There is a basic recusive descent parser for Pages format. The task is to rewrite this to use a single parser loop and a "context" class for every element. This involves:

* Deriving the parser (source is src/lib/PAGParser.*) from IWORKParser.
* Defining Pages-specific dictionary and parser state--see KEY2ParserState.h and KEYDictionary.h. They will not have any members initially, but they will likely be needed later.
* Defining helper base classes for contexts--see KEY2XMLContextBase.h.
* Replacing every parseFoo function by a context class. These define IWORKXMLContext API and each one of them should be derived from one of the helper base classes from previous point.

Note that sf:text-body element is already handled by IWORKTextBodyContext, so the text-related parser functions do not have to be converted, but just dropped.