You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@turbine.apache.org by Rodney Schneider <ro...@actf.com.au> on 2002/12/10 04:42:56 UTC

intake and turbine 2 issues (was Re: TDK 2.2 glitches)

On Tue, 10 Dec 2002 13:24, you wrote:

> if anything is set up at all wrong you get a null pointer exception;

Would it be possible for you to either supply a patch or at least log 
an issue in Scarab for this problem?

> the property names you specify in your inteke.xml get capitalised
> (i.e. if you have <field name="someName" ... you have to refer to it
> as SomeName or you get exceptions);

I have always used <field name="SomeName"> (note the capitals).  This 
should be mentioned in the Intake Howto.

> IntakeTool.get() is declared 'throws Exception' so you have to wrap
> calls to it, although it only throws TurbineException -- and when it
> does so, e.g. due to a typo in a template, the log just tells you
> that a TurbineException was thrown, without the exception message, so
> you have no idea why... And that's just what I ran into in one
> evening! :-)

The exception and error handling in Turbine leaves much to be desired.  
Nearly everything 'throws Exception'.  I don't think the Turbine 
developers had much experience with exception handling when they wrote 
the code.  Torque has recently been changed to throw TorqueException. 
The parts of Turbine that currently throw Exception should either throw 
TurbineException or a subclass of it.  I have attached a patch to 
TurbineException for you to retrieve its nested exception so you can 
log a decent error message.

Now that Turbine 2.2 has been released, expect the Turbine 2.x codebase 
to be majorly overhauled and cleaned up.  You are more than welcome to 
help with this effort :)

> I like what Intake provides, I just think it could be a lot easier to
> setup and use. That exception from IntakeTool.get() is a real killer.
>
> > Also, someone needs to backport the fixes that have been made to
> > the Fulcrum version of Intake to Turbine 2.x.  I have been meaning
> > to do this for a while, but haven't found the time.
>
> Wouldn't it make more sense for Turbine just to switch to the Fulcrum
> version, if that's being actively developed? I thought Turbine was
> migrating to decoupled services in general. Or was that just for
> Torque?

Eventually the services will be decoupled from Turbine, but how this is 
going to be done has not yet been decided.  Torque has already been 
completely decoupled.

Regards,

-- Rodney