/*
 * This test must be run in a database with UTF-8 encoding,
 * because other encodings don't support all the characters used.
 */

SELECT getdatabaseencoding() <> 'UTF8'
       AS skip_test \gset
\if :skip_test
\quit
\endif

set client_encoding = utf8;

set standard_conforming_strings = on;


-- Run the Tcl test cases that require Unicode

-- expectMatch 9.44 EMP* {a[\u00fe-\u0507][\u00ff-\u0300]b} \
--  "a\u0102\u02ffb" "a\u0102\u02ffb"
select * from test_regex('a[\u00fe-\u0507][\u00ff-\u0300]b', E'a\u0102\u02ffb', 'EMP*');

-- expectMatch 13.27 P "a\\U00001234x" "a\u1234x" "a\u1234x"
select * from test_regex('a\U00001234x', E'a\u1234x', 'P');
-- expectMatch 13.28 P {a\U00001234x} "a\u1234x" "a\u1234x"
select * from test_regex('a\U00001234x', E'a\u1234x', 'P');
-- expectMatch 13.29 P "a\\U0001234x" "a\u1234x" "a\u1234x"
-- Tcl has relaxed their code to allow 1-8 hex digits, but Postgres hasn't
select * from test_regex('a\U0001234x', E'a\u1234x', 'P');
-- expectMatch 13.30 P {a\U0001234x} "a\u1234x" "a\u1234x"
-- Tcl has relaxed their code to allow 1-8 hex digits, but Postgres hasn't
select * from test_regex('a\U0001234x', E'a\u1234x', 'P');
-- expectMatch 13.31 P "a\\U000012345x" "a\u12345x" "a\u12345x"
select * from test_regex('a\U000012345x', E'a\u12345x', 'P');
-- expectMatch 13.32 P {a\U000012345x} "a\u12345x" "a\u12345x"
select * from test_regex('a\U000012345x', E'a\u12345x', 'P');
-- expectMatch 13.33 P "a\\U1000000x" "a\ufffd0x" "a\ufffd0x"
-- Tcl allows this as a standalone character, but Postgres doesn't
select * from test_regex('a\U1000000x', E'a\ufffd0x', 'P');
-- expectMatch 13.34 P {a\U1000000x} "a\ufffd0x" "a\ufffd0x"
-- Tcl allows this as a standalone character, but Postgres doesn't
select * from test_regex('a\U1000000x', E'a\ufffd0x', 'P');


-- Additional tests, not derived from Tcl

-- Exercise logic around high character ranges a bit more
select * from test_regex('a
  [\u1000-\u1100]*
  [\u3000-\u3100]*
  [\u1234-\u25ff]+
  [\u2000-\u35ff]*
  [\u2600-\u2f00]*
  \u1236\u1236x',
  E'a\u1234\u1236\u1236x', 'xEMP');

select * from test_regex('[[:alnum:]]*[[:upper:]]*[\u1000-\u2000]*\u1237',
  E'\u1500\u1237', 'ELMP');
select * from test_regex('[[:alnum:]]*[[:upper:]]*[\u1000-\u2000]*\u1237',
  E'A\u1239', 'ELMP');
select * from test_regex('[[:alnum:]]*[[:upper:]]*[\u1000-\u2000]*\u1237',
  E'\u1500\u1237', 'iELMP');

-- systematically test char classes
select * from test_regex('[[:alnum:]]+',  E'x*\u1500\u1237', 'L');
select * from test_regex('[[:alpha:]]+',  E'x*\u1500\u1237', 'L');
select * from test_regex('[[:ascii:]]+',  E'x\u1500\u1237', 'L');
select * from test_regex('[[:blank:]]+',  E'x \t\u1500\u1237', 'L');
select * from test_regex('[[:cntrl:]]+',  E'x\u1500\u1237', 'L');
select * from test_regex('[[:digit:]]+',  E'x9\u1500\u1237', 'L');
select * from test_regex('[[:graph:]]+',  E'x\u1500\u1237', 'L');
select * from test_regex('[[:lower:]]+',  E'x\u1500\u1237', 'L');
select * from test_regex('[[:print:]]+',  E'x\u1500\u1237', 'L');
select * from test_regex('[[:punct:]]+',  E'x.\u1500\u1237', 'L');
select * from test_regex('[[:space:]]+',  E'x \t\u1500\u1237', 'L');
select * from test_regex('[[:upper:]]+',  E'xX\u1500\u1237', 'L');
select * from test_regex('[[:xdigit:]]+',  E'xa9\u1500\u1237', 'L');
select * from test_regex('[[:word:]]+',  E'x_*\u1500\u1237', 'L');