CVE-2007-1661

La librería Perl-Compatible Regular Expression (PCRE) anterior a 7.3 vuelve demasiado atrás cuando compara determinados bytes de entrada con algunos patrones de expresiones regulares en modo no-UTF-8, lo cual permite a atacantes locales o remotos (dependiendo del contexto) obtener información sensible o provocar una denegación de servicio (caída), como se ha demostrado mediante los patrones "\X?\d" y "\P{L}?\d".

Tavis Ormandy (Google Security) discovered multiple vulnerabilities in PCRE. He reported an error when processing \Q\E sequences with unmatched \E codes that can lead to the compiled bytecode being corrupted. PCRE does not properly calculate sizes for unspecified multiple forms of character class, which triggers a buffer overflow. Further improper calculations of memory boundaries were reported when matching certain input bytes against regex patterns in non UTF-8 mode and when searching for unmatched brackets or parentheses. Multiple integer overflows when processing escape sequences may lead to invalid memory read operations or potentially cause heap-based buffer overflows. PCRE does not properly handle \P and \P{x} sequences which can lead to heap-based buffer overflows or trigger the execution of infinite loops, PCRE is also prone to an error when optimizing character classes containing a singleton UTF-8 sequence which might lead to a heap-based buffer overflow. Versions less than 7.3-r1 are affected.