On 21-06-2008 19:47:23 +0200, Stefan Manegold wrote:
> > 3: since mclient does some rendering of the mapi results, jdbcclient
> > has been obsoleted. However, it still is pretty useful for testing
> > JDBC, hence it still exists, but isn't built/distributed by default.
>
> Thanks!
>
> Isn't there also a version of jdbcclient in monetdb-X.Y-jdbc.jar?
> If so, shouldn't we consider modifying the respective tests to use that one,
> and hence completely remove both jdbcclient.jar & mjclient?
> ("Officially", neither is distributed/released ...)
No. The way it is built is as follows:
- mcl (the classes that deal with the wire protocol)
- jdbc (the classes that implement the JDBC interfaces and their logic)
- jdbcclient (the client class that emulates an interactive console)
monetdb-X.Y-jdbc.jar: mcl + jdbc
jdbcclient.jar: mcl + jdbc + jdbcclient
Hence, you can *always* use jdbcclient.jar for everything, since it just
contains everything (except xrpc stuff, which I leave out of
consideration here).
> Otherwise, would it be an option to replace mjclient/jdbcclient.jar by
> mclient in the respetive tests?
No, because that wouldn't test the jdbc + mcl components. Some tests,
however, are a mix of testing jdbcclient and jdbc features, like the
dump/restore test, e.g. reconstructing indexes and constraints with
multi-columns. Tests that only address jdbcclient (like the encoding
tests, even though one could argue that tests jdbc as well) could be
disabled.
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Monetdb-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-developers
