-- regression tests for icu_ext
CREATE EXTENSION icu_ext;
-- Check that the database has the built-in ICU collations
-- required by the tests
SELECT collname FROM pg_collation WHERE collname IN
 ('und-x-icu', 'en-x-icu')
ORDER BY collname;
 collname  
-----------
 en-x-icu
 und-x-icu
(2 rows)

-- icu_char_name
SELECT c, to_hex(ascii(c)), icu_char_name(c)
   FROM regexp_split_to_table('El Niño', '') as c;
 c | to_hex |          icu_char_name          
---+--------+---------------------------------
 E | 45     | LATIN CAPITAL LETTER E
 l | 6c     | LATIN SMALL LETTER L
   | 20     | SPACE
 N | 4e     | LATIN CAPITAL LETTER N
 i | 69     | LATIN SMALL LETTER I
 ñ | f1     | LATIN SMALL LETTER N WITH TILDE
 o | 6f     | LATIN SMALL LETTER O
(7 rows)

-- icu_character_boundaries
SELECT * FROM icu_character_boundaries('Ete'||E'\u0301', 'fr') as chars;
 chars 
-------
 E
 t
 é
(3 rows)

-- icu_collation_attributes
SELECT * FROM icu_collation_attributes('en') WHERE attribute <> 'version';
  attribute  |  value   
-------------+----------
 displayname | English
 kn          | false
 kb          | false
 kk          | false
 ka          | noignore
 ks          | level3
 kf          | false
 kc          | false
(8 rows)

-- icu_compare
SELECT icu_compare('abcé', 'abce', 'en@colStrength=primary;colCaseLevel=yes');
 icu_compare 
-------------
           0
(1 row)

SELECT icu_compare('Abcé', 'abce' COLLATE "en-x-icu");
 icu_compare 
-------------
           1
(1 row)

-- icu_confusable_strings_check
SELECT txt, icu_confusable_strings_check('phil', txt) AS confusable
    FROM (VALUES ('phiL'), ('phiI'), ('phi1'), (E'ph\u0131l')) AS s(txt);
 txt  | confusable 
------+------------
 phiL | f
 phiI | t
 phi1 | t
 phıl | t
(4 rows)

-- icu_line_boundaries
SELECT *,convert_to( contents, 'utf-8')
FROM icu_line_boundaries(
$$Thus much let me avow
You are not wrong, who deem
That my days have been a dream;
Yet if hope has flown away
In a night, or in a day,$$
, 'en');
 tag | contents |    convert_to    
-----+----------+------------------
   0 | Thus     | \x5468757320
   0 | much     | \x6d75636820
   0 | let      | \x6c657420
   0 | me       | \x6d6520
 100 | avow    +| \x61766f770a
     |          | 
   0 | You      | \x596f7520
   0 | are      | \x61726520
   0 | not      | \x6e6f7420
   0 | wrong,   | \x77726f6e672c20
   0 | who      | \x77686f20
 100 | deem    +| \x6465656d0a
     |          | 
   0 | That     | \x5468617420
   0 | my       | \x6d7920
   0 | days     | \x6461797320
   0 | have     | \x6861766520
   0 | been     | \x6265656e20
   0 | a        | \x6120
 100 | dream;  +| \x647265616d3b0a
     |          | 
   0 | Yet      | \x59657420
   0 | if       | \x696620
   0 | hope     | \x686f706520
   0 | has      | \x68617320
   0 | flown    | \x666c6f776e20
 100 | away    +| \x617761790a
     |          | 
   0 | In       | \x496e20
   0 | a        | \x6120
   0 | night,   | \x6e696768742c20
   0 | or       | \x6f7220
   0 | in       | \x696e20
   0 | a        | \x6120
   0 | day,     | \x6461792c
(31 rows)

-- icu_number_spellout
SELECT loc, icu_number_spellout(1234, loc)
    FROM (values ('en'),('fr'),('de'),('ru'),('ja')) AS s(loc);
 loc |            icu_number_spellout            
-----+-------------------------------------------
 en  | one thousand two hundred thirty-four
 fr  | mille deux cent trente-quatre
 de  | ein­tausend­zwei­hundert­vier­und­dreißig
 ru  | одна тысяча двести тридцать четыре
 ja  | 千二百三十四
(5 rows)

-- icu_replace
SELECT n,
   icu_replace(
      n,
     'jeanrene',
     '{firstname}',
     'und@colStrength=primary;colAlternate=shifted')
FROM (values('jeanrenédupont'),('Jean-René  Dupont')) as s(n)
ORDER BY n COLLATE "C";
         n         |     icu_replace     
-------------------+---------------------
 Jean-René  Dupont | {firstname}  Dupont
 jeanrenédupont    | {firstname}dupont
(2 rows)

-- icu_sentence_boundaries
SELECT * FROM icu_sentence_boundaries('Call me Mr. Brown. It''s a movie.',
 'en@ss=standard');
 tag |      contents       
-----+---------------------
   0 | Call me Mr. Brown. 
   0 | It's a movie.
(2 rows)

-- icu_strpos
SELECT v,icu_strpos('hey rene', v, 'und@colStrength=primary;colAlternate=shifted')
FROM (VALUES ('René'), ('rené'), ('Rene'), ('n'), ('në'), ('no'), (''), (null))
 AS s(v)
ORDER BY v COLLATE "C";
  v   | icu_strpos 
------+------------
      |          1
 Rene |          5
 René |          5
 n    |          7
 no   |          0
 në   |          7
 rené |          5
      |           
(8 rows)

-- icu_transform
SELECT icu_transform('10\N{SUPERSCRIPT MINUS}\N{SUPERSCRIPT FOUR}'
		   '\N{MICRO SIGN}m = 1 \N{ANGSTROM SIGN}',
		 'Name-Any');
 icu_transform 
---------------
 10⁻⁴µm = 1 Å
(1 row)

SELECT icu_transform('Ich muß essen.', '[:^ascii:]; Hex');
    icu_transform    
---------------------
 Ich mu\u00DF essen.
(1 row)

-- icu_word_boundaries
SELECT * FROM icu_word_boundaries($$Do you like O'Reilly books?$$, 'en');
 tag | contents 
-----+----------
 200 | Do
   0 |  
 200 | you
   0 |  
 200 | like
   0 |  
 200 | O'Reilly
   0 |  
 200 | books
   0 | ?
(10 rows)