5.3. Pcre

UTF-8 checking now references RFC 3629 instead of RFC 2279. This makes it more restrictive in the strings that it accepts. For example, the UTF-8 character ordinal value is now limited to 0x0010FFFF:

echo -ne "\x00\x11\xff\xff" | recode UCS-4-BE..UTF8 | pcregrep --utf-8 '.'

$ echo -ne "\x00\x11\xff\xff" | recode UCS-4-BE..UTF8 | pcregrep --utf-8 '.'
pcregrep: pcre_exec() error -10 while matching this line:

See the RFC for more details: http://tools.ietf.org/html/rfc3629#section-12.

Saved patterns that were compiled by earlier versions of PCRE must be recompiled. This affects applications that serialize pre-compiled PCRE expressions to external memory (for example, a file) and load them later. This is usually done for performance reasons, for example in large spam filters.