summaryrefslogtreecommitdiff
path: root/unittest/strings/strings-t.c
diff options
context:
space:
mode:
Diffstat (limited to 'unittest/strings/strings-t.c')
-rw-r--r--unittest/strings/strings-t.c114
1 files changed, 114 insertions, 0 deletions
diff --git a/unittest/strings/strings-t.c b/unittest/strings/strings-t.c
new file mode 100644
index 00000000000..278f42e56a5
--- /dev/null
+++ b/unittest/strings/strings-t.c
@@ -0,0 +1,114 @@
+/* Copyright 2000, 2010, Oracle and/or its affiliates. All rights reserved.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; version 2 of the License.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */
+
+#include <tap.h>
+#include <my_global.h>
+#include <my_sys.h>
+
+
+/*
+ Test that like_range() returns well-formed results.
+*/
+static int
+test_like_range_for_charset(CHARSET_INFO *cs, const char *src, size_t src_len)
+{
+ char min_str[32], max_str[32];
+ size_t min_len, max_len, min_well_formed_len, max_well_formed_len;
+ int error= 0;
+
+ cs->coll->like_range(cs, src, src_len, '\\', '_', '%',
+ sizeof(min_str), min_str, max_str, &min_len, &max_len);
+ diag("min_len=%d\tmax_len=%d\t%s", (int) min_len, (int) max_len, cs->name);
+ min_well_formed_len= cs->cset->well_formed_len(cs,
+ min_str, min_str + min_len,
+ 10000, &error);
+ max_well_formed_len= cs->cset->well_formed_len(cs,
+ max_str, max_str + max_len,
+ 10000, &error);
+ if (min_len != min_well_formed_len)
+ diag("Bad min_str: min_well_formed_len=%d min_str[%d]=0x%02X",
+ (int) min_well_formed_len, (int) min_well_formed_len,
+ (uchar) min_str[min_well_formed_len]);
+ if (max_len != max_well_formed_len)
+ diag("Bad max_str: max_well_formed_len=%d max_str[%d]=0x%02X",
+ (int) max_well_formed_len, (int) max_well_formed_len,
+ (uchar) max_str[max_well_formed_len]);
+ return
+ min_len == min_well_formed_len &&
+ max_len == max_well_formed_len ? 0 : 1;
+}
+
+
+static CHARSET_INFO *charset_list[]=
+{
+#ifdef HAVE_CHARSET_big5
+ &my_charset_big5_chinese_ci,
+ &my_charset_big5_bin,
+#endif
+#ifdef HAVE_CHARSET_euckr
+ &my_charset_euckr_korean_ci,
+ &my_charset_euckr_bin,
+#endif
+#ifdef HAVE_CHARSET_gb2312
+ &my_charset_gb2312_chinese_ci,
+ &my_charset_gb2312_bin,
+#endif
+#ifdef HAVE_CHARSET_gbk
+ &my_charset_gbk_chinese_ci,
+ &my_charset_gbk_bin,
+#endif
+#ifdef HAVE_CHARSET_latin1
+ &my_charset_latin1,
+ &my_charset_latin1_bin,
+#endif
+#ifdef HAVE_CHARSET_sjis
+ &my_charset_sjis_japanese_ci,
+ &my_charset_sjis_bin,
+#endif
+#ifdef HAVE_CHARSET_tis620
+ &my_charset_tis620_thai_ci,
+ &my_charset_tis620_bin,
+#endif
+#ifdef HAVE_CHARSET_ujis
+ &my_charset_ujis_japanese_ci,
+ &my_charset_ujis_bin,
+#endif
+#ifdef HAVE_CHARSET_utf8
+ &my_charset_utf8_general_ci,
+ &my_charset_utf8_unicode_ci,
+ &my_charset_utf8_bin,
+#endif
+};
+
+
+int main()
+{
+ size_t i, failed= 0;
+
+ plan(1);
+ diag("Testing my_like_range_xxx() functions");
+
+ for (i= 0; i < array_elements(charset_list); i++)
+ {
+ CHARSET_INFO *cs= charset_list[i];
+ if (test_like_range_for_charset(cs, "abc%", 4))
+ {
+ ++failed;
+ diag("Failed for %s", cs->name);
+ }
+ }
+ ok(failed == 0, "Testing my_like_range_xxx() functions");
+ return exit_status();
+}