Some vulnerabilities have been reported in PCRE, which can be exploited by malicious people to cause a DoS (Denial of Service), disclose sensitive information, or potentially compromise an application using the library.

1) An error exists in the processing of "\Q\E" sequences with orphan "\E" codes. This can be exploited to desynchronize the compiled regular expression and execute corrupted bytecode.

2) An error in the processing of multiple unspecified character classes can be exploited to trigger an insufficient memory allocation.

3) An error exists in the processing of multiple "\X?\d" and "\P{L}?\d" patterns in non-UTF-8 mode. This can be exploited to crash an affected process or to disclose restricted memory.

4) An error exists in multiple unspecified routines when searching for unmatched brackets or parentheses. This can be exploited to crash an affected application via a specially crafted string.

5) Some integer overflow errors when processing escape sequences can be exploited to crash an affected application or potentially cause heap-based buffer overflows.

6) Some errors exist in the processing of "\P" and "\P{x}" sequences. This can be exploited to cause heap-based buffer overflows or trigger the execution of infinite loops.

7) An error in the optimization of character classes containing a lone unicode sequence can be exploited to cause a heap-based buffer overflow.

The vulnerabilities are reported in versions prior to 7.3.

Solution
Update to version 7.3 or later.

Provided and/or discovered by
Debian credits Tavis Ormandy, Google Security Team

Original Advisory
http://lists.debian.org/debian-security-announce/debian-security-announce-2007/msg00177.html

Deep Links
Links available in Customer Area