[ruby-core:28336] [Feature #2152] Split functionality of Float#inspect and Float#to_s

Issue #2152 has been updated by Benoit Daloze.


Sorry to answer so late.

I think the change of Array#to_s is not comparable here. Many people did use the fact that Array#to_s did what Array#join does now.

But here it's very different: Float are almost not used for their real representation like it is now default with #to_s (it is not intended for them to get a lot of decimals I mean). I think in any app you want to code, you will want to use a fixed representation ('%.2f'), or a nice one (without ...9999999). So, I think not a lot of people use Float#to_s in 1.9.2, because it's not appropriate to show the data, except if you absolutely want to show the errors in Float. I think changing #to_s to the old behavior will break a very few codes. And would even improve compatibility with earlier versions!

Here is a simple example:
ruby -e 'p 729.0/10'

ruby-1.8.7-p249: 72.9
ruby-1.9.2-head: 72.90000000000001

"While I can appreciate the desire to make floats a little clearer to beginners."
You seems to not understand that the behavior of Float#to_s has changed, and that we propose to make #to_s like before, while keeping #inspect for the full representation.

This change would improve compatibility and tracking Float errors for beginners (and make you still wise of what is a Float when you are 'inspecting' it).

It would also, according to me, be more consistent with Ruby standards of #to_s and #inspect (I said upper why).
----------------------------------------
http://redmine.ruby-lang.org/issues/show/2152

----------------------------------------
http://redmine.ruby-lang.org