diff options
| author | Alexander Barkov <bar@mariadb.org> | 2015-07-07 09:15:58 +0400 |
|---|---|---|
| committer | Alexander Barkov <bar@mariadb.org> | 2015-07-07 09:15:58 +0400 |
| commit | e4f8cea35627a8383d5d2d9b985038c960c0b19f (patch) | |
| tree | bfc3f99f0c38c692f673463c2e43b1d5ce2f59a6 /unittest | |
| parent | a5f4412bd44a16ba4d7ed31194716c0b59eecfeb (diff) | |
| download | mariadb-git-e4f8cea35627a8383d5d2d9b985038c960c0b19f.tar.gz | |
MDEV-8419 utf32: compare broken bytes as "greater than any non-broken character"
Diffstat (limited to 'unittest')
| -rw-r--r-- | unittest/strings/strings-t.c | 54 |
1 files changed, 54 insertions, 0 deletions
diff --git a/unittest/strings/strings-t.c b/unittest/strings/strings-t.c index 65a7f1e1155..fe595a5c303 100644 --- a/unittest/strings/strings-t.c +++ b/unittest/strings/strings-t.c @@ -537,6 +537,55 @@ static STRNNCOLL_PARAM strcoll_utf16le_general_ci[]= {CSTR("\x00\xD8\x00\xDC"), CSTR("\xFF\xDB\xFF\xDF"), 0},/* Non-BMP MB4 vs non-BMP MB4 */ {CSTR("\x00\x00"), CSTR("\x00\xD8\x01\xDC"), -1},/* U+0000 vs non-BMP MB4 */ {CSTR("\x00\x00"), CSTR("\xFF\xDB\xFF\xDF"), -1},/* U+0000 vs non-BMP MB4 */ + + {NULL, 0, NULL, 0, 0} +}; + + +static STRNNCOLL_PARAM strcoll_utf32_common[]= +{ + /* Minimum character: U+0000 == _utf32 0x00000000 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\xFF"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00\x00\x00"),-1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00\x20\x00\x00"),-1},/* MB4 vs broken MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\xFF\xFF\xFF\xFF"),-1},/* MB4 vs broken MB4 */ + + /* Minimum non-BMP character: U+10000 == _utf32 0x00010000 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\xFF"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\x00\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\x00\x00\x00"),-1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\x00\x20\x00\x00"),-1},/* MB4 vs broken MB4 */ + {CSTR("\x00\x01\x00\x00"), CSTR("\xFF\xFF\xFF\xFF"),-1},/* MB4 vs broken MB4 */ + + /* Maximum character: U+10FFFF == _utf32 0x0010FFFF */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\xFF"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\x00\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\x00\x00\x00"), -1}, /* MB4 vs incomplete MB4 */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\x20\x00\x00\x00"),-1},/* MB4 vs broken MB3 */ + {CSTR("\x00\x10\xFF\xFF"), CSTR("\xFF\xFF\xFF\xFF"),-1},/* MB4 vs broken MB4 */ + + + /* Broken MB4 vs incomplete/broken MB3 */ + {CSTR("\x00\x20\x00\x00"), CSTR("\x00"), 1}, /* Broken MB4 vs incomplete MB4 */ + {CSTR("\x00\x20\x00\x00"), CSTR("\x00\x00"), 1}, /* Broken MB4 vs incomplete MB4 */ + {CSTR("\x00\x20\x00\x00"), CSTR("\x00\x00\x00"), 1}, /* Broken MB4 vs incomplete MB4 */ + {CSTR("\x00\x20\x00\x00"), CSTR("\x00\x20\x00\x01"),-1},/* Broken MB4 vs broken MB4 */ + + {NULL, 0, NULL, 0, 0} +}; + + +static STRNNCOLL_PARAM strcoll_utf32_general_ci[]= +{ + /* Two non-BMP characters are compared as equal */ + {CSTR("\x00\x01\x00\x00"), CSTR("\x00\x01\x00\x01"), 0}, /* non-BMP MB4 vs non-BMP MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00\x01\x00\x00"), -1}, /* U+0000 vs non-BMP MB4 */ + {CSTR("\x00\x00\x00\x00"), CSTR("\x00\x01\x00\x01"), -1}, /* U+0000 vs non-BMP MB4 */ + {NULL, 0, NULL, 0, 0} }; @@ -688,6 +737,11 @@ test_strcollsp() failed+= strcollsp(&my_charset_utf16le_bin, strcoll_utf16le_space); failed+= strcollsp(&my_charset_utf16le_bin, strcoll_utf16le_common); #endif +#ifdef HAVE_CHARSET_utf32 + failed+= strcollsp(&my_charset_utf32_general_ci, strcoll_utf32_common); + failed+= strcollsp(&my_charset_utf32_general_ci, strcoll_utf32_general_ci); + failed+= strcollsp(&my_charset_utf32_bin, strcoll_utf32_common); +#endif #ifdef HAVE_CHARSET_utf8 failed+= strcollsp(&my_charset_utf8_general_ci, strcoll_utf8mb3_common); failed+= strcollsp(&my_charset_utf8_general_mysql500_ci, strcoll_utf8mb3_common); |