Because there's non-programmers in this community, if you aren't sure what this means but are too afraid to ask, it's a Regular Expression that better represents the terms "Linux" and "Unix."
Though if we're going to be that pedantic, it would be
[nN][uiI][xX]$. That extra pipe wouldn't actually do anything in the last example, because regexp picks one character from the set by default.And if we want to be really pedantic,
(?!nix)[nN][uI][xX]$Would be the most accurate.
Edit: based on comments, I think...
(nux|NIX)$...would be the best. Then you don't wind up with weird matches with things like
UNiX.We're talking about Unix so being as pedantic as possible is actually required.
Actually
*nixisn't a Regular Expression, because the star operator*requires a preceding character or object to apply to. This is a wildcard for the shell style globbing, where a single star doesn't require a second object.Isn't
(I|U)equivalent to([IU])?Yes, but you can really only do that with single characters, since your first example is an ordered group and the second is an unordered set in a capturing group. The equivalency drops off when you include more characters.
Plus, you can do things like
[a-zA-Z], and you can't do that with the former example.I would imagine there's a difference in computing overhead, too, but I have no idea which is more performant.
*nixis more likely to be a glob, therefore an accurate version would be*n?xEdit: global -> glob dang autocorrect
All that effort, when you could've just called it LUNIX.
By the way, does this expression match LUNIX? But if so, won't it also match Binux or Bunix?
Yep, it would match LUNIX and Binux, but it would not match Bunix because of the negative lookahead.
There's a good idea and some real potential here, but it didn't quite land for me.
Unix
Minix
AIX
Irix
HP/UX
Ultrix
OSF/1
Linux
Those are just the ones I remember because I've used them
