        - [ ] sixth bullet
    TEXT

    vim.edit filename
    vim.command "let g:bullets_checkbox_markers=' .:oOX'"
    vim.normal '2j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 1/6 = 16%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [.] top bullet
        - [X] first bullet
        - [ ] second bullet
        - [ ] third bullet
        - [ ] fourth bullet
        - [ ] fifth bullet
        - [ ] sixth bullet

    TEXT

    vim.normal 'j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 2/6 = 33%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [:] top bullet
        - [X] first bullet
        - [X] second bullet
        - [ ] third bullet
        - [ ] fourth bullet
        - [ ] fifth bullet
        - [ ] sixth bullet

    TEXT

    vim.normal 'j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 3/6 = 50%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [:] top bullet
        - [X] first bullet
        - [X] second bullet
        - [X] third bullet
        - [ ] fourth bullet
        - [ ] fifth bullet
        - [ ] sixth bullet

    TEXT

    vim.normal 'j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 4/6 = 66%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [o] top bullet
        - [X] first bullet
        - [X] second bullet
        - [X] third bullet
        - [X] fourth bullet
        - [ ] fifth bullet
        - [ ] sixth bullet

    TEXT

    vim.normal 'j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 5/6 = 83%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [O] top bullet
        - [X] first bullet
        - [X] second bullet
        - [X] third bullet
        - [X] fourth bullet
        - [X] fifth bullet
        - [ ] sixth bullet

    TEXT

    vim.normal 'j'
    vim.command 'ToggleCheckbox'
    vim.write

    file_contents = IO.read(filename)

    # 6/6 = 100%
    expect(file_contents).to eq normalize_string_indent(<<-TEXT)
      # Hello there
      - [X] top bullet
        - [X] first bullet
        - [X] second bullet
        - [X] third bullet
        - [X] fourth bullet
        - [X] fifth bullet
        - [X] sixth bullet

    TEXT
  end

[mstojanovic]

on it, i will do it asap just to clear one thing so that it makes more sense to all the users.
since i forked the project because i needed this feature ive noticed that the rounding doesnt really work as said.
i sad that the including border value is on the left (e.g. [25,50>) but after some usage and recalculations with this code its on the right (e.g. <25,50]). the issue is when the number_of_markers-2 == number_of_cb. and probably for some multiples k*number_of_markers-2 (honestly im too lazy to calculate now the correct formula for that).
basically my question boils down to decide what is more readable/understandable/natural (suppose 6 markers ' .:oOX')
a) right border included <50,75]

- [o] lowercase 
  - [X]
  - [X]
  - [X]
  - [ ]

b) left border included [75,100>

- [O] uppercase 
  - [X]
  - [X]
  - [X]
  - [ ]

[harshad1]

@mstojanovic This still merges and passes existing tests. Do you plan to add tests or should I just merge it?