Spring Core 3.2.5

[Rod Widdowson]

Based on Ian's suggestion, what I have done is
 
1) Enter IDP338 for this case.
2) Checked in a rather grotesque fix to the code: any bean which needs to be
looked up by class can be added by reference to a no-op bean.  The parsing
code looks for the no-op bean
Before it starts looking for beans by Class.
3) Checked in a "canary test", which will fail when we upgrade to a version
of spring which passes this test.

All of this is (of course) easily reverted.

Rod

> On Behalf Of Ian Young
> My opinions:
> 
> If there is a way for us to work round the new (mis-)behaviour and move
> forward, I'd recommend we do that.  We should only revert to a previous
> version if we can't do that.
> 
> Either way we should be logging the bug in the Spring code. I don't think
> we're the best people to provide a patch, though I'd probably want to go
> down that route if they didn't correct it in their next stable release.
> 
> 	-- Ian
> 
>