wiki:MicroXmlCowan

Version 6 (modified by cowan, 6 weeks ago) (diff)

--

Events

A MicroXML event is either an end-of-file object or a list representing a parsing event, in one of the following formats. Stack is a list of element names currently being processed; the first element of the list is the current element.

($start stack attr-list)

Represents a start-tag. Attr-list is a JSO representing the attributes.

($end stack)

Represents an end-tag.

($text text)

Represents character content. Text is the character content as a string.

($error stack error-code . other)

Represents a parsing error. Error-code is a symbol. Other is implementation-dependent.

($pi stack target content)

Represents a processing instruction (which is not part of MicroXML). Target is a symbol; content is a string. Parsers may return $error objects instead of $pi objects; they are provided as a way of allowing a MicroXML parser to read some (but not all) XML documents. Note that any XML declaration is taken to be a processing instruction, as in SGML.

SXML

MicroXML uses a simplified version of SXML as the internal representation of documents. Each SXML element is a list whose first member is a symbol representing the element name, whose second member is a JSO mapping the attribute names (as symbols) to their values (as strings), and whose remaining members (if any) are either SXML elements or strings. The prototype of a JSO representing an attribute list is ignored. There is no representation of comments or processing instructions in this version of SXML, and no notion of document objects (a document is just an element that has no parent). The parent of an element that has one is stored as an attribute named $parent whose value is a SRFI 111 box containing the parent element. This pseudo-attribute is ignored by all other attribute processing. The box is used to prevent SXML elements from being circular tree structure.

Parsing, building, and printing procedures

Options is a list of symbols that control how MicroXML is written. All implementations recognize apos to wrap attribute values in apostrophes, end-tags to write end-tags for empty elements, and ascii to escape all characters outside the ASCII range. Other symbols are also allowed.

(uxml->sxml port [ handler ])

Reads all the characters from the textual input port port as a MicroXML document and returns the SXML equivalent. The procedure handler is invoked when a $error or $pi event is produced. The default handler signals an error that satisfies uxml-error?. If the handler is #f, errors and processing instructions are ignored; this relaxed parsing mode allows some XML documents that are not well-formed MicroXML to be parsed.

(sxml->uxml element port options)

Writes the SXML element in MicroXML format to the textual output port port, using the symbols in options.

(make-uxml-generator port)

Returns a generator of event objects representing a MicroXML document read from the textual input port port. Processing continues no matter how many errors there are until all characters have been read.

(make-sxml-generator element)

Returns a generator of event objects representing the SXML element.

(event-generator->uxml gen port options)

Invokes the generator gen to obtain event objects and writes the corresponding MicroXML document to the textual output port port, using the symbols in options. If the resulting document would not be well-formed MicroXML, an error is signaled that satisfies uxml-error?.

(event-generator->sxml gen)

Invokes the generator gen to obtain event objects, constructs the corresponding SXML element, and returns it. If the resulting object would not be well-formed SXML, an error is signaled that satisfies uxml-error?.

(write-uxml port options)

Returns a procedure that accepts an event object. When invoked repeatedly, it writes the corresponding MicroXML representation to the textual output port port using the symbols in options. If the resulting document would not be well-formed MicroXML, an error is signaled that satisfies uxml-error?.

(build-sxml)

Returns a builder procedure that accepts an event object or an end of file object. When invoked repeatedly, it builds the corresponding SXML representation. If the object is an end of file object, the procedure returns the SXML element; if not, it returns an unspecified value. If the resulting document would not be well-formed MicroXML, an error is signaled that satisfies uxml-error?.

Predicates

(sxml-element? obj)

Returns #t if obj is an SXML element and #f otherwise. The procedure checks that obj is a list whose first element is a symbol and whose second element is a JSO; further elements of the list are not examined.

(sxml-empty? element)

Returns #t if element is an empty SXML element and #f otherwise.

(sxml-wf-element-name? string attribute)

Returns #t if string matches the MicroXML name production; returns #f otherwise.

(sxml-wf-attribute-name? string attribute)

Returns #t if string matches the MicroXML attribute name production; returns #f otherwise.

(sxml-wf-string? string)

Returns #t if all the characters in string are allowed in MicroXML character content and #f otherwise.

(sxml-wf-element? element)

Returns #t if element is well-formed. The first element of element must be a symbol whose print name satisfies sxml-wf-element-name?. The second element of element must be a JSO that maps symbols whose print name satisfies sxml-wf-attribute-name? to strings that satisfy sxml-wf-string?. The remaining elements must be either strings that satisfy sxml-wf-string? or lists that satisfy sxml-wf-element?.

(sxml-attribute? element attr-name)

Returns #t if attr-name (a symbol) is an attribute of element and #f otherwise.

(sxml-id-valid? element id-mapping idref-list)

Returns #t if all idref attributes contain valid ids. An id is valid if it appears as a key in id-mapping (see make-id-mapping). Idref-list is a list of 2-element sublists, where the first element of each sublist is an element name and the second element is an attribute name.

(sxml-language? element language)

Returns #t if the language of element, as specified by the value of a lang or xml:lang attribute (the latter is not well-formed MicroXML but is supported for backward compatibility) matches language. If element has no such attribute, the language of the nearest ancestor of element that has such an attribute is used. If there is no such attribute at all, then sxml-language? returns #f.

The attribute value matches language if, in a case-insensitive comparison, language exactly equals the attribute value, or if language exactly equals a prefix of the attribute value such that the first character following the prefix is "-".

Element procedures

(sxml-name element)

Returns the name of element as a symbol.

(sxml-set-name element name)

(sxml-set-name! element name)

Changes the name of element to the symbol name, purely or with mutation.

(sxml-attr-list element)

Returns the attribute list of element as a JSO.

(sxml-set-attr-list element jso)

(sxml-set-attr-list! element jso)

Changes the attribute list of element to jso, purely or with mutation.

(sxml-contentelement)

Returns the name of element as a list.

(sxml-set-content element list)

(sxml-set-content! element list)

Changes the content of element to the list, purely or with mutation.

(sxml-value element)

Returns the results of concatenating all string content children in element and all its descendants in depth-first left-to-right preorder.

(sxml-defaults element attribute-defaults element-defaults inherited-attributes)

(sxml-defaults! element attribute-defaults element-defaults inherited-attributes)

Returns element with default values expanded in itself and all its descendants (purely or with mutation) as follows:

  • attribute-defaults is a list of 3-element sublists. Each sublist contains an element name (a symbol), an attribute name (a symbol), and a default value (a string). All elements with those names are checked for the presence of the corresponding attribute. If it is missing, the attribute is added with the specified default value.
  • element-defaults is a list of 2-element sublists. Each sublist contains an element name (a symbol) and a default value (a string). All empty elements with any of those names have the default value installed as the only content child.
  • inherited-attributes is a list of symbols. All elements are checked for an attribute whose name is one of the list. If absent, then the most recent ancestor of the element that has this attribute is found (note that no parent map is required), and the attribute is added to the element being processed with the same value as in the ancestor.

(sxml-element position element parent-map)

Returns the position of element among the element children of the parent of element as an exact integer, with 1 meaning the first element child. If there is no parent, return 0.

(sxml-element-size element)

Return the number of content children of element as an exact integer.

(sxml-normalize-element! element)

Returns a normalized version (using mutation) of an SXML element that does not necessarily conform to the definition. In particular, at least the following repairs are made:

  • If the name is a string, it is converted to a symbol with string-symbol.
  • If the attribute-list is missing, an empty JSO is provided.
  • If the attribute-list does not begin with an @ element, one is provided.
  • If one of the content children or an attribute value is a number, it is converted to a string with number->string.
  • If one of the content children or an attribute value is a boolean, it is converted to a string with boolean->string.
  • If one of the content children or an attribute value is a symbol, it is converted to a string with symbol->string.
  • If one of the content children is some other type of Scheme object, it is converted to a string by some implementation-defined means or else removed.
  • If an attribute value is some other type of Scheme object, it is converted to a string by some implementation-defined means or else that attribute is removed.
  • If after the above transformations are completed, two or more consecutive content children are strings, they are consolidated.
  • If any content children are elements, they are recursively normalized.

(sxml-trace element traceproc)

Displays information on (current-error-port) about element. The precise nature of the information displayed is undefined, except that it should end with a newline. Element is returned.

(sxml-root element)

Returns the root element of element.

(sxml-parent element)

Uses the $parent pseudo-attribute to determine the parent of element and returns it, or #f if there is none.

(sxml-detach-parent! element parent-mapping)

Removes the mapping from element to its parent from parent-mapping. If element does not have a parent, nothing is done. Returns an unspecified value.

(sxml-make-id-mapping document)

Creates an id mapping based on the SXML element document. An id mapping is an opaque object that maps an id (a symbol) to an element. The element and all its descendants are checked for the presence of an attribute named id or xml:id (the latter is not well-formed MicroXML but is allowed in SXML for backward compatibility). If found, an entry is created in the id mapping that maps the corresponding attribute value as a symbol to the element. Returns the id mapping.

(sxml-id id id-mapping)

Looks up the symbol id in id-mapping and returns the corresponding element, or #f if there is none.

String procedures

(uxml-escape-string string attribute? apos? ascii?)

Converts string to contain the necessary entity references for MicroXML. In all cases, the characters < & > are escaped with entity references. If attribute? is true, then if apos? is true, ' is escaped, but if apos? is false, then " is escaped, in both cases with an entity reference. Finally, if ascii? is true, non-ASCII characters are escaped with numeric character references. All other characters are left unchanged. The escaped result is returned.

(uxml-unescape-string string)

Converts string by translating all MicroXML escapes, both entity references and numeric character references, to single characters. All other characters are left unchanged. The result is returned.

(uxml-normalize-space string)

Returns a string that is equal to string, but with all leading and trailing whitespace removed, and all other consecutive whitespace characters replaced by a single space.

Boolean conversions

These use the conventions of XPath and XML Schema.

(sxml-string->boolean string)

Converts the strings "1" and "true" to #t, and the strings "0" and false to #f. If any other string is passed, an error is signaled that satisfies uxml-error?.

(sxml-boolean->string boolean)

Converts #t to "true" and #f to "false".

(sxml-number->boolean number)

If number returns #t when zero? is applied to it, returns #f; otherwise returns #t;

(sxml-boolean->number boolean)

If boolean is true, returns 1, otherwise returns 0.

Axis procedures

The following procedures are generator operations: they accept a generator (of SXML elements) and return a generator (also of SXML elements). After the sxml- prefix, they begin with g, using the convention of SRFI 121 for generator operations.

(sxml-gparent gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their parent elements on successive invocations.

(sxml-gancestor gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their ancestor elements from parent to root on successive invocations.

(sxml-gancestor-or-self gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their element itself and then its ancestor elements from parent to root on successive invocations.

(sxml-gchild gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their descendant elements in depth-first order from left to right on successive invocations.

(sxml-gdescendant gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their child elements from left to right on successive invocations.

(sxml-gdescendant-or-self gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns their element itself and then its child elements from left to right on successive invocations.

(sxml-gfollowing gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns all of their following elements in document order on successive invocations.

(sxml-gfollowing-or-self gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns the elements themselves and then all of their following elements in document order on successive invocations.

(sxml-gpreceding gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns all of their preceding elements in reverse document order on successive invocations.

(sxml-gpreceding-or-self gen)

Returns a generator of SXML elements which invokes SXML elements from gen and returns the elements themselves and then all of their preceding elements in reverse document order on successive invocations.

Paths

TBD

Error handling

Errors are signaled using objects of a disjoint type. They contain an $error or $pi event.

(uxml-error? obj)

Returns #t if obj belongs to the error type, and #f otherwise.

(uxml-error-event xml-error)

Returns an $error or $pi event (i.e. a list) encapsulated in an error object.