I want to name or reference a grouped pattern for reuse without enforcing a similar match across this pattern to shorten my regex.
I'm sure there are examples with more obvious benefit to doing something like this, but let's say I want to match something like a string composed of six 2-digit hex numbers separated by colons, such as 38:f8:b7:90:45:92.
The pattern I came up with for this is (?:[0-9A-Fa-f]{2}:){5}[0-9A-Fa-f]{2}
[0-9A-Fa-f]{2} is used to represent a 2-digit hex number, which I'll call the hex pattern. Instead of repeating this a second time after the colon, I'd like a way to name this or something so I can shorten my regex.
I've already tried (?:(?<x>[0-9A-Fa-f]{2}):){5}\k<x> but rather than being the original hex pattern I made, it seems to only match with the last match the hex pattern found. For example, running this regex on 38:f8:b7:90:45:92 will basically turn the pattern into ([0-9A-Fa-f]{2}):){5}45 since 45 is the last match the original hex pattern found.
Therefore only something like 00:18:12:c1:5a:5a, where the last two 2-digit numbers are the same, will match.
Is there a way to name a pattern for complete reuse?
If supported, you could make use of repeating the subpattern
That might look like:
(?:([0-9A-Fa-f]{2}):){5}(?1)
^ ^^^^
Or by name:
(?:(?<x>[0-9A-Fa-f]{2}):){5}(?&x)
^^^^^ ^^^^^