This file documents the features and limitations of the ODBSequoia driver version v0.5a.
The ODBSequoia driver is designed as a thin layer above the Carob static library. Most of the work is actually performed by Carob. To simplify the whole thing is called "the driver" below.
Ability to connect to a sequoia controller, send a write query, get the update count, send a read request, browse the resultset, disconnect. You can actually do something useful with this preview. However no advanced features are implemented yet and your usual ODBC application is likely to miss some. For instance dynamic database browsers will very partially work at best because they try to access database and resultset metadata, which is not implemented yet.
Among others that means that not yet implemented features and other failures will NOT leave you in the dark wondering what happened. There are also no "half-implemented" features that could put the driver in an inconsistent state or crash. Giving a try to this driver will not eat your time; if your application has good error reporting then you will be able to know right away where this preview stands.
From your ODBC .ini configuration file you can also enable a useful log feature (using log4cxx or to a more basic stderr).
When some ODBC function has a 'W' (wide char) version, then ODBSequoia implements it using the 'W' version, so no worry for your unicode application. But this also imply that you need an Unicode Driver Manager. Check the UNICODE file for more details.
If you report a problem to the mailing-list and the fix is simple enough, it will be committed to CVS in the next days. Feature wishes reported on the mailing-list will always get a higher priority than others: we need your feedback to know which are the ODBC features that matter. And do not forget to check our JIRA task/bug tracker.
Obviously patches & other contributions are highly welcome (under the Apache 2.0 licence currently used).
Almost every line of this code has been manually tested, meaning it has actually run at least once. But the next developement step is definitely to implement some exhaustive, automated and regularly run non-regression test-suite.
Calling these functions will simply fail.
$Date: 2006/05/24 13:25:04 $ $Revision: 1.3 $