Hi,
I modified two files in postgresql-7.1.3/src/backend/regex/ and in
postgresql-7.1.3/src/include/regex/
so 'character class' (eg. [[:alnum:]], [[:alpha:]], etc.) now should
support 'locale' settings.
http://galaxy.metacerca.it/~anto/pgslq_7_1_3_regex_locale.tar.gz (~14
KB, 2 files)
It is not a great work and do not support multibyte, but, for me, is
sufficient to isolate, for example, an italian word containing ascii
char > 127.
For example: select T from tab where T ~*
'(^|[^[:alnum:]]+)citt[[:alnum:]]*([^[:alnum:]]+|$)';
now match the word 'città' in a string like 'vado in città', 'città'
etc..
PS: excuse my poor english
Regards
Antonello
--
_______________________________________________________
Antonello Nocchi CERCA.COM S.r.l
Via dello Stadio, 19 Tel. +39-0578-75.77.77
53045 - Montepulciano (Siena) Tel. +39-0578-71.67.09
ITALY Fax. +39-0578-71.51.89
antonello@cerca.com http://www.cerca.com