This is a package of low-to-high level lexing and parsing procedures that can be combined to yield a SAX, a DOM, a validating parser, or a parser intended for a particular document type. The procedures in the package can be used separately to tokenize or parse various pieces of XML documents. The package supports XML Namespaces, internal and external parsed entities, user-controlled handling of whitespace, and validation. This module therefore is intended to be a framework, a set of "Lego blocks" you can use to build a parser following any discipline and performing validation to any degree. As an example of the parser construction, this file includes a semi-validating SXML parser.
The present XML framework has a "sequential" feel of SAX yet a "functional style" of DOM. Like a SAX parser, the framework scans the document only once and permits incremental processing. An application that handles document elements in order can run as efficiently as possible. Unlike a SAX parser, the framework does not require an application register stateful callbacks and surrender control to the parser. Rather, it is the application that can drive the framework – calling its functions to get the current lexical or syntax element. These functions do not maintain or mutate any state save the input port. Therefore, the framework permits parsing of XML in a pure functional style, with the input port being a monad (or a linear, read-once parameter).
Besides the port, there is another monad – seed. Most of the
middle- and high-level parsers are single-threaded through the seed. The
functions of this framework do not process or affect the seed in any way:
they simply pass it around as an instance of an opaque datatype. User functions,
on the other hand, can use the seed to maintain user's state, to accumulate
parsing results, etc. A user can freely mix his own functions with those of the
framework. On the other hand, the user may wish to instantiate a high-level
parser: SSAX:make-elem-parser
or SSAX:make-parser
. In the latter
case, the user must provide functions of specific signatures, which are called
at predictable moments during the parsing: to handle character data, element
data, or processing instructions (PI). The functions are always given the
seed, among other parameters, and must return the new seed.
From a functional point of view, XML parsing is a combined pre-post-order traversal of a "tree" that is the XML document itself. This down-and-up traversal tells the user about an element when its start tag is encountered. The user is notified about the element once more, after all element's children have been handled. The process of XML parsing therefore is a fold over the raw XML document. Unlike a fold over trees defined in [1], the parser is necessarily single-threaded – obviously as elements in a text XML document are laid down sequentially. The parser therefore is a tree fold that has been transformed to accept an accumulating parameter [1,2].
Formally, the denotational semantics of the parser can be expressed as
parser:: (Start-tag -> Seed -> Seed) -> (Start-tag -> Seed -> Seed -> Seed) -> (Char-Data -> Seed -> Seed) -> XML-text-fragment -> Seed -> Seed parser fdown fup fchar "<elem attrs> content </elem>" seed = fup "<elem attrs>" seed (parser fdown fup fchar "content" (fdown "<elem attrs>" seed)) parser fdown fup fchar "char-data content" seed = parser fdown fup fchar "content" (fchar "char-data" seed) parser fdown fup fchar "elem-content content" seed = parser fdown fup fchar "content" ( parser fdown fup fchar "elem-content" seed)
Compare the last two equations with the left fold
fold-left kons elem:list seed = fold-left kons list (kons elem seed)
The real parser created by SSAX:make-parser
is slightly more complicated,
to account for processing instructions, entity references, namespaces,
processing of document type declaration, etc.
The XML standard document referred to in this module ishttp://www.w3.org/TR/1998/REC-xml-19980210.html
The present file also defines a procedure that parses the text of an XML document or of a separate element into SXML, an S-expression-based model of an XML Information Set. SXML is also an Abstract Syntax Tree of an XML document. SXML is similar but not identical to DOM; SXML is particularly suitable for Scheme-based XML/HTML authoring, SXPath queries, and tree transformations. See SXML.html for more details. SXML is a term implementation of evaluation of the XML document [3]. The other implementation is context-passing.
The present frameworks fully supports the XML Namespaces Recommendation:http://www.w3.org/TR/REC-xml-names/ Other links:
-- Scheme Procedure: pair? x Return `#t' if X is a pair; otherwise return `#f'.
-- Scheme Procedure: null? x Return `#t' iff X is the empty list, else `#f'.