PCRE

Describes how recheck checks a regular expression of the pcre dialect.

Engine

recheck follows the syntax and the semantics of PCRE2 10.49 as PHP's preg_* functions use it.

Settings

The pcre dialect has no settings.

CLI

$ recheck --dialect=pcre '/^(a|a)*$/'

JavaScript

import { check } from "recheck";

await check("^(a|a)*$", "", { dialect: "pcre" });

Scala

import codes.quine.labs.recheck.ReDoS
import codes.quine.labs.recheck.common.Parameters
import codes.quine.labs.resyntax.dialect.Dialect

ReDoS.check("^(a|a)*$", "", Parameters(dialect = Dialect.PCRE))

Flags and Units

The matching unit is byte, or scalar-value with the u flag or (*UTF) (see Matching Units).

recheck supports the following pattern modifiers of PHP.

Flag	PCRE2 option	Inline	Notes
i	PCRE2_CASELESS	(?i)
m	PCRE2_MULTILINE	(?m)
n	PCRE2_NO_AUTO_CAPTURE	(?n)	PHP 8.2 or later.
r	PCRE2_EXTRA_CASELESS_RESTRICT	(?r)	PHP 8.4 or later.
s	PCRE2_DOTALL	(?s)
u	PCRE2_UTF and PCRE2_UCP	(*UTF)(*UCP)
x	PCRE2_EXTENDED	(?x)	xx is x too, while (?xx) is PCRE2_EXTENDED_MORE.
A	PCRE2_ANCHORED		The match starts at the start of the input.
D	PCRE2_DOLLAR_ENDONLY		$ matches only at the end of the input, unless with m.
J	PCRE2_DUPNAMES	(?J)
U	PCRE2_UNGREEDY	(?U)
S			No effect (PHP 7.3 or later).
X			No effect (PHP 7.3 or later, where an unknown escape is always an error).

Syntax Limitations

The following constructs are not supported.

• Recursion and subroutine calls (e.g. (?R), (?1), (?&name) and \g<name>), including (?(DEFINE)...) in a pattern with calls, are reported as unsupported.
• Callouts (e.g. (?C1)), (*scan_substring:...) and script runs (e.g. (*sr:...)) are reported as unsupported.
• Extended bracketed character classes ((?[ ... ])) are not supported yet, and they are reported as invalid.

Approximations

• PCRE2 has auto-possessification and JIT that recheck does not model, so an attack that recheck reports may not work on PCRE2, and the runtime-attack verification checks whether it does.
• \G matches only at the start of the input, since recheck models a search from there, not one from a later position (e.g. the second match of preg_match_all).
• PCRE2 skips the attempts that cannot match by its start-of-match optimizations (e.g. (*COMMIT)abc matches xyzabc in PCRE2, unless the pattern starts with (*NO_START_OPT)), and recheck does not model them.

Runtime-Attack Verification

• Command: php.
• Matching: preg_match.
• Engine limits: the backtrack limit.
• Engine errors: the recursion limit and the JIT stack limit.
• Environment variable: RECHECK_PHP.

The program runs with PHP's default settings of pcre.backtrack_limit, pcre.recursion_limit and pcre.jit.