Fix regression tests of unaccent to work without UTF8 support
The tests of unaccent rely on UTF8 characters, and unlike any other test
suite in the tree (fuzzystrmatch, citext, hstore, etc.), they would fail
if run on a database that does not support UTF8 encoding.
This commit fixes the tests of unaccent so as these are skipped when run
on a database without UTF8 support, using the same method as the other
test suits based on \if, getdatabaseencoding() and an alternate output
file.
This has been broken for a long time, but nobody has complained about
that either, so no backpatch is done. This can be reproduced with
something like REGRESS_OPTS="--no-locale --encoding=sql_ascii", for
instance. To defend against that, this module's Makefile and
meson.build enforced a UTF8 encoding without locales, but it did not
offer protection for options given by REGRESS_OPTS. This switch makes
this regression test suite more consistent with all the others, as
well.
Reviewed-by: Peter Eisentraut
Discussion: https://postgr.es/m/ZIq1HUnIV2ksW85x@paquier.xyz
2023-07-04 01:05:00 +02:00
|
|
|
|
/*
|
|
|
|
|
* 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
|
2009-08-18 12:34:39 +02:00
|
|
|
|
|
Fix regression tests of unaccent to work without UTF8 support
The tests of unaccent rely on UTF8 characters, and unlike any other test
suite in the tree (fuzzystrmatch, citext, hstore, etc.), they would fail
if run on a database that does not support UTF8 encoding.
This commit fixes the tests of unaccent so as these are skipped when run
on a database without UTF8 support, using the same method as the other
test suits based on \if, getdatabaseencoding() and an alternate output
file.
This has been broken for a long time, but nobody has complained about
that either, so no backpatch is done. This can be reproduced with
something like REGRESS_OPTS="--no-locale --encoding=sql_ascii", for
instance. To defend against that, this module's Makefile and
meson.build enforced a UTF8 encoding without locales, but it did not
offer protection for options given by REGRESS_OPTS. This switch makes
this regression test suite more consistent with all the others, as
well.
Reviewed-by: Peter Eisentraut
Discussion: https://postgr.es/m/ZIq1HUnIV2ksW85x@paquier.xyz
2023-07-04 01:05:00 +02:00
|
|
|
|
CREATE EXTENSION unaccent;
|
2009-08-18 18:00:50 +02:00
|
|
|
|
|
2019-01-02 18:36:05 +01:00
|
|
|
|
SET client_encoding TO 'UTF8';
|
2009-08-18 12:34:39 +02:00
|
|
|
|
|
|
|
|
|
SELECT unaccent('foobar');
|
2019-01-02 18:36:05 +01:00
|
|
|
|
SELECT unaccent('ёлка');
|
|
|
|
|
SELECT unaccent('ЁЖИК');
|
2019-01-10 06:10:21 +01:00
|
|
|
|
SELECT unaccent('˃˖˗˜');
|
2019-02-01 15:23:01 +01:00
|
|
|
|
SELECT unaccent('À'); -- Remove combining diacritical 0x0300
|
2022-07-05 09:17:51 +02:00
|
|
|
|
SELECT unaccent('℃℉'); -- degree signs
|
|
|
|
|
SELECT unaccent('℗'); -- sound recording copyright
|
2009-08-18 12:34:39 +02:00
|
|
|
|
|
|
|
|
|
SELECT unaccent('unaccent', 'foobar');
|
2019-01-02 18:36:05 +01:00
|
|
|
|
SELECT unaccent('unaccent', 'ёлка');
|
|
|
|
|
SELECT unaccent('unaccent', 'ЁЖИК');
|
2019-01-10 06:10:21 +01:00
|
|
|
|
SELECT unaccent('unaccent', '˃˖˗˜');
|
2019-02-01 15:23:01 +01:00
|
|
|
|
SELECT unaccent('unaccent', 'À');
|
2022-07-05 09:17:51 +02:00
|
|
|
|
SELECT unaccent('unaccent', '℃℉');
|
|
|
|
|
SELECT unaccent('unaccent', '℗');
|
2009-08-18 12:34:39 +02:00
|
|
|
|
|
|
|
|
|
SELECT ts_lexize('unaccent', 'foobar');
|
2019-01-02 18:36:05 +01:00
|
|
|
|
SELECT ts_lexize('unaccent', 'ёлка');
|
|
|
|
|
SELECT ts_lexize('unaccent', 'ЁЖИК');
|
2019-01-10 06:10:21 +01:00
|
|
|
|
SELECT ts_lexize('unaccent', '˃˖˗˜');
|
2019-02-01 15:23:01 +01:00
|
|
|
|
SELECT ts_lexize('unaccent', 'À');
|
2022-07-05 09:17:51 +02:00
|
|
|
|
SELECT ts_lexize('unaccent', '℃℉');
|
|
|
|
|
SELECT ts_lexize('unaccent', '℗');
|
|
|
|
|
|
|
|
|
|
-- Controversial case. Black-Letter Capital H (U+210C) is translated by
|
|
|
|
|
-- Latin-ASCII.xml as 'x', but it should be 'H'.
|
|
|
|
|
SELECT unaccent('ℌ');
|