The Wsdl parsers

The key of theses parsers is to do the minimum treatments meaning that we must load the minimum amount of nodes in order to speed it up and to low the memory usage.

Neverthless, the goals of these parsers are various:

• Get the maximum amount of informations about each structs and operations
• Consolidate the informations parsed by the SoapClient parsers with informations they can't see as: - SoapHeaders (header tags) - Enumerations values - Restrictions on parameters - Default parameter value - Abstract elements - Input/Output parameters type - Values of type array - Inheritance between elements

Knowing this, it is simpler to understand why simpleType are not parsed as parsing them would mean that:

• We would retrieve each tag
• For each tag we would apply various methods to test the presence of each possible information we want to get (possibly none) and each possibility
• For each child tag, parse its informations and its own children recursively

This shows that potentially we would load lots of nodes for nothing if they don't contain anything interesting. We simply do the opposite by:

• Retrieve each tag that provide additional usefull informations
• For each retrieved tag, climb to its parent
• Parse the tag and consolite its parent's informations with it

So, if we load all the documentation nodes that contain textual informations about its container, if documentations are numerous it's good because it means that the Web Service is well documented.

On the other hand, if there is no documentation node, then we won't do anything meaning that we won't loose time to parse any node.

After all!

After all, if you think it is not a good choice to not parse simpleType and so on, please let me know the reasons.