Permalink
Commits on Nov 3, 2011
Commits on Oct 27, 2011
  1. Issue 170: Fixed memory overwrite when converting Unicode to SQLWCHAR.

    Many thanks to davidp.r...@gmail.com for finding this.
    Michael Kleehammer committed Oct 27, 2011
Commits on Oct 17, 2011
  1. fix: Function sequence error with prepared execute, SQLTables, same e…

    …xecute.
    
    The saved prepared statement was not cleaned up when a SQLTables or similar function was used.
    If the next execute used the same SQL, pyodbc did not call SQLPrepare.
    Michael Kleehammer committed Oct 17, 2011
Commits on Sep 14, 2011
  1. Issue 145: Update to handle failure.

    The parameter array is initialized to zero in case of failure during the binding process.
    Therefore it is possible that some values are still zero in FreeInfos.
    committed Sep 14, 2011
Commits on Sep 13, 2011
  1. Issue 145: cursor.execute leaks parameters

    Thanks to luke (see issue) for finding, debugging, and providing a fix.
    Michael Kleehammer committed Sep 13, 2011
Commits on Aug 21, 2011
  1. Issue 199: with cxn: doesn't commit on success

    (cherry picked from commit a6b400f)
    committed Aug 21, 2011
Commits on Jul 16, 2011
  1. Issue 188: Segfault when fetching NVARCHAR(MAX) data over 511 bytes

    The provided fix suggestion looks correct and I can see how it would crash, but was unable to
    reproduce with the unit test.  Probably related to Microsoft vs. FreeTDS differences.
    committed Jul 16, 2011
Commits on Jul 11, 2011
  1. LONGVARCHAR/LONGWVARCHAR fix

    Was passing the incorrect length when binding, causing SQLPut errors when using long strings
    (usually when inserting long strings) for some drivers.
    committed Jul 11, 2011
Commits on Jul 8, 2011
  1. Cleaned up compiler warnings

    committed May 20, 2011
Commits on Jul 4, 2011
  1. Issue 186: Added SQLite tests

    The issue was actually in the ODBC driver, which has been fixed in version
    0.91 of the driver.
    
    There are a couple of issues left, but I'll need to work through them with the driver author:
    reading a zero-byte blob and reading text values in 2 passes.
    committed Jul 4, 2011
Commits on Nov 25, 2010
  1. Fixed printf in new pyodbcconf.

    Eliminated unused code.
    Eliminated compiler warning.
    committed Nov 25, 2010
Commits on Nov 24, 2010
  1. Added pyodbc.drivers() for Windows builds.

    In the future, I will try to add the method to other platforms by reading the iODBC and
    unixODBC configuration files.
    committed Nov 24, 2010
Commits on Nov 23, 2010
Commits on Nov 21, 2010
  1. Issue 121: update with large parameter + no rows updated = error

    Issue and patch (!) from luke@deller.id.au.  Thanks.
    committed Nov 21, 2010
  2. Issue 129: ("select ?", None) fails

    If SQLDescribeParam cannot tell us the type of a column for a NULL parameter, we'll assume it
    is VARCHAR.  We already do this if SQLDescribeParam is not supported.
    committed Nov 21, 2010
Commits on Sep 6, 2010
  1. Allocated own string for decimals.

    Was using str(decimal), but it can return items in exponential notation, which I'm not sure
    many databases would handle properly.
    committed Sep 6, 2010
  2. More Unicode fixes. Updated mysql tests.

    Michael Kleehammer committed Sep 6, 2010
Commits on Sep 5, 2010
  1. Issue 91: decimal fix; Completely reworked parameters; added leakcheck

    Fixed issue 91 - handling of decimals was incorrect (off by 1).  Added lots of SQL Server unit
    tests.
    
    To properly fix this, reworked and simplified parameter binding.  Instead of separating the
    memory requirement and the actual memory preparation, it is now performed in 1 step and stored
    in the new ParamInfo structures.
    
    I added a lot of optional malloc calls (not usually used), so I added a PYODBC_LEAK_CHECK
    parameter which causes pyodbc_malloc and pyodbc_free to track allocations.  This is not thread
    safe and is intended only for troubleshooting.
    committed Sep 5, 2010
Commits on Sep 4, 2010
  1. Issue 86: getinfo reading into wrong data type.

    Was using PyArg_ParseTuple to read a long into a SQLUSMALLINT, which is completely wrong.
    Changed dadta type to unsigned long then cast as necessary later.
    Michael Kleehammer committed with Aug 30, 2010
  2. Issue 83/90: Fixed manifest so sdist works on non-Windows.

    Michael Kleehammer committed with Aug 30, 2010
  3. Added more decimal tests (which fail right now).

    Michael Kleehammer committed with Aug 30, 2010
  4. Implemented Row comparisons.

    Michael Kleehammer committed with Aug 30, 2010
  5. Issue 101: Added connect timeout keyword.

    Michael Kleehammer committed with Aug 30, 2010