Splitting out front matter from a Twig template with PCRE

Question

Let's say I want to split the metadata from the Twig template

---
some metadata
multiple lines
---
Twig template
More data

I came up with /\A---\R(.+?\R)?---\R(.*)\Z/s which does it more or less but I am wondering whether it can become pathological in backtracking.

Answer 1

Your regex seems working well.

You may make it somewhat more efficient if you "unroll" the first lazy dot pattern like

/\A---\R(.*(?:\R(?!---\R).*)*\R)?---\R(?s)(.*)\Z/

See the regex demo. Note: no mofifiers are necessary, there is one inline (?s) modifier inside the pattern.

Details

• \A - start of string
• ---\R - a full --- line with a linebreak
• (.*(?:\R(?!---\R).*)*\R)? - an optional Capturing group 1:
  • .* - the whole line
  • (?:\R(?!---\R).*)* - 0 or more repetitions of
    • \R(?!---\R) - a line break that is not followed with a --- line followed with a linebreak
    • .* - the whole line
  • \R - a linebreak sequence
• ---\R - a full --- line with a linebreak
• (?s) - an inline DOTALL modifier making the dots to the right match line break chars
• (.*) Group 2: any 0+ chars as many as possible
• \Z - end of string.