The ODBC defines SQL_GUID datatype in sqlext.h. It is used
e.g. for transfer of UUID datatype from PostgreSQL.
Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
This prevents valgrind reports like:
==18426== Conditional jump or move depends on uninitialised value(s)
==18426== at 0x6423EB5: Poco::Data::StatementImpl::execute(bool const&) (in /usr/lib/libPocoDatad.so.46)
==18426== by 0x641DB5F: Poco::Data::Statement::execute(bool) (in /usr/lib/libPocoDatad.so.46)
==18426== by 0x632A0C: Poco::Data::Keywords::now(Poco::Data::Statement&) (Statement.h:443)
==18426== by 0x641C8E5: Poco::Data::Statement::operator,(void (*)(Poco::Data::Statement&)) (in /usr/lib/libPocoDatad.so.46)
...
==18426== Uninitialised value was created by a stack allocation
==18426== at 0x6A1A170: Poco::Data::ODBC::ODBCStatementImpl::affectedRowCount() const (in /usr/lib/libPocoDataODBCd.so.46)
Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
This fixes a bug where Poco would return a string 1 character shorter than is actually in the DB (when the string in the DB uses the full length of a CHAR(N) or VARCHAR(N) column).
The SQLGetConnectAttr() function can write up to 64-bit integers, as documented by MSDN here: http://msdn.microsoft.com/en-us/library/ms710297%28v=vs.85%29.aspx
This fix uses 64-bit integers instead of 32-bit ones, the original code corrupted the stack when used on win7-x64 with MSSQL 2008.
- removed naked pointers from Data interfaces
- fixed GH #82: name conflict in Data::Keywords::bind
- fixed GH #157: MySQL: cannot bind to 'long' data type on
Windows/Visual C++
- fixed GH #158: MySQL: MYSQL_BIND 'is_unsigned' member is not set
http://support.microsoft.com/kb/294169
added Buffer::size/capacityBytes returning length of buffer in bytes
only windows Unicode ODBC tested
reverting the SF #506 fix (non-conforming driver, causes problems with
other ODBC drivers)
- session handle property for SQLite and ODBC
- SQLite database utility functions and test:
(1) loading from disk to memory
(2) saving from memory to disk
