users@jpa-spec.java.net

[jpa-spec users] [jsr338-experts] Re: converters

From: Bernd Müller <bernd.mueller_at_ostfalia.de>
Date: Wed, 08 Feb 2012 19:08:17 +0100

Am 08.02.2012 18:13, schrieb michael keith:
> On second thought, we might need to be a little more discriminating
> about autoApply and schema generation.
> Perhaps the autoApply would need to be defined in XML if the converter
> does type changing? Seems unkind to require that the provider rifle
> through all of the classes looking for every converter before it can
> process any entity. What do others think?
>
JSF has to scan an archive too,because of the @FacesConverter
annotation.
CDI has to scan an archive too, because all beans are injectable
anywhere.
But: this has to be done only one time, at application start.
I think this is justifiable.
Probably, it should be done only for archives with a persistence.xml,
like beans.xml for CDI.
Otherwise additional jars can break an application. Should this
be added to the spec?

regards

Bernd