commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jon Tirsén <>
Subject RE: runtime attributes
Date Wed, 20 Nov 2002 08:18:40 GMT

> -----Original Message-----
> From: Ara Abrahamian [] 
> Sent: den 19 november 2002 21:45
> To: 'Jon Tirsén';
> Cc:;; 
> 'Jakarta Commons Developers List'
> Subject: RE: runtime attributes
> Well then let's see what happens if someone serializes a 
> ClassMetaData with Nanning, and reads it with an 
> Xdoclet-based runtime api. Assume Nanning uses a different 
> syntax than xdoclet and we just store the raw text. Now the 
> Xdoclet runtime reads it and can't understand it! It can 
> happen quite easily, maybe a third party uses Nanning and not 
> Xdoclet like you do? So the lowest denominator is 
> getProperty/setProperty and no raw text, no syntax, just the api.

Yeah, I've been thinking about this too. This is one of the main reasons
why commons-attributes should be quite high-level. I'm leaning more
towards Aras way at the moment, if we just dump the serialization.

> So we can have different parsers, different syntaxs, but as 
> far as the api is concerned there's no difference for the 
> user of the api.

Yeah, the parsers and the runtime-plugin can do whatever they like if
they implement the API. So an example: A library like OJB mandates the
following: I need to have an attribute of type
org.apache.jakarta.ojb.FieldMapping on each field and an attribute of
type org.apache.jakarta.ojb.ClassMapping on each class. I don't care how
you do it, and what parser you're using to do it, I just need to access
those attributes through the commons-attribute-API. This is how you do
it using the default-parser: (example code).

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message