Today I was browsing Rails codebase (sometimes people do weird things for fun), and found an interesting piece of code:
This sparkled my interest: what is divmod, especially - its pretty harsh form divmod(1)? Quick search through documentation - and it starts to make perfect sense:
Returns an array containing the quotient and modulus obtained by dividing num by numeric; especially if you instantly remember what is
modulus. Looks like example table in documentation is designed to confuse (including expected output of multiple methods), but for the special case of a positive float.divmod(1) there is much simpler explanation: it just splits a positive floating point number to its integer and fractional parts (+/- error).
Or, in other “words” (this one is true for negative numbers too):
1 2 3 4 5 6 7 8
Let look at other surprising Numeric methods (ones that I either never saw or successfully forgot about).
Returns square of self. Nothing to say, other than interesting choice of name. I would never guess what it means.
Returns an array of a target and an argument if they are of the same type, converts both to Float if they have different types. It would be interesting to see a practical (not in the core code) use of this.
Returns self for Numeric. Probably - no use for non-Complex numbers
I am excited about this pair, and really want to find a practical use for them. Documentation is extremely clear: it states that denominator returns denominator, and numerator - surprise! - numerator. Makes perfect sense for Rational, but sounds weird for all other types. In reality these methods return numerator and denominator of a Numeric converted to a Rational
Taking out of the context - try to make sense out of these results:
1 2 3 4
Together - much better
1 2 3 4 5 6
I could find only one “valid” use for #denominator for non-rational numbers - simplify (is it really simplification?) the check if any number is “practically an integer”
1 2 3 4 5 6
For #numerator I was able to find only pretty bad usage (not sure who started it as a pattern):
1 2 3
- for a non-manipulated Date input, it is no better than #.to_i
- for calculated dates - it is much worse than #to_i as it will purely lie to you.
Float division. Not sure yet, but I think I like it better than its practical equivalents a / b.to_f or a.to_f / b - seems more explicit.
Thank you, Ruby, for consistency! All other xxx? methods return true or false. But not this one. Returns self where you’d expect true, and nil for false. Usage example in documentation
1 2 3
does not give any clarity for why this inconsistency is necessary. This metthod would make much more sense to me without “?””