[[:punct:]] matches differently in irb and rails test [duplicate] - ruby

Question

This question already has an answer here:
Regex "punct" character class matches different characters depending on Ruby version
(1 answer)
Closed 5 years ago.
[[:punct:]] doesn't match any punctuation when it's called by a rails model test. Using the following code
test "punctuation matched by [[:punct:]]" do
punct_match = /\A[[:punct:]]+\Z/.match('[\]\[!"#$%&\'()*+,./:;<=>?#\^_`{|}~-]')
puts punct_match
puts punct_match.class
end
this outputs a non-printable character and NilClass.
However, if I execute the same statement
punct_match = /\A[[:punct:]]+\Z/.match('[\]\[!"#$%&\'()*+,./:;<=>?#\^_`{|}~-]')
in irb matches correctly and outputs
[\]\[!"#$%&'()*+,./:;<=>?#\^_`{|}~-]
=> nil
What am I missing?
Ruby version => 2.2.4,
Rails version => 4.2.6

Answer 1

The behaviour of /[[:punct:]]/ changed slightly in ruby version 2.4.0.
This bug was raised in the ruby issues, which links back to this (much older) issue in Onigmo - the regexp engine used since Ruby 2.0+.
The short answer is, these characters were not matched by /[[:punct:]]/ in ruby versions <2.4.0, and are now matched:
$+<=>^`|~
You must be running irb in a newer ruby version than this rails application, which is why it matches there.
On a separate note, you should alter your code slightly to:
/\A[[:punct:]]+\z/.match('[]!"#$%&\'()*+,./:;<=>?#^_`{|}~-]')
Use \z, not \Z. There is a slight difference: \Z will also match a new line at the end of the string.
You have unnecessary back-slashes in the string, such as '\^'
You have repeated a [ character: '[\]\['