Inconsistency between String#to_d and String#to_f about bigdecimal HOT 5 CLOSED

Also another (potentially serious) issue. If you pass "1.", to_f & to_i accept it, but to_d throws an exception.

2.4.1 :001 > require 'bigdecimal'
2.4.1 :002 > require 'bigdecimal/util'
2.4.1 :003 > "1.".to_i
=> 1
2.4.1 :004 > "1.".to_f
=> 1.0
2.4.1 :005 > "1.".to_d
ArgumentError: invalid value for BigDecimal(): "1."

from bigdecimal.

I'm seeing something even more worrying with bigdecimal-1.3.2

2.4.1 :001 > require 'bigdecimal'
=> true
2.4.1 :002 > require 'bigdecimal/util'
=> true
2.4.1 :006 > "1.".to_i
=> 1
2.4.1 :007 > "1.".to_f
=> 1.0
2.4.1 :008 > "1.".to_d
=> 0.0

from bigdecimal.

I'm also seeing this issue. Looks like it's somewhat related to this fix: #55

In ruby 2.3.3:

>> "30.".to_d
=> #<BigDecimal:55adc4a76918,'0.3E2',9(18)>

In ruby 2.4.1 (and updating bigdecimal to 1.3.2):

>> "30.".to_d
=> 0.0

from bigdecimal.

also

> "1.23other".to_d
=> 0.123e1
> "1.23degree".to_d
=> 0.0

from bigdecimal.

I guess the problem it that:
to_d tries to operate with "e"(also does the same with "d") on the string to multiplies or divide by 10 times the value that succeds the letter, but when encounters another letter, it somehow sets the value to 0.0 while to_f ignores the word.

that's why stuff like that happens:

> "1.23other".to_d
=> 0.123e1
> "1.23degree".to_d
=> 0.0
> "1.23kelvin".to_d
=> 0.123e1
> "1.23degree".to_f
=> 1.23

I'm having some struggle with a validation, because when it tries to validates an string with my regex it refuses stuff like "41avocado" but shows to the user the value "41.00" with error, cause after the typecast, the word is just ignored, but if the letter after the number was "e" or "d" the value is set to "0.00", wich is more user-friendly to be understood...

from bigdecimal.