Re: [cp-patches] RFC: preview of XMLEncoder

[Mark Wielaard]

Hi Robert,

[sorry for the double message, I still cannot sent mail from my primary
address to the list]

On Tue, 2005-11-22 at 15:56 +0100, Robert Schuster wrote:
> as requested on IRC I put the patch that implements parts of the XMLEncoder on
> the list. This code compiles, handles the easier XML serialization situations
> but breaks on the more complicated ones (arrays, GUI components).

Cool! Very good to post this since now people know what the status is of
your work.

> Furthermore I am not really content with some parts of the encoder's 
> architecture.
> Please tell me if you want this in anyway.

There are some clear parts that should go in:

>        * java/beans/Statement.java: Doc fixes.
>         (doExecute): Workaround for Class.forName call.
>         (toString): Made output look more like on the JDK.
>         * java/beans/Expression.java: Doc fixes.
>         (toString): Made output look more like on the JDK.

These look like clear bug fixes, documentation updates. Please commit!

>         * java/beans/PersistenceDelegate.java,
>         java/beans/DefaultPersistenceDelegate.java,
>         java/beans/Encoder.java,
>         java/beans/XMLEncoder.java: New file.
>         * gnu/java/beans/encoder/Action.java,
>         gnu/java/beans/encoder/ActionIssuer.java,
>         gnu/java/beans/encoder/ArrayInstantiationScannerState.java,
>         gnu/java/beans/encoder/ArraySetScannerState.java,
>         gnu/java/beans/encoder/ArrayPersistenceDelegate.java,
>         gnu/java/beans/encoder/ClassPersistenceDelegate.java,
>         gnu/java/beans/encoder/GenericScannerState.java,
>         gnu/java/beans/encoder/IgnoringScannerState.java,
>         gnu/java/beans/encoder/PrimitivePersistenceDelegate.java,
>         gnu/java/beans/encoder/ReportingScannerState.java,
>         gnu/java/beans/encoder/ScanEngine.java,
>         gnu/java/beans/encoder/ScannerState.java,
>         gnu/java/beans/encoder/StAXActionIssuer.java: New file.

These are very sparse on documentation. Making it hard to follow. If you
could document the design a bit, plus some some notes on how you think
the architecture could be improved then it might be a good addition if
someone wants to work on improving it.

Dare I ask about test cases?

Cheers,

Mark