> Remy Maucherat wrote:
> >
> > > Hi all,
> > >
> > > As suggested, I've converted JDBCDescriptorsStore to using prepared
> > > statements so as to have the neccesary escaping done for us. I haven't
> > > tested it yet, though.
> >
> > Great !
> >
> > > I intend to finish off the conversion and test on friday, then commit.
> > > If anyone wants to speak against this, please do so and I'll hold off.
> >
> > I'm definitely +1 for this.
> >
>
> ok. Well, it's mostly working now, but I'm not sure it's completely
> reliable just yet. It seems to be, though.
>
> The problem I'm looking at right now is retrieveRevisionDescriptor() in
> the store being called with a null revisioNumber.
That looks like a bug ...
The Memory store (which I'm using most of the time) is robust too (it will
throw a RevNotFound). I think I'll add some debug statements to try to see
when it happens.
> I'm not sure what the
> right thing to do in this situation is. I'm going to try just checking
> for this at the start of the method, and throwing a
> RevisionDescriptorNotFoundException if it's null (this seems
> reasonable),
That's what the memory store is doing (by accident).
> but I'm not sure whether that's the right thing. I suspect that this was
> previously hidden behind java's automatic-stringification magic.
Yes it was ;)
> (update after testing: this works, but I'm still not sure if it's
> 'correct')
>
> OK, should be committed in a few minutes. Please, everybody that uses
> the current JDBCDescriptorsStore, TEST!
Remy
