axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Glen Daniels <gdani...@macromedia.com>
Subject RE: Some source
Date Tue, 24 Apr 2001 15:49:50 GMT

The reason the tests don't break is that nothing in them is using my code
yet. :)

All the classes I built were named differently than the ones the current
codebase is using, so to hook them in we'll need to change the current code.
However, they're really not ready for this yet, primarily because they don't
serialize - deserialization works fine, but there's no output code yet
(making two-way communication a bit challenging :)).

If people are psyched that this is the way to go, I'll take care of adding
this, renaming the classes, and hooking it in to the actual framework.  In
the meanwhile, I'd certainly encourage you to build on the framework and
flesh out the deserialization - but hold off on checking it in for a little
bit.

One of the other big things I need to add is multiref handling, which I know
how to do but haven't implemented yet.

--Glen


> -----Original Message-----
> From: Sam Ruby [mailto:rubys@us.ibm.com]
> Sent: Tuesday, April 24, 2001 10:46 AM
> To: axis-dev@xml.apache.org
> Subject: Re: Some source
> 
> 
> Glen,
> 
> It seems to compile cleanly and doesn't break tests (I'll 
> admit that I've
> only tried the stock quote, but that one seemed to work).
> 
> My focus for the near term is to flush out the 
> (de)serializers - at least
> to the point where the N+I demo can be implemented without 
> unnatural acts.
> I see that the support is incomplete but adequately roughed in.
> 
> My take is that this work should be committed so that we can 
> complete and
> build upon it.  Particularly as it appears to be the most promising
> technical approach that we know of at this time for achieving 
> performance
> and scalability.
> 
> - Sam Ruby
> 

Mime
View raw message