Question

I am not a regex ninja. I’ve been tweaking this for an hour, and I’m sure someone on SO can do it more effectively.

This is a regex for a username with some slightly peculiar requirements (to accomodate legacy usernames).

Rules:

• length: 24 charaters maximum, 3 characters minimum
• must not start or end with a space
• any number of the following (up to the max field lenght of 24):
  • 0-9
  • A-Z
  • a-z
  • . (dot)
  • ‘ ‘ (space)
• zero or one of each from following set:
  • @ (at)
  • _ (underscore)
  • – (hyphen)
  • ‘ (apostrophy)

Here’s what I have so far:

^[^ ](?=[A-Za-z0-9. @_\-]{1,24}$)[a-zA-Z0-9_. ]*\.?[a-zA-Z0-9_]*[^ ]$

but it is not quite right. I’m not sure how to escape the ‘ (apostrophe)

also, should I use another lookahead for the characters allowed zero or one times?

Thanks.

Update:
Note that target is the .Net 4.0 regex libraries with C#

Answer 1

This would be a regex solution, decide on your own if this is readable/usable/maintainable for you.

(?!^.*?([@'_-]).*\1.*$)^(?! )[\w. @'-]{3,24}(?<! )$

See it here on Regexr

(?!^.*?([@'_-]).*\1.*$) is a negative lookahead, if one of the characters you named is found it is put into capturing group 1 and to ensure this is not repeating using the backreference \1.

^(?! ) is a negative lookahead to ensure there is no space after the start of the string.

[\w. @'-]{3,24} The characters you allow at least 3 at most 24 of them

(?<! )$ is a negative lookbehind to ensure there is no space before the end of the string.