XML Interface

Table 9-1 summarizes the methods of available through the XML interface.

XmlAccess

Sets the open/read/close callbacks used to load data for a specific URL access method. Overrides the built-in data loading functions for HTTP, FTP, and so on, or provides functions to handle new types, such as UNKNOWN.

xmlerr XmlAccess(
   xmlctx *xctx, 
   xmlurlacc access, 
   void *userctx,
   XML_ACCESS_OPEN_F(
      (*openf),
      ctx,
      uri,
      parts,
      length,
      uh),
   XML_ACCESS_READ_F(
      (*readf),
      ctx,
      uh,
      data,
      nraw,
      eoi),
   XML_ACCESS_CLOSE_F(
      (*closef), 
      ctx,
      uh));

xctx

IN

access

IN

userctx

IN

openf

IN

readf

IN

closef

IN

(xmlerr) numeric error code, XMLERR_OK [0] on success

XmlCreate

Create an XML Developer's Toolkit xmlctx. Properties common to all xmlctx's (both XDK and XMLType) are:

• "data_encoding", name of data encoding) The encoding in which XML data will be presented through DOM and SAX. If not specified, the default is UTF-8 (or UTF-E on EBCDIC platforms). Note that single-byte encodings such as EBCDIC or ISO-8859 are substantially faster than multibyte encodings like UTF-8; Unicode (UTF-16) uses more memory but has better performance than multibyte.

• BEGIN_NO_DOC ("data_lid", data encoding lid) The data encoding specified as an NLS lx_langid; the matching NLS global area must also be specified. END_NO_DOC

• "default_input_encoding", name of default input encoding) If the encoding of an input document cannot be automatically determined through BOM, XMLDecl, protocol header, and so on, then this encoding will be assumed.

• BEGIN_NO_DOC ("default_input_lid", default input encoding lid) The default input encoding specified as an NLS lx_langid; the matching NLS global area must also be specified. END_NO_DOC

• "error_language", error language or language.encoding) The language (and optional encoding) in which error messages are created. The default is American with UTF-8 encoding. To specify just the language, give the name of the language and nothing else ("American"); To also specify the encoding, add a dot and the Oracle name of the encoding ("American.WE8ISO8859P1").

• "error_handler", function pointer, see XML_ERRMSG_F) Default behavior on errors is to output the formatted message to stderr. If an error handler is provided, the formatted message will be passed to it instead of being printed.

• "error_context", user-defined context for error handler) This is a context pointer to be passed to the error handler function. It's meaning is user-defined; it is just specified here and passed along when an error occurs.

• "input_encoding", name of forced input encoding) The forced input encoding for input documents. Used to override a document's XMLDecl, and so on, and always interpret it in the given encoding. Use of this feature is strongly discouraged. It should be not necessary in normal use, as BOMs, XMLDecls, and so on, when existing, should be correct.

• BEGIN_NO_DOC ("input_lid", INLID, POINTER), The forced input encoding

• ("lpu_context", lpu context) The LPU context used for URL data loading and access-method hooking. If one is not provided, it will be made for you.

• ("lml_context", LMLCTX, POINTER), The LML context used for low-level memory allocation. If not provided, one will be made. From the outside, end-users have to set memory_alloc, memory_free, and so on. END_NO_DOC

• ("memory_alloc", low-level memory allocation function) Low-level memory allocation function, if malloc is not to be used. If provided, the matching free function must also be given. See XML_ALLOC_F.

• ("memory_free", low-level memory freeing function) Low-level memory freeing function, if free is not to be used. Matches the alloc function.

• ("memory_context", user-defined memory context) User-defined memory context which is passed to the alloc and free functions. Its definition and use is entirely up to the user; it is just set here and passed to the callbacks.

• BEGIN_NO_DOC ("nls_global_area", NLS global area, lx_glo) If any encoding are specified as NLS lids, the matching NLS global area must also be specified. END_NO_DOC

• The XDK has properties of its own, that only apply to an XDK type xmlctx (the previous properties were all general and applied to all xmlctx's).

• ("input_buffer_size", size in characters of input buffer) This is the basic I/O buffer size. Default is 256K, minimum is 4K and maximum is 4MB. Depending on the encoding, 1, 2 or 3 of these buffers may be needed. Note size is in characters, not bytes. If the buffer holds Unicode data, it will be twice as large.

• ("memory_block_size", size in bytes of memory allocation unit) This is the size of chunk the high-level memory package will request from the low-level allocator; i.e., the basic unit of memory allocation. Default is 64K, minimum is 16K and maximum is 256K.

xmlctx *XmlCreate(
   xmlerr *err, 
   oratext *name,
   list);

err

OUT

access

IN

list

IN

(xmlctx *) created xmlctx [or NULL on error with err set]

XmlCreateDTD

Create DTD.

xmldocnode* XmlCreateDTD(
   xmlctx *xctx
   oratext *qname,
   oratext *pubid,
   oratext *sysid,
   xmlerr *err)

xctx

IN

qname

IN

pubid

IN

sysid

IN

err

OUT

(xmldtdnode *) new DTD node

XmlCreateDocument

Creates the initial top-level DOCUMENT node and its supporting infrastructure. If a qualified name is provided, a an element with that name is created and set as the document's root element.

xmldocnode* XmlCreateDocument(
   xmlctx *xctx,
   oratext *uri,
   oratext *qname, 
   xmldtdnode *dtd,
   xmlerr *err)

xctx

IN

uri

IN

qname

IN

dtd

IN

err

OUT

(xmldocnode *) new Document object.

XmlDestroy

Destroys an xmlctx

void XmlDestroy(
   xmlctx *xctx)

xctx

IN

XmlFreeDocument

Destroys a document created by XmlCreateDocument or through one of the Load functions. Releases all resources associated with the document, which is then invalid.

void XmlFreeDocument(
   xmlctx *xctx,
   xmldocnode *doc)

xctx

IN

doc

IN

XmlGetEncoding

Returns data encoding in use by XML context. Ordinarily, the data encoding is chosen by the user, so this function is not needed. However, if the data encoding is not specified, and allowed to default, this function can be used to return the name of that default encoding.

oratext *XmlGetEncoding(
   xmlctx *xctx);

xctx

IN

(oratext *) name of data encoding

XmlHasFeature

Determine if a DOM feature is implemented. Returns TRUE if the feature is implemented in the specified version, FALSE otherwise.

In level 1, the legal values for package are 'HTML' and 'XML' (case-insensitive), and the version is the string "1.0". If the version is not specified, supporting any version of the feature will cause the method to return TRUE.

• DOM 1.0 features are "XML" and "HTML".

• DOM 2.0 features are "Core", "XML", "HTML", "Views", "StyleSheets", "CSS", "CSS2", "Events", "UIEvents", "MouseEvents", "MutationEvents", "HTMLEvents", "Range", "Traversal"

boolean XmlHasFeature(
   xmlctx *xctx,
   oratext *feature,
   oratext *version)

xctx

IN

feature

IN

version

IN

(boolean) feature is implemented?

XmlIsSimple

Returns a flag saying whether the context's data encoding is "simple", single-byte for each character, like ASCII or EBCDIC.

boolean XmlIsSimple(
   xmlctx *xctx);

xctx

IN

(boolean) TRUE of data encoding is "simple", FALSE otherwise

XmlIsUnicode

Returns a flag saying whether the context's data encoding is Unicode, UTF-16, with two-byte for each character.

boolean XmlIsUnicode(
   xmlctx *xctx);

xctx

IN

(boolean) TRUE of data encoding is Unicode, FALSE otherwise

XmlLoadDom

Loads (parses) an XML document from an input source and creates a DOM. The root document node is returned on success, or NULL on failure (with err set).

The function takes two fixed arguments, the xmlctx and an error return code, then zero or more (property, value) pairs, then NULL.

SOURCE Input source is set by one of the following mutually exclusive properties (choose one):

• ("uri", document URI) [compiler encoding]

• ("file", document filesystem path) [compiler encoding]

• ("buffer", address of buffer, "buffer_length", # bytes in buffer)

• ("stream", address of stream object, "stream_context", pointer to stream object's context)

• ("stdio", FILE* stream)

PROPERTIES Additional properties:

• ("dtd", DTD node) DTD for document

• ("base_uri", document base URI) for documents loaded from other sources than a URI, sets the effective base URI. the document's base URI is needed in order to resolve relative URI include, import, and so on.

• ("input_encoding", encoding name) forced input encoding [name]

• ("default_input_encoding", encoding_name) default input encoding to assume if document is not self-describing (no BOM, protocol header, XMLDecl, and so on)

• ("schema_location", string) schemaLocation of schema for this document. used to figure optimal layout when loading documents into a database

• ("validate", boolean) when TRUE, turns on DTD validation; by default, only well-formedness is checked. note that schema validation is a separate beast.

• ("discard_whitespace", boolean) when TRUE, formatting whitespace between elements (newlines and indentation) in input documents is discarded. by default, ALL input characters are preserved.

• ("dtd_only", boolean) when TRUE, parses an external DTD, not a complete XML document.

• ("stop_on_warning", boolean) when TRUE, warnings are treated the same as errors and cause parsing, validation, and so on, to stop immediately. by default, warnings are issued but the game continues.

• ("warn_duplicate_entity", boolean) when TRUE, entities which are declared more than once will cause warnings to be issued. the default is to accept the first declaration and silently ignore the rest.

• ("no_expand_char_ref", boolean) when TRUE, causes character references to be left unexpanded in the DOM data. ordinarily, character references are replaced by the character they represent. however, when a document is saved those characters entities do not reappear. to way to ensure they remain through load and save is to not expand them.

• ("no_check_chars", boolean) when TRUE, omits the test of XML [2] Char production: all input characters will be accepted as valid

xmldocnode *XmlLoadDom(
   xmlctx *xctx, 
   xmlerr *err, 
   list);

xctx

IN

err

OUT

list

IN

(xmldocnode *) document node on success [NULL on failure with err set]

XmlLoadSax

Loads (parses) an XML document from an input source and generates a set of SAX events (as user callbacks). Input sources and basic set of properties is the same as for XmlLoadDom.

xmlerr XmlLoadSax(
   xmlctx *xctx,
   xmlsaxcb *saxcb,
   void *saxctx, 
   list);

xctx

IN

saxcb

IN

saxctx

IN

list

IN

(xmlerr) numeric error code, XMLERR_OK [0] on success

XmlLoadSaxVA

Loads (parses) an XML document from an input source and generates a set of SAX events (as user callbacks). Input sources and basic set of properties is the same as for XmlLoadDom.

xmlerr XmlLoadSaxVA(
   xmlctx *xctx, 
   xmlsaxcb *saxcb, 
   void *saxctx, 
   va_list va);

xctx

IN

saxcb

IN

saxctx

IN

va

IN

(xmlerr) numeric error code, XMLERR_OK [0] on success

XmlSaveDom

Serializes document or subtree to the given destination and returns the number of bytes written; if no destination is provided, just returns formatted size but does not output.

If an output encoding is specified, the document will be re-encoded on output; otherwise, it will be in its existing encoding.

The top level is indented step*level spaces, the next level step*(level+1) spaces, and so on.

When saving to a buffer, if the buffer overflows, 0 is returned and err is set to XMLERR_SAVE_OVERFLOW.

DESTINATION Output destination is set by one of the following mutually exclusive properties (choose one):

• ("uri", document URI) POST, PUT? [compiler encoding]

• ("file", document filesystem path) [compiler encoding]

• ("buffer", address of buffer, "buffer_length", # bytes in buffer)

• ("stream", address of stream object, "stream_context", pointer to stream object's context)

PROPERTIES Additional properties:

• ("output_encoding", encoding name) name of final encoding for document. unless specified, saved document will be in same encoding as xmlctx.

• ("indent_step", unsigned) spaces to indent each level of output. default is 4, 0 means no indentation.

• ("indent_level", unsigned) initial indentation level. default is 0, which means no indentation, flush left.

• ("xmldecl", boolean) include an XMLDecl in the output document. ordinarily an XMLDecl is output for a compete document (root node is DOC).

• ("bom", boolean) input a BOM in the output document. usually the BOM is only needed for certain encodings (UTF-16), and optional for others (UTF-8). causes optional BOMs to be output.

• ("prune", boolean) prunes the output like the unix 'find' command; does not not descend to children, just prints the one node given.

ubig_ora XmlSaveDom(
   xmlctx *xctx,
   xmlerr *err,
   xmlnode *root,
   list);

xctx

IN

err

OUT

root

IN

list

IN

(ubig_ora) number of bytes written to destination

XmlVersion

Returns the version string for the XDK

oratext *XmlVersion();

(oratext *) version string