Unaccent dictionary.

Author: feodor

contrib/Makefile

@@ -1,4 +1,4 @@
-# $PostgreSQL: pgsql/contrib/Makefile,v 1.88 2009/08/07 20:50:21 petere Exp $
+# $PostgreSQL: pgsql/contrib/Makefile,v 1.89 2009/08/18 10:34:39 teodor Exp $
 
 subdir = contrib
 top_builddir = ..
@@ -39,6 +39,7 @@ SUBDIRS = \
 		tablefunc	\
 		test_parser	\
 		tsearch2	\
+		unaccent	\
 		vacuumlo
 
 ifeq ($(with_openssl),yes)

contrib/README

@@ -169,6 +169,10 @@ tsearch2 -
 	Pavel Stehule <[email protected]>, based on code originally by
 	Teodor Sigaev <[email protected]> and Oleg Bartunov <[email protected]>.
 
+unaccent -
+	Unaccent dictionary for text search
+	Teodor Sigaev <[email protected]> and Oleg Bartunov <[email protected]>.
+
 uuid-ossp -
 	UUID generation functions
 	by Peter Eisentraut <[email protected]>

contrib/unaccent/Makefile

@@ -0,0 +1,24 @@
+# $PostgreSQL: pgsql/contrib/unaccent/Makefile,v 1.1 2009/08/18 10:34:39 teodor Exp $
+
+MODULE_big = unaccent
+OBJS = unaccent.o
+
+DATA_built = unaccent.sql
+DATA = uninstall_unaccent.sql
+DATA_TSEARCH = unaccent.rules
+REGRESS = unaccent
+
+
+ifdef USE_PGXS
+PG_CONFIG = pg_config
+PGXS := $(shell $(PG_CONFIG) --pgxs)
+include $(PGXS)
+else
+subdir = contrib/pg_trgm
+top_builddir = ../..
+include $(top_builddir)/src/Makefile.global
+include $(top_srcdir)/contrib/contrib-global.mk
+endif
+
+#redefine REGRESS_OPTS because of needings of UTF8 database
+REGRESS_OPTS = --dbname=$(CONTRIB_TESTDB) --multibyte=UTF8 --no-locale 

contrib/unaccent/expected/unaccent.out

@@ -0,0 +1,58 @@
+SET client_min_messages = warning;
+\set ECHO none
+RESET client_min_messages;
+SET client_encoding TO 'KOI8';
+SELECT unaccent('foobar');
+ unaccent 
+----------
+ foobar
+(1 row)
+
+SELECT unaccent('����');
+ unaccent 
+----------
+ ����
+(1 row)
+
+SELECT unaccent('����');
+ unaccent 
+----------
+ ����
+(1 row)
+
+SELECT unaccent('unaccent', 'foobar');
+ unaccent 
+----------
+ foobar
+(1 row)
+
+SELECT unaccent('unaccent', '����');
+ unaccent 
+----------
+ ����
+(1 row)
+
+SELECT unaccent('unaccent', '����');
+ unaccent 
+----------
+ ����
+(1 row)
+
+SELECT ts_lexize('unaccent', 'foobar');
+ ts_lexize 
+-----------
+ 
+(1 row)
+
+SELECT ts_lexize('unaccent', '����');
+ ts_lexize 
+-----------
+ {����}
+(1 row)
+
+SELECT ts_lexize('unaccent', '����');
+ ts_lexize 
+-----------
+ {����}
+(1 row)
+

contrib/unaccent/sql/unaccent.sql

@@ -0,0 +1,19 @@
+SET client_min_messages = warning;
+\set ECHO none
+\i unaccent.sql
+\set ECHO all
+RESET client_min_messages;
+
+SET client_encoding TO 'KOI8';
+
+SELECT unaccent('foobar');
+SELECT unaccent('Ｌ肆');
+SELECT unaccent('出殡');
+
+SELECT unaccent('unaccent', 'foobar');
+SELECT unaccent('unaccent', 'Ｌ肆');
+SELECT unaccent('unaccent', '出殡');
+
+SELECT ts_lexize('unaccent', 'foobar');
+SELECT ts_lexize('unaccent', 'Ｌ肆');
+SELECT ts_lexize('unaccent', '出殡');