-
09-25-2007 4:11 PM
|
|
-
Robert Simpson
  - Joined on 08-27-2005
- Posts 1,701
|
- Breaking change in GetSchema("Indexes") -- MetaDataCollections restrictions and identifier parts counts were wrong for this schema and I was using the wrong final parameter as the final restriction. Meaning, if you use the Indexes schema and are querying for a specific index the array should now be {catalog, null, table, index } instead of {catalog, null, table, null, index}
- Fixed some errors in the encryption module, most notably when a non-default page size is specified in the connection string.
- Fixed SQLiteDataReader to better handle type-less usage scenarios, which also fixes problems with null values and datetimes.
- Fixed the leftover temp files problem on WinCE
- Added connection pooling. The default is disabled for now, but may change in the future. Set "Pooling=True" in the connection string to enable it.
- Sped up SQLiteConnection.Open() considerably.
- Added some more robust cleanup code regarding SQLiteFunctions.
- Minor additions to the code to allow for future LINQ integration into the main codebase.
- Fixed a long-standing bug in the Open() command of SQLiteConnection which failed to honor the documented default behavior of the SQLite.NET provider to open the database in "Synchronous=Normal" mode. The default was "Full".
- If Open() fails, it no longer sets the connection state to Broken. It instead reverts back to Closed, and cleans up after itself.
- Added several new parameters to the ConnectionString for setting max page count, legacy file format, and another called FailIfMissing to raise an error rather than create the database file automatically if it does not already exist.
- Fixed some designer toolbox references to the wrong version of the SQLite.Designer
- Fixed a bug in the mergebin utility with regards to COR20 metadata rowsize computations.
- Minor documentation corrections
|
|
Page 1 of 1 (1 items)