Project

General

Profile

Actions

Bug #13116

closed

modulo, divmod range problem: float_val % 1 may return 1.0

Added by tompng (tomoya ishida) over 4 years ago. Updated over 4 years ago.

Status:
Rejected
Priority:
Normal
Assignee:
-
Target version:
-
ruby -v:
ruby 2.4.0p0 (2016-12-24 revision 57164) [x86_64-darwin16]
[ruby-core:79014]

Description

mod, divmod range problem: float_val % 1 may return 1.0

x % y sometimes returns y (not in 0 <= result < y)

float_val = 0.0.prev_float
float_val % 1 #=> 1.0
Math.sin(2*Math::PI) % 10 #=> 10.0

I not sure this is a bug or a feature
but I think returning x.prev_float rather than x would be appropriate.

# when would this feature cause problem?:
require 'chunky_png'
image = ChunkyPNG::Image.new 100, 100
points = 100.times.map{|i|[10*Math.sin(Math::PI*i/10), i, 0xaabbccff]}
# render points to an image object(circular boundary)
points.each do |x, y, color|
  image[(x % image.width).floor, (y % image.height).floor] = color
  # => ChunkyPNG::OutOfBounds: Coordinates (100,20) out of bounds!
end
image.save 'out.png'

# to avoid error:
(x % w % w).floor
[(x % w).floor, w - 1].min
xx = (x % w).floor; xx -= 1 if xx == w
# but it's a little bit weird
# sample implementation in ruby:
class Float
  alias_method :original_modulo, :%
  alias_method :original_divmod, :divmod
  def divmod x
    div, mod = original_divmod x
    if mod == x
      mod = x > 0 ? mod.prev_float : mod.next_float
    end
    [div, mod]
  end

  def % x
    mod = original_modulo x
    return mod if mod != x
    x > 0 ? mod.prev_float : mod.next_float
  end
end
Actions

Also available in: Atom PDF