Regex Dialect & Syntax

regex.zig supports an expressive subset of standard regular expression syntax optimized for linear-time processing and complete ReDoS immunity.

---

Metacharacters & Escaping

The following metacharacters have special syntactic meaning and must be escaped with \ to be treated as literal characters:

\ . + * ? ( ) [ ] { } | ^ $

Use regex.escape(allocator, raw_str) or regex.escapeBuffer(&buf, raw_str) to programmatically escape user input.

---

Character Classes

Bracket expressions define custom character sets:

• [abc]: Matches any character in the set (a, b, or c).
• [^abc]: Negated set — matches any character not in a, b, or c.
• [a-z0-9]: Range expressions matching lowercase letters and digits.
• [\u{3040}-\u{309F}]: Unicode codepoint ranges up to 21 bits.

Predefined Character Shorthands

Shorthand	Meaning	ASCII Equivalent
\d	Digit	[0-9]
\D	Non-digit	[^0-9]
\w	Word character	[a-zA-Z0-9_]
\W	Non-word character	[^a-zA-Z0-9_]
\s	Whitespace	[ \t\r\n\x0c\x0b]
\S	Non-whitespace	[^ \t\r\n\x0c\x0b]

POSIX Character Classes

Inside bracket expressions: [:alnum:], [:alpha:], [:ascii:], [:blank:], [:cntrl:], [:digit:], [:graph:], [:lower:], [:print:], [:punct:], [:space:], [:upper:], [:word:], [:xdigit:].

---

Quantifiers & Repetition

Quantifiers define how many times the preceding atom may match:

Quantifier	Meaning	Greediness
*	0 or more times	Greedy
*?	0 or more times	Lazy
+	1 or more times	Greedy
+?	1 or more times	Lazy
?	0 or 1 time	Greedy
??	0 or 1 time	Lazy
{n}	Exactly $n$ times	Bounded
{n,}	At least $n$ times	Bounded
{n,m}	Between $n$ and $m$ times	Bounded
{n,m}?	Between $n$ and $m$ times	Lazy bounded

---

Anchors & Zero-Width Assertions

• ^: Beginning of line (in multiline mode) or beginning of text.
• $: End of line (in multiline mode) or end of text.
• \A: Absolute beginning of input string.
• \z: Absolute end of input string.
• \b: ASCII word boundary (\w adjacent to \W or string bounds).
• \B: Non-word boundary.

---

Grouping & Captures

• (expr): Numbered capturing group. Submatch offsets are recorded into capture slots $1, $2, etc.
• (?:expr): Non-capturing group. Groups sub-expressions without consuming capture slot indices.
• (?P<name>expr): Named capturing group. Retrieve matched slice by identifier with caps.name("name").
• (?i) / (?-i): Inline case-insensitive flag toggles.
• (?m) / (?-m): Inline multiline flag toggles.
• (?s) / (?-s): Inline dot-all flag toggles.

---

Unsupported Constructs

To guarantee linear-time execution and prevent arbitrary computational complexity, the following non-regular constructs are intentionally rejected at compile time:

• Backreferences (e.g. \1, \2): Backreferences transform pattern matching into an NP-complete problem.
• Lookaround Assertions (e.g. (?=...), (?!...)): Lookahead and lookbehind assertions complicate non-backtracking NFA engines.