summaryrefslogtreecommitdiffstats
path: root/test/icu.test
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-13 14:07:11 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-13 14:07:11 +0000
commit63847496f14c813a5d80efd5b7de0f1294ffe1e3 (patch)
tree01c7571c7c762ceee70638549a99834fdd7c411b /test/icu.test
parentInitial commit. (diff)
downloadsqlite3-63847496f14c813a5d80efd5b7de0f1294ffe1e3.tar.xz
sqlite3-63847496f14c813a5d80efd5b7de0f1294ffe1e3.zip
Adding upstream version 3.45.1.upstream/3.45.1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'test/icu.test')
-rw-r--r--test/icu.test167
1 files changed, 167 insertions, 0 deletions
diff --git a/test/icu.test b/test/icu.test
new file mode 100644
index 0000000..644cbb1
--- /dev/null
+++ b/test/icu.test
@@ -0,0 +1,167 @@
+# 2007 May 1
+#
+# The author disclaims copyright to this source code. In place of
+# a legal notice, here is a blessing:
+#
+# May you do good and not evil.
+# May you find forgiveness for yourself and forgive others.
+# May you share freely, never taking more than you give.
+#
+#***********************************************************************
+#
+# $Id: icu.test,v 1.2 2008/07/12 14:52:20 drh Exp $
+#
+
+set testdir [file dirname $argv0]
+source $testdir/tester.tcl
+
+ifcapable !icu&&!icu_collations {
+ finish_test
+ return
+}
+
+# Create a table to work with.
+#
+execsql {CREATE TABLE test1(i1 int, i2 int, r1 real, r2 real, t1 text, t2 text)}
+execsql {INSERT INTO test1 VALUES(1,2,1.1,2.2,'hello','world')}
+proc test_expr {name settings expr result} {
+ do_test $name [format {
+ lindex [db eval {
+ BEGIN;
+ UPDATE test1 SET %s;
+ SELECT %s FROM test1;
+ ROLLBACK;
+ }] 0
+ } $settings $expr] $result
+}
+
+ifcapable icu {
+
+ # Tests of the REGEXP operator.
+ #
+ test_expr icu-1.1 {i1='hello'} {i1 REGEXP 'hello'} 1
+ test_expr icu-1.2 {i1='hello'} {i1 REGEXP '.ello'} 1
+ test_expr icu-1.3 {i1='hello'} {i1 REGEXP '.ell'} 0
+ test_expr icu-1.4 {i1='hello'} {i1 REGEXP '.ell.*'} 1
+ test_expr icu-1.5 {i1=NULL} {i1 REGEXP '.ell.*'} {}
+
+ # Some non-ascii characters with defined case mappings
+ #
+ set ::EGRAVE "\xC8"
+ set ::egrave "\xE8"
+
+ set ::OGRAVE "\xD2"
+ set ::ograve "\xF2"
+
+ # That German letter that looks a bit like a B. The
+ # upper-case version of which is "SS" (two characters).
+ #
+ set ::szlig "\xDF"
+
+ # Tests of the upper()/lower() functions.
+ #
+ test_expr icu-2.1 {i1='HellO WorlD'} {upper(i1)} {HELLO WORLD}
+ test_expr icu-2.2 {i1='HellO WorlD'} {lower(i1)} {hello world}
+ test_expr icu-2.3 {i1=$::egrave} {lower(i1)} $::egrave
+ test_expr icu-2.4 {i1=$::egrave} {upper(i1)} $::EGRAVE
+ test_expr icu-2.5 {i1=$::ograve} {lower(i1)} $::ograve
+ test_expr icu-2.6 {i1=$::ograve} {upper(i1)} $::OGRAVE
+ test_expr icu-2.3 {i1=$::EGRAVE} {lower(i1)} $::egrave
+ test_expr icu-2.4 {i1=$::EGRAVE} {upper(i1)} $::EGRAVE
+ test_expr icu-2.5 {i1=$::OGRAVE} {lower(i1)} $::ograve
+ test_expr icu-2.6 {i1=$::OGRAVE} {upper(i1)} $::OGRAVE
+
+ test_expr icu-2.7 {i1=$::szlig} {upper(i1)} "SS"
+ test_expr icu-2.8 {i1='SS'} {lower(i1)} "ss"
+
+ do_execsql_test icu-2.9 {
+ SELECT upper(char(0xfb04,0xfb04,0xfb04,0xfb04));
+ } {FFLFFLFFLFFL}
+
+ # In turkish (locale="tr_TR"), the lower case version of I
+ # is "small dotless i" (code point 0x131 (decimal 305)).
+ #
+ set ::small_dotless_i "\u0131"
+ test_expr icu-3.1 {i1='I'} {lower(i1)} "i"
+ test_expr icu-3.2 {i1='I'} {lower(i1, 'tr_tr')} $::small_dotless_i
+ test_expr icu-3.3 {i1='I'} {lower(i1, 'en_AU')} "i"
+}
+
+#--------------------------------------------------------------------
+# Test the collation sequence function.
+#
+do_test icu-4.1 {
+ execsql {
+ CREATE TABLE fruit(name);
+ INSERT INTO fruit VALUES('plum');
+ INSERT INTO fruit VALUES('cherry');
+ INSERT INTO fruit VALUES('apricot');
+ INSERT INTO fruit VALUES('peach');
+ INSERT INTO fruit VALUES('chokecherry');
+ INSERT INTO fruit VALUES('yamot');
+ }
+} {}
+do_test icu-4.2 {
+ execsql {
+ SELECT icu_load_collation('en_US', 'AmericanEnglish');
+ SELECT icu_load_collation('lt_LT', 'Lithuanian');
+ }
+ execsql {
+ SELECT name FROM fruit ORDER BY name COLLATE AmericanEnglish ASC;
+ }
+} {apricot cherry chokecherry peach plum yamot}
+
+
+# Test collation using Lithuanian rules. In the Lithuanian
+# alphabet, "y" comes right after "i".
+#
+do_test icu-4.3 {
+ execsql {
+ SELECT name FROM fruit ORDER BY name COLLATE Lithuanian ASC;
+ }
+} {apricot cherry chokecherry yamot peach plum}
+
+#-------------------------------------------------------------------------
+# Test that it is not possible to call the ICU regex() function with
+# anything other than exactly two arguments. See also:
+#
+# http://src.chromium.org/viewvc/chrome/trunk/src/third_party/sqlite/icu-regexp.patch?revision=34807&view=markup
+#
+ifcapable icu {
+ do_catchsql_test icu-5.1 { SELECT regexp('a[abc]c.*', 'abc') } {0 1}
+ do_catchsql_test icu-5.2 {
+ SELECT regexp('a[abc]c.*')
+ } {1 {wrong number of arguments to function regexp()}}
+ do_catchsql_test icu-5.3 {
+ SELECT regexp('a[abc]c.*', 'abc', 'c')
+ } {1 {wrong number of arguments to function regexp()}}
+ do_catchsql_test icu-5.4 {
+ SELECT 'abc' REGEXP 'a[abc]c.*'
+ } {0 1}
+ do_catchsql_test icu-5.5 {SELECT 'abc' REGEXP } {1 {incomplete input}}
+ do_catchsql_test icu-5.6 {SELECT 'abc' REGEXP, 1} {1 {near ",": syntax error}}
+
+ do_malloc_test icu-6.10 -sqlbody {
+ SELECT upper(char(0xfb04,0xdf,0xfb04,0xe8,0xfb04));
+ }
+}
+
+# 2020-03-19
+# The ESCAPE clause on LIKE takes precedence over wildcards
+#
+do_execsql_test idu-6.0 {
+ DROP TABLE IF EXISTS t1;
+ CREATE TABLE t1(id INTEGER PRIMARY KEY, x TEXT);
+ INSERT INTO t1 VALUES
+ (1,'abcde'),
+ (2,'abc_'),
+ (3,'abc__'),
+ (4,'abc%'),
+ (5,'abc%%');
+ SELECT id FROM t1 WHERE x LIKE 'abc%%' ESCAPE '%';
+} {4}
+do_execsql_test icu-6.1 {
+ SELECT id FROM t1 WHERE x LIKE 'abc__' ESCAPE '_';
+} {2}
+
+finish_test