Erroneous regular expressions

I've noticed errors in two regular expressions used in the validators:

In SEPA::CreditorIdentifierValidator:

REGEX = /\A[a-zA-Z]{2,2}[0-9]{2,2}([A-Za-z0-9]|[\+|\?|\/|\-|\:|\(|\)|\.|,|']){3,3}([A-Za-z0-9]|[\+|\?|\/|\-|:|\(|\)|\.|,|']){1,28}\z/
#                                             ^   ^  ^  ^  ^  ^  ^  ^  ^ ^
#                                             A   B  B  B  ...

Here, | is used as a meta-character for alternation to combine character classes (A) which is correct. But | is also used within a character class to combine its characters (B) which is wrong. Within a character class, | is treated as a literal character. You can't use it as a meta-character here.

In SEPA::MandateIdentifierValidator:

REGEX = /\A([A-Za-z0-9]|[\+|\?|\/|\-|\:|\(|\)|\.|\,|\']){1,35}\z/

• Again, | inside the character class (| is not allowed per spec) • The "space" character is missing, which according to the spec is a valid character

Apart from the above errors, the regular expressions seem overly complex and are quite hard to read. For example, instead of defining two separate character classes which are joined via |, you can define the allowed characters in one character class which allows you to omit the capture group. You can also remove all of the escape characters by using %r|...| and by moving - to the end. Applied to MandateIdentifierValidator's regex:

REGEX = %r|\A[A-Za-z0-9 +?/:().,'-]{1,35}\z|

salesking / sepa_king

Erroneous regular expressions #94