Import SQLite 3.10.2.

third_party/sqlite/src/test/fts4unicode.test

 # 2012 May 25
 #
 # 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.
 #
 #*************************************************************************
@@ skipping 344 matching lines @@
   set txt "${whitespace}hello${whitespace}world${whitespace}"
   uplevel [list do_test $tn [list do_tokenize $tokenizer $txt] {hello world}]
 }
 
 set tokenizers [list unicode61]
 ifcapable icu { lappend tokenizers icu }
 
 # Some tests to check that the tokenizers can both identify white-space 
 # codepoints. All codepoints tested below are of type "Zs" in the
 # UnicodeData.txt file.
+#
+# Note that codepoint 6158 has changed from Zs to Cf in recent versions
+# of UnicodeData.txt.  So take that into account for the "icu" tests.
+#
 foreach T $tokenizers {
   do_isspace_test 6.$T.1 $T    32
   do_isspace_test 6.$T.2 $T    160
   do_isspace_test 6.$T.3 $T    5760
-  do_isspace_test 6.$T.4 $T    6158
+  if {$T!="icu"} {
+    do_isspace_test 6.$T.4 $T    6158
+  }
   do_isspace_test 6.$T.5 $T    8192
   do_isspace_test 6.$T.6 $T    8193
   do_isspace_test 6.$T.7 $T    8194
   do_isspace_test 6.$T.8 $T    8195
   do_isspace_test 6.$T.9 $T    8196
   do_isspace_test 6.$T.10 $T    8197
   do_isspace_test 6.$T.11 $T    8198
   do_isspace_test 6.$T.12 $T    8199
   do_isspace_test 6.$T.13 $T    8200
   do_isspace_test 6.$T.14 $T    8201
   do_isspace_test 6.$T.15 $T    8202
   do_isspace_test 6.$T.16 $T    8239
   do_isspace_test 6.$T.17 $T    8287
   do_isspace_test 6.$T.18 $T   12288
 
-  do_isspace_test 6.$T.19 $T   {32 160 5760 6158}
+  if {$T!="icu"} {
+    do_isspace_test 6.$T.19 $T   {32 160 5760 6158}
+  } else {
+    do_isspace_test 6.$T.19 $T   {32 160 5760 8192}
+  }
   do_isspace_test 6.$T.20 $T   {8192 8193 8194 8195}
   do_isspace_test 6.$T.21 $T   {8196 8197 8198 8199}
   do_isspace_test 6.$T.22 $T   {8200 8201 8202 8239}
   do_isspace_test 6.$T.23 $T   {8287 12288}
 }
 
 #-------------------------------------------------------------------------
 # Test that the private use ranges are treated as alphanumeric.
 #
 foreach {tn1 c} {
@@ skipping 153 matching lines @@
 do_execsql_test 11.1 {
   CREATE VIRTUAL TABLE ft1 USING fts3tokenize(
     "unicode61", "tokenchars=@.", "separators=1234567890"
   );
   SELECT token FROM ft1 WHERE input = 'berlin@street123sydney.road';
 } {
   berlin@street sydney.road
 }
 
 finish_test