Re: [Simple-support] Support for almost-primitive types, Converters in attributes and Arrays in attributes

[Raphael J. <rap...@gm...>]

Hi,

It seems that it does not work for attributes.

Attributes require to parse / generate primitive types (and not array of
it), whereas there should not be  any fondamental difference from parsing /
seriliazing the "text" value of a single element.

Raphael

2011/5/20 Niall Gallagher <gal...@ya...>

> Hi,
>
> These capabilities are all ready there, take a look at the
> Matcher/Transform for transforming custom types and attributes. Also,
> @Elementon an array produces a comma separated list, all read and written
> via the Transform objects resolve to the types.
>
> Niall
>
> --- On *Thu, 19/5/11, Raphael Jolivet <rap...@gm...>* wrote:
>
>
> From: Raphael Jolivet <rap...@gm...>
> Subject: [Simple-support] Support for almost-primitive types, Converters in
> attributes and Arrays in attributes
> To: sim...@li...
> Received: Thursday, 19 May, 2011, 6:10 AM
>
>
> Hello there,
>
> I am using the Simple XML framework extensively for all my XML parsing /
> generating tasks and I am very happy with it.
> However, I would like to propose some enhancements that would enable me to
> serialize / deserialize almost all my classes without the need of further
> work, except for some annotations.
>
> = Almost-primitive types
>
> First, I would like to ask for the support of "almost-primitive" types.
>
> In the Jersey REST Servlet (based on JAX-RS API), there is a convention
> that states : Any class providing a "public Constructor(String)" or a static
> method "static public fromString(String)" can be considered as a native type
> and is parsed, serialized without having to write a custom Converter /
> Mapper. The toString() method is using for serialization.
>
> This is very handy, as writing a single static method for a class is less
> work than writing a Converter and having to register it, or to annotate
> every occurrence of attributes of its type.
>
> = Only primitive types supported for attributes ?
>
> Also, I have experienced that custom converters (either registered in a
> strategy, or specified as annotations with the AnnotationStrategy) are not
> supported for attributes.
>
> It seems that attributes explicitly require the use of Java primitive
> types.
> Therefore, I had to write custom getters / setters with Strings to overcome
> this limitation.
>
> = Unique annotation to parse Lists and Arrays, and new annotation to parse
> arrays of attributes (with a separator)
>
> I think that the distinction between ElementArray and ElementList is not
> necessary, as it is semantically the same structure of objects ( a sequence
> of objects) : The framework should handle transparently Arrays and List
> without having to specify it in the annotation. let's have a unique
> "ElementSequence" annotation.
>
> Also, this is a very common situation to have several values wrapped into a
> unique attributes, as comma separated values.
> <task times="2 s, 3 h, 4m" >
>
> I would like to be able to write something like
>
> AttributeSequence(name="times", separator=",")
> Duration[] durations;
>
> Duration being a class that provides a static "fromString(String)" method,
> to parse Strings.
> With that, the framework should be able to serialize / parse this code
> without the need of any other information / code.
>
> What do you think about those requests ?
> Some of them may be already supported : Please tell me if I missed
> something.
>
> Thanks again for your great work,
> Best regards,
>
> Raphael JOLIVET
>
>
>
> -----Inline Attachment Follows-----
>
>
> ------------------------------------------------------------------------------
> What Every C/C++ and Fortran developer Should Know!
> Read this article and learn how Intel has extended the reach of its
> next-generation tools to help Windows* and Linux* C/C++ and Fortran
> developers boost performance applications - including clusters.
> http://p.sf.net/sfu/intel-dev2devmay
>
> -----Inline Attachment Follows-----
>
> _______________________________________________
> Simple-support mailing list
> Sim...@li...<http://mc/compose?to=...@li...>
> https://lists.sourceforge.net/lists/listinfo/simple-support
>
>