Regex for UUIDs: matching the 8-4-4-4-12 layout

The layout is fixed and easy to state: eight hex digits, then three groups of four, then twelve, separated by hyphens. The preset says exactly that and nothing more, which makes it a good scraper. Running it over a log file pulls out every identifier regardless of which version generated it, and that is usually what you want when you are looking for one.

Two nibbles carry meaning. The first character of the third group is the version, so a version 4 identifier has a 4 there. The first character of the fourth group encodes the variant, and for the layout in normal use it is 8, 9, a or b. Pinning both turns a layout check into a check that the value really is a random version 4 identifier.

The distinction matters when the identifier is doing security work. If your system issues version 4 identifiers because they are unpredictable, accepting a version 1 identifier undermines that, since a version 1 value encodes a timestamp and a node address and is guessable. A stricter pattern is the cheapest way to reject one.

The preset carries no anchors, which is right for finding identifiers inside text and wrong for validating a field. Adding ^ and $ turns it into a validator, and while you are there the case insensitive flag is worth a thought too: the canonical form is lowercase, and accepting uppercase silently creates two spellings of the same key.