Refactor Serializer to support iteratation

[Paul Egan]

The [source:django/trunk/django/core/serializers/base.py Serializer] interface forces the result to be fully materialised before use. This prevents it's use for streaming responses.

Currently the closest you can come is something like the following, but the response doesn't return until the xml document is complete:

    response = HttpResponse(content_type='text/xml')
    serializer = serializers.get_serializer("xml")()
    serializer.serialize(queryset.iterator(), stream=response)
    return response

The attached patch adds an iterator method to Serializer which yields the serialised data in chunks. This allows the serialisation of large query sets to be streamed to a response without using much memory:

    serializer = serializers.get_serializer("xml")()
    xml_iter = serializer.iterator(queryset.iterator(), chunk_size=100)
    return HttpResponse(xml_iter, content_type='text/xml')

[Paul Egan]

Patch to add iterator support to Serializer

[Eric Holscher]

Setting to DDN because it is a design decision. That said, I think this is a solid improvement and would be a nice thing to have in 1.1

[Russell Keith-Magee]

Duplicate of #5423 (although this patch like it could be more promising).