What is the correct Raku recursive regex syntax?

Question

What or how is Raku recursive regex syntax and all match variable in Raku as on try

'hellohelloworldworld' ~~ m{ ^(h\w+?o) (?0) world  };
say "\n=$&"

seems to not work

Please help out solving these.

@raiph This question should have an answer regarding "publication of a match variable." Please consider reposting, thx. — jubilatious1
– jubilatious1, Commented Feb 1, 2022 at 21:46
@raiph yes but you're citing an answer to a question initially posted 3.5 years ago. I for one expect the compiler to have evolved since then, so that what may have been considered a bug back then has either 1) been enshrined in the docs, or 2) has been obviated. Which is to say...going forward, a 2022 updated answer from you regarding "Publication" of match variables by Rakudo would be most prized. Regards. — jubilatious1
– jubilatious1, Commented Feb 1, 2022 at 22:51
@jubilatious I think it was changed in the question, so not your mistake :-) — Sebastian
– Sebastian, Commented Feb 2, 2022 at 16:46
@jubilatious1 "Still hoping to confirm/refute ..." Publication of match variables is irrelevant to recursion. That's one of the reasons I deleted my answer. (I don't think you can see my answer. You need 10K+ reputation to see deleted answers. — raiph
– raiph, Commented Feb 13 at 16:39
Having reviewed this SO right now (due to you linking to it in your recent SO) I see my answer (deleted at the time I write this comment) contained the seeds of a good answer. It correctly demonstrates both general recursion and <~~> recursion. I think I will reinstate it with the mistakes removed and a fresh attempt at directly answering the original question added. I recall deleting my answer because I had decided .@codesections' answer was better. I now realize his answer is wrong. I'll add a comment below his answer to clarify that. — raiph
– raiph, Commented Feb 13 at 16:47

codesections · Accepted Answer · 2022-01-30 21:41:33Z

8

Raku has dedicated syntax for anonymous recursive regexes :<~~>.

Using this syntax, you could write the regex in your question as:

'hellohelloworldworld' ~~ m{ ^(h\w+?o) <~~>? world  };

say $/; # OUTPUT: «｢hellohelloworld｣␤
        #          0 => ｢hellohello｣␤»

answered Jan 30, 2022 at 21:41

codesections

9,62018 silver badges53 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

raiph Over a year ago

"I didn't plagiarize, honest teach!" :) (I prepared an edit of my answer covering self-recursion, and a bogus-but-maybe-interesting "Answer #3", hours ago. But I lost net connection (it's yoyoing today due to Storm Corrie), and then got waylaid, before publishing it (just now).)

jubilatious1 Over a year ago

Confused, because without <~~>? the code say $/ if 'hellohelloworldworld' ~~ m{ ^(h\w+?o) world }; still returns the same result. I'm on REPL version moar (2021.06).

jubilatious1 · Accepted Answer · 2022-02-02 16:35:19Z

The two answers I expected to see have already been posted, they are:

" {} publication" of a match variable for use later within the same regex/matching operation (technically a backreference):

    > say $/ if 'hellohelloworldworld' ~~ m/ ^(h\w+?o) {} $0 world /;
    ｢hellohelloworld｣
     0 => ｢hello｣
    > say $/ if 'hellohelloworldworld' ~~ m/ ^(h\w+?o) world /;
    ｢hellohelloworld｣
     0 => ｢hellohello｣

and,

use of Raku's dedicated " <~~> recursing-match" operator within the regex.

In true TMTOWTDI-spirit, there is however a third option, using Raku's :nd() adverb to achieve a sort of "poor-man's" recursion. Starting from the ['(' \w* ] grouping, you can successively pull out ｢(bird｣, ｢(in｣, and ｢(nest｣ from the input string (bird(in(nest))). Or all three at once (last example):

In the Raku REPL:

> my $nested = "(bird(in(nest)))";
(bird(in(nest)))
> say $nested;
(bird(in(nest)))
> say  $nested ~~ m:1st/ ['(' \w* ] /;
｢(bird｣
> say  $nested ~~ m:2nd/ ['(' \w* ] /;
｢(in｣
> say  $nested ~~ m:3rd/ ['(' \w* ] /;
｢(nest｣
> say  $nested ~~ m:nd(1..3)/ ['(' \w* ] /;
(｢(bird｣ ｢(in｣ ｢(nest｣)
>

Behind the scenes this is most likely using Raku's :position adverb or :continue adverb, in conjunction with Raku's $/.to match variable:

> say  $nested ~~ m/ ['(' \w* ] /;
｢(bird｣
> say  $nested ~~ m:pos($/.to)/ ['(' \w* ] / given $nested ~~ m/ ['(' \w* ] /;
｢(in｣
> say  $nested ~~ m:pos($/.to)/ ['(' \w* ] / given $nested ~~ (m/ ['(' \w* ] / && m:pos($/.to)/ ['(' \w* ] /);
｢(nest｣
>

Again, Raku gives you a lot of different ways to approach the problem, which is one of the niceties of the language.

Collectives™ on Stack Overflow

What is the correct Raku recursive regex syntax?

2 Answers 2

2 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related