#3290: Can't convert sjis&ujis half-width katakana correctly

5 files changed, 96 insertions, 1 deletions

diff --git a/mysql-test/include/have_sjis.inc b/mysql-test/include/have_sjis.inc
new file mode 100644
index 00000000000..0d580a3a232
--- /dev/null
+++ b/mysql-test/include/have_sjis.inc
@@ -0,0 +1,4 @@
+-- require r/have_sjis.require
+disable_query_log;
+show collation like "sjis_japanese_ci";
+enable_query_log;
diff --git a/mysql-test/r/ctype_sjis.result b/mysql-test/r/ctype_sjis.result
new file mode 100644
index 00000000000..1e3e28784a5
--- /dev/null
+++ b/mysql-test/r/ctype_sjis.result
@@ -0,0 +1,43 @@
+drop table if exists t1;
+set names sjis;
+select 'a' like 'a';
+'a' like 'a'
+1
+select 'A' like 'a';
+'A' like 'a'
+1
+select 'A' like 'a' collate sjis_bin;
+'A' like 'a' collate sjis_bin
+0
+set @sjis1= _sjis   0xa1a2a3a4a5a6a7a8a9aaabacadaeaf;
+set @sjis2= _sjis 0xb0b1b2b3b4b5b6b7b8b9babbbcbdbebf;
+set @sjis3= _sjis 0xc0c1c2c3c4c5c6c7c8c9cacbcccdcecf;
+set @sjis4= _sjis 0xd0d1d2d3d4d5d6d7d8d9dadbdcdddedf;
+set @utf81= CONVERT(@sjis1 USING utf8);
+set @utf82= CONVERT(@sjis2 USING utf8);
+set @utf83= CONVERT(@sjis3 USING utf8);
+set @utf84= CONVERT(@sjis4 USING utf8);
+select hex(@utf81);
+hex(@utf81)
+EFBDA1EFBDA2EFBDA3EFBDA4EFBDA5EFBDA6EFBDA7EFBDA8EFBDA9EFBDAAEFBDABEFBDACEFBDADEFBDAEEFBDAF
+select hex(@utf82);
+hex(@utf82)
+EFBDB0EFBDB1EFBDB2EFBDB3EFBDB4EFBDB5EFBDB6EFBDB7EFBDB8EFBDB9EFBDBAEFBDBBEFBDBCEFBDBDEFBDBEEFBDBF
+select hex(@utf83);
+hex(@utf83)
+EFBE80EFBE81EFBE82EFBE83EFBE84EFBE85EFBE86EFBE87EFBE88EFBE89EFBE8AEFBE8BEFBE8CEFBE8DEFBE8EEFBE8F
+select hex(@utf84);
+hex(@utf84)
+EFBE90EFBE91EFBE92EFBE93EFBE94EFBE95EFBE96EFBE97EFBE98EFBE99EFBE9AEFBE9BEFBE9CEFBE9DEFBE9EEFBE9F
+select hex(CONVERT(@utf81 USING sjis));
+hex(CONVERT(@utf81 USING sjis))
+A1A2A3A4A5A6A7A8A9AAABACADAEAF
+select hex(CONVERT(@utf82 USING sjis));
+hex(CONVERT(@utf82 USING sjis))
+B0B1B2B3B4B5B6B7B8B9BABBBCBDBEBF
+select hex(CONVERT(@utf83 USING sjis));
+hex(CONVERT(@utf83 USING sjis))
+C0C1C2C3C4C5C6C7C8C9CACBCCCDCECF
+select hex(CONVERT(@utf84 USING sjis));
+hex(CONVERT(@utf84 USING sjis))
+D0D1D2D3D4D5D6D7D8D9DADBDCDDDEDF
diff --git a/mysql-test/r/have_sjis.require b/mysql-test/r/have_sjis.require
new file mode 100644
index 00000000000..72ce8ec88fe
--- /dev/null
+++ b/mysql-test/r/have_sjis.require
@@ -0,0 +1,2 @@
+Collation	Charset	Id	Default	Compiled	Sortlen
+sjis_japanese_ci	sjis	13	Yes	Yes	1
diff --git a/mysql-test/t/ctype_sjis.test b/mysql-test/t/ctype_sjis.test
new file mode 100644
index 00000000000..1dd363c4910
--- /dev/null
+++ b/mysql-test/t/ctype_sjis.test
@@ -0,0 +1,34 @@
+-- source include/have_sjis.inc
+
+#
+# Tests with the sjis character set
+#
+--disable_warnings
+drop table if exists t1;
+--enable_warnings
+
+set names sjis;
+
+select 'a' like 'a';
+select 'A' like 'a';
+select 'A' like 'a' collate sjis_bin;
+
+set @sjis1= _sjis   0xa1a2a3a4a5a6a7a8a9aaabacadaeaf;
+set @sjis2= _sjis 0xb0b1b2b3b4b5b6b7b8b9babbbcbdbebf;
+set @sjis3= _sjis 0xc0c1c2c3c4c5c6c7c8c9cacbcccdcecf;
+set @sjis4= _sjis 0xd0d1d2d3d4d5d6d7d8d9dadbdcdddedf;
+
+set @utf81= CONVERT(@sjis1 USING utf8);
+set @utf82= CONVERT(@sjis2 USING utf8);
+set @utf83= CONVERT(@sjis3 USING utf8);
+set @utf84= CONVERT(@sjis4 USING utf8);
+
+select hex(@utf81);
+select hex(@utf82);
+select hex(@utf83);
+select hex(@utf84);
+
+select hex(CONVERT(@utf81 USING sjis));
+select hex(CONVERT(@utf82 USING sjis));
+select hex(CONVERT(@utf83 USING sjis));
+select hex(CONVERT(@utf84 USING sjis));
diff --git a/strings/ctype-sjis.c b/strings/ctype-sjis.c
index c337b8122fb..8ba650d1486 100644
--- a/strings/ctype-sjis.c
+++ b/strings/ctype-sjis.c
@@ -4437,6 +4437,12 @@ my_wc_mb_sjis(CHARSET_INFO *cs  __attribute__((unused)),
   if (!(code=func_uni_sjis_onechar(wc)))
     return MY_CS_ILUNI;
   
+  if (code>=0xA1 && code <= 0xDF)
+  {
+    s[0]= code;
+    return 1;
+  }
+  
   if (s+2>e)
     return MY_CS_TOOSMALL;
   
@@ -4454,12 +4460,18 @@ my_mb_wc_sjis(CHARSET_INFO *cs  __attribute__((unused)),
   if (s >= e)
     return MY_CS_TOOFEW(0);
   
-  if (hi<0x80)
+  if (hi < 0x80)
   {
     pwc[0]=hi;
     return 1;
   }
   
+  if (hi >= 0xA1 && hi <= 0xDF)
+  {
+    pwc[0]= func_sjis_uni_onechar(hi);
+    return 1;
+  }
+  
   if (s+2>e)
     return MY_CS_TOOFEW(0);