On Thu, 12 Mar 2009, Dimitrios wrote:
> Uppon further investigation, it is apparent that the PCRE library
> requires some extra "logic", or more specificaly a table with the
> extra Greek characters which are causing the problem.
PCRE is case-insensitive for Greek characters if it is compiled with
UTF-8 support and Unicode Property Support, and it is then run in UTF-8
mode. Here is a simple test for this:
PCRE version 7.8 2008-09-05
/\x{391}/8i <== the pattern, UTF-8, case-insensitive
\x{391} <== a subject string
0: \x{391} <== it matches
\x{3B1} <== a second subject
0: \x{3b1} <== it also matches
The character U+391 is a Greek capital Alpha; the character U+3B1 is a
lower case alpha. The test matches both. Without the /i option, it does
not match the second string.
I do not know how the version of PCRE that PHP uses is compiled.
Philip
--
Philip Hazel
