1 files changed, 143 insertions, 0 deletions

diff --git a/contrib/unaccent/expected/unaccent.out b/contrib/unaccent/expected/unaccent.out
new file mode 100644
index 0000000..ee0ac71
--- /dev/null
+++ b/contrib/unaccent/expected/unaccent.out
@@ -0,0 +1,143 @@
+CREATE EXTENSION unaccent;
+-- must have a UTF8 database
+SELECT getdatabaseencoding();
+ getdatabaseencoding 
+---------------------
+ UTF8
+(1 row)
+
+SET client_encoding TO 'UTF8';
+SELECT unaccent('foobar');
+ unaccent 
+----------
+ foobar
+(1 row)
+
+SELECT unaccent('ёлка');
+ unaccent 
+----------
+ елка
+(1 row)
+
+SELECT unaccent('ЁЖИК');
+ unaccent 
+----------
+ ЕЖИК
+(1 row)
+
+SELECT unaccent('˃˖˗˜');
+ unaccent 
+----------
+ >+-~
+(1 row)
+
+SELECT unaccent('À');  -- Remove combining diacritical 0x0300
+ unaccent 
+----------
+ A
+(1 row)
+
+SELECT unaccent('℃℉'); -- degree signs
+ unaccent 
+----------
+ °C°F
+(1 row)
+
+SELECT unaccent('℗'); -- sound recording copyright
+ unaccent 
+----------
+ (P)
+(1 row)
+
+SELECT unaccent('unaccent', 'foobar');
+ unaccent 
+----------
+ foobar
+(1 row)
+
+SELECT unaccent('unaccent', 'ёлка');
+ unaccent 
+----------
+ елка
+(1 row)
+
+SELECT unaccent('unaccent', 'ЁЖИК');
+ unaccent 
+----------
+ ЕЖИК
+(1 row)
+
+SELECT unaccent('unaccent', '˃˖˗˜');
+ unaccent 
+----------
+ >+-~
+(1 row)
+
+SELECT unaccent('unaccent', 'À');
+ unaccent 
+----------
+ A
+(1 row)
+
+SELECT unaccent('unaccent', '℃℉');
+ unaccent 
+----------
+ °C°F
+(1 row)
+
+SELECT unaccent('unaccent', '℗');
+ unaccent 
+----------
+ (P)
+(1 row)
+
+SELECT ts_lexize('unaccent', 'foobar');
+ ts_lexize 
+-----------
+ 
+(1 row)
+
+SELECT ts_lexize('unaccent', 'ёлка');
+ ts_lexize 
+-----------
+ {елка}
+(1 row)
+
+SELECT ts_lexize('unaccent', 'ЁЖИК');
+ ts_lexize 
+-----------
+ {ЕЖИК}
+(1 row)
+
+SELECT ts_lexize('unaccent', '˃˖˗˜');
+ ts_lexize 
+-----------
+ {>+-~}
+(1 row)
+
+SELECT ts_lexize('unaccent', 'À');
+ ts_lexize 
+-----------
+ {A}
+(1 row)
+
+SELECT ts_lexize('unaccent', '℃℉');
+ ts_lexize 
+-----------
+ {°C°F}
+(1 row)
+
+SELECT ts_lexize('unaccent', '℗');
+ ts_lexize 
+-----------
+ {(P)}
+(1 row)
+
+-- Controversial case.  Black-Letter Capital H (U+210C) is translated by
+-- Latin-ASCII.xml as 'x', but it should be 'H'.
+SELECT unaccent('ℌ');
+ unaccent 
+----------
+ x
+(1 row)
+