From 44e73a498c5f3ad7e39dc70821c2d8316d3207f2 Mon Sep 17 00:00:00 2001
From: Michael Paquier <michael@paquier.xyz>
Date: Tue, 4 Jul 2023 08:05:00 +0900
Subject: [PATCH] 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
---
 contrib/unaccent/Makefile                |  4 ----
 contrib/unaccent/expected/unaccent.out   | 16 +++++++++-------
 contrib/unaccent/expected/unaccent_1.out |  8 ++++++++
 contrib/unaccent/meson.build             |  1 -
 contrib/unaccent/sql/unaccent.sql        | 14 +++++++++++---
 5 files changed, 28 insertions(+), 15 deletions(-)
 create mode 100644 contrib/unaccent/expected/unaccent_1.out

diff --git a/contrib/unaccent/Makefile b/contrib/unaccent/Makefile
index d6c466e07a..652a3e774c 100644
--- a/contrib/unaccent/Makefile
+++ b/contrib/unaccent/Makefile
@@ -12,10 +12,6 @@ PGFILEDESC = "unaccent - text search dictionary that removes accents"
 
 REGRESS = unaccent
 
-# We need a UTF8 database
-ENCODING = UTF8
-NO_LOCALE = 1
-
 ifdef USE_PGXS
 PG_CONFIG = pg_config
 PGXS := $(shell $(PG_CONFIG) --pgxs)
diff --git a/contrib/unaccent/expected/unaccent.out b/contrib/unaccent/expected/unaccent.out
index ee0ac71a1c..f080707c4a 100644
--- a/contrib/unaccent/expected/unaccent.out
+++ b/contrib/unaccent/expected/unaccent.out
@@ -1,11 +1,13 @@
+/*
+ * 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
 CREATE EXTENSION unaccent;
--- must have a UTF8 database
-SELECT getdatabaseencoding();
- getdatabaseencoding 
----------------------
- UTF8
-(1 row)
-
 SET client_encoding TO 'UTF8';
 SELECT unaccent('foobar');
  unaccent 
diff --git a/contrib/unaccent/expected/unaccent_1.out b/contrib/unaccent/expected/unaccent_1.out
new file mode 100644
index 0000000000..37aead89c0
--- /dev/null
+++ b/contrib/unaccent/expected/unaccent_1.out
@@ -0,0 +1,8 @@
+/*
+ * 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
diff --git a/contrib/unaccent/meson.build b/contrib/unaccent/meson.build
index bd629ec090..a93bf10468 100644
--- a/contrib/unaccent/meson.build
+++ b/contrib/unaccent/meson.build
@@ -37,6 +37,5 @@ tests += {
     'sql': [
       'unaccent',
     ],
-    'regress_args': ['--no-locale', '--encoding=UTF8'],
   },
 }
diff --git a/contrib/unaccent/sql/unaccent.sql b/contrib/unaccent/sql/unaccent.sql
index 3fc0c706be..663646c1ac 100644
--- a/contrib/unaccent/sql/unaccent.sql
+++ b/contrib/unaccent/sql/unaccent.sql
@@ -1,7 +1,15 @@
-CREATE EXTENSION unaccent;
+/*
+ * This test must be run in a database with UTF-8 encoding,
+ * because other encodings don't support all the characters used.
+ */
 
--- must have a UTF8 database
-SELECT getdatabaseencoding();
+SELECT getdatabaseencoding() <> 'UTF8'
+       AS skip_test \gset
+\if :skip_test
+\quit
+\endif
+
+CREATE EXTENSION unaccent;
 
 SET client_encoding TO 'UTF8';