Friends, I've been wondering if it's possible to c...

# getting-started

[Paulo Cereda]

Friends, I've been wondering if it's possible to compose regexes regexen regex things, e,g

val words = Regex("""\w+""")
val digits = Regex("""\d+""")

combined into \w+|\d+ via words | digits or similar. Any suggestions are welcome. Thanks! 🙂

[Vampire]

You already wrote how to do it.

[Vampire]

Besides that it is quite useless as \w already includes \d

[Vampire]

So \w+|\d+ is identical to \w+

[ephemient]

I tend to find that the extension looks nicer,

val words = """\w+""".toRegex()

[ephemient]

also it is generally difficult to combine regexes. you could

val both = "${words.pattern}|${digits.pattern)".toRegex()

but that will break back-references within each pattern, and doesn't handle differing regex options either

[ephemient]

for example, it won't work to combine these two

val re1 = """([a-z]+)""".toRegex(RegexOption.IGNORE_CASE)
val re2 = """(['"]).*\1""".toRegex(RegexOption.DOT_MATCHES_ALL)

[Michael de Kaste]

I've wanted to play around with a regexbuilder before where you could do something akin to

buildRegex{
    either{
        capture("""\w+""", name = "words")
        capture("""\d+""", name = "numbers")
        just("""\s+""")
    }
}

you could maybe make that?

[Vampire]

Ah, you meant combining from the variables, sorry, I got that wrong.
I thought you meant writing a regex that matches both.
Then what @ephemient said.
It is probably better to not try to do it unless you know for sure the patterns are combinable.

[ephemient]

or build a parser combinator instead of regex

[Paulo Cereda]

Thanks for the pointers, friends! I was just playing with my code for a talk and wondered if I could reuse/combine some regex I had (and thus have more problems 😁). A parser combinator sounds ideal. Cheers!

[ephemient]

a sketch of a primitive parser combinator: https://pl.kotl.in/9TI8Rl6dD

[Paulo Cereda]

@ephemient wow, this is brilliant, thank you! gratitude thank you