#3 Cannot read strings as UTF-8 bytes

[Thomas Phan]

This is an API that I have been looking for since I
started using C# with mysql, great :)

---

in MySqlCommand::ParseCommand, will there be a
support to use '?' for the param to be replaced? this
works in OdbcCommand

---

Will there be an UTF-8 support?

• such as the feature,
  useUnicode=true&characterEncoding=UTF-8, offered by
  MySQL Connector/J (the official JDBC driver) My
  application uses both C# and Java

• basically, it ignores the system's character set, and
  converts C# unicode strings into UTF-8 bytes, then
  push the bytes into a VARCHAR column, and vice versa

I attempted the changes in MySql.cs, and
MySqlDataReader.cs:

public static string RealEscapeString(IntPtr db, string
str) {
return RealEscapeString(db,
System.Text.Encoding.UTF8.GetBytes(str));
}

public string GetString(int i) {
TryReadField(i);
ArrayList res = new ArrayList(256);
unsafe
{
byte ptr = (byte) _currentRow[i];
while (ptr != 0) res.Add(ptr++);
}
return System.Text.Encoding.UTF8.GetString((byte[])
res.ToArray(typeof(byte)), 0, res.Count);
}

The modified GetString() works ok, but the
RealEscapeString() doesn't. then I tried

command.Parameters.Add("@name",
MySqlDbType.VarCharBinary).Value =
System.Text.Encoding.UTF8.GetBytes(<chinese
characters, &#20013;&#25991;, i.e. a 3-byte unicode>);

It gives:

MySqlSharp.Provider.MySqlException: Query failed
You have an error in your SQL syntax near ''&#37533;&#21084;' at line
1
at MySqlSharp.Provider.MySqlDataReader.NextResult
() in c:\tmp\mysqlsharp1.0be
ta1-src\src\provider\mysqldatareader.cs:line 191
at
MySqlSharp.Provider.MySqlCommand.ExecuteReader
(CommandBehavior cb) in c:\t
mp\mysqlsharp1.0beta1-
src\src\provider\mysqlcommand.cs:line 250
at Test.Test.Main(String[] args) in
c:\tmp\mysqlsharp1.0beta1-src\src\test\te
st.cs:line 52

the SQL in cmds[0] was SELECT * FROM Lookup
WHERE name='&#37533;&#21084;?', which is a correct UTF-8
VARCHAR encoding

Is that becoz the native mysql dll converts the encoded
string using the system's character set? In myOdbc, I
used

command.Parameters.Add(null,
OdbcType.VarBinary).Value =
System.Encoding.UTF8.GetBytes("&#20013;&#25991;");

and this gives the same output as the JDBC driver

Thanks

Thomas

[Thomas Phan]

Logged In: YES
user_id=409109

Looked at mysql's CAPI, we may need to use
mysql_escape_string for bytes

[DllImport("libmySQL.dll",
CharSet=System.Runtime.InteropServices.CharSet.Ansi,
EntryPoint="mysql_escape_string", ExactSpelling = true)]
private static extern int EscapeStringInternal(IntPtr to, IntPtr
from, uint length);

and the SQL error is due to the difference between
command's unicode and char* lengths.

MySql.RealQuery(_connection._mysql, _cmds
[_currentCommand], (uint)_cmds[_currentCommand].Length)
^^^
MySqlDataReader::NextResult

For bytes that may include 0 (end of string char), I suggest
that MySqlCommand::ParseCommand() to return byte[][] (or
a collection of byte[])

Thanks

Thomas