create mssql.mxodbc dialect

[Former user]

Create an mssql dialect which uses the mxodbc connector, which will require getting the existing (untested, not being used) connector.mxodbc module to be revamped.

Work on this started at the PyCon 2010 SQLAlchemy sprint; basic functionality is in place for creating connections, handling inserts (using pyodbc ScopeIdentity approach), but many tests are still failing and one of the tests causes a deadlock (test_dialect.test_mssql.TypesTest.test_autoincrement).

Base tests are passing 100% Engine tests have 15 errors and 8 failures SQL tests have 26 errors and 3 failures Dialect tests have 2 errors, 2 failures, and one SkipTest (to prevent deadlock)

Total tests using mssql+mxodbc

Ran 2158 tests in 175.665s

FAILED (SKIP=1, errors=198, failures=14)

[Mike Bayer]

correct patch format

[Mike Bayer]

First version of this is in 46fa536ba4c2542bcac41346bb113a08ebcf342c. Fixed a few bugs and it is passing a good deal of tests (but haven't yet done a full run). It likely is on par with the current state of the pyodbc driver.

[Mike Bayer]

this dialect is in and has decent test support now. Still needs some work with unicode data, those issues can be separate tickets if needed.

[Mike Bayer]

Removing milestone: 0.6.0 (automated comment)