Sometimes the DOM tree output is just too large to fit
reasonablyintomemory. In that case (and if you don't expect to save back the
XMLdocumentloaded using libxml), it's better to use the SAX interface of
libxml.SAX isa callback-based interfaceto the parser.
Beforeparsing,the application layer registers a customized set of callbacks
whicharecalled by the library as it progresses through the XML input. To get more detailed step-by-step guidance on using the SAX
interfaceoflibxml, see the nicedocumentation.writtenby
JamesHenstridge. You can debug the SAX behaviour by using
thetestSAXprogram located in the gnome-xml module (it's
usuallynot shipped in thebinary packages of libxml, but you can find it in
the tarsourcedistribution). Here is the sequence of callbacks that would be
reportedbytestSAX when parsing the example XML document shown earlier: SAX.setDocumentLocator()
SAX.startDocument()
SAX.getEntity(amp)
SAX.startElement(EXAMPLE, prop1='gnome is great', prop2='& linux too')
SAX.characters( , 3)
SAX.startElement(head)
SAX.characters( , 4)
SAX.startElement(title)
SAX.characters(Welcome to Gnome, 16)
SAX.endElement(title)
SAX.characters( , 3)
SAX.endElement(head)
SAX.characters( , 3)
SAX.startElement(chapter)
SAX.characters( , 4)
SAX.startElement(title)
SAX.characters(The Linux adventure, 19)
SAX.endElement(title)
SAX.characters( , 4)
SAX.startElement(p)
SAX.characters(bla bla bla ..., 15)
SAX.endElement(p)
SAX.characters( , 4)
SAX.startElement(image, href='linus.gif')
SAX.endElement(image)
SAX.characters( , 4)
SAX.startElement(p)
SAX.characters(..., 3)
SAX.endElement(p)
SAX.characters( , 3)
SAX.endElement(chapter)
SAX.characters( , 1)
SAX.endElement(EXAMPLE)
SAX.endDocument() Most of the other interfaces of libxml2 are based on the
DOMtree-buildingfacility, so nearly everything up to the end of this
documentpresupposes theuse of the standard DOM tree build. Note that the DOM
treeitself is built bya set of registered default callbacks, without
internalspecificinterface. Daniel Veillard |
