SuperBlob

A SuperBlob is basically just a redirection mechanism. It just contains the names of the products that we wish to include in the generated XML.:

 {
    "name": "fake",
    "schema_version": 4000,
    "products": [
        "c",
        "d"
    ]
}

This superblob is called fake. It just asks to look at the products named c and d. A rule can be set to point at this SuperBlob if we wish the response product to have all the files listed in products c and d.

How does this work?

The web interface, while processing requests, checks if the rule evaluates to a SuperBlob. If it is, it gets the product names from the SuperBlob and gets the corresponding blobs by evaluating the query with the requested product changed. So, in our example, the web interface will evaluate the query with the product name changed to c and d and obtain the resultant blobs. It will pick up the header and footer XML from the output of processing the blob obtained from the first product and the inner XML will include the concatenated inner XMLs of all the blobs obtained from the products listed in the SuperBlob.

So, if product a gave:

and product b gave:

the SuperBlob will give:

So, we can now have one fixed rule for all GMP responses and have several rules for each constituent plugin without having to worry about the combinations like we had to do earlier.

While working on this, I also removed the createXML method in favour of three methods that return the header, the inner XML and the footer respectively. This helped in seperating the various components of the XML output without having to parse it. The XML generation logic has moved to the client view.