Always add trailing `.0` in `Float#inspect`; ref #4225

Trailing `.0` is removed from `Float#to_s` and `Float#inspect` at
9d08025b. However, I think the more human-readable format is better
for `Float#inspect`.

For example, in the `Float#to_s` format, the failure message is not
well understood when testing values including types by `eql?` (e.g.
`Numeric#step` test).

  ```ruby
  assert "example" do
    exp = 1.0
    act = 1
    assert_operator(exp, :eql?, act)  #=> Expected 1 to be eql? 1.
  end
  ```

1 files changed, 5 insertions, 0 deletions

diff --git a/test/t/float.rb b/test/t/float.rb
index 1cbc90532..e2a8a5088 100644
--- a/test/t/float.rb
+++ b/test/t/float.rb
@@ -239,6 +239,11 @@ assert('Float#to_s') do
   end
 end
 
+assert('Float#inspect') do
+  assert_equal("-3.25", -3.25.inspect)
+  assert_equal("50.0", 50.0.inspect)
+end
+
 assert('Float#eql?') do
   assert_operator(5.0, :eql?, 5.0)
   assert_not_operator(5.0, :eql?, 5)