Class PullParser

temporary array of current attributes

index for last attribute in attrPos array

size of attrPos array

Should attribute uniqueness be checked for attributes as in specified XML and NS specifications?

Content of current element if in CONTENT state

temprary array to keep ElementContent stack

how many elements are on elStack

size of elStack array

Have we read empty element?

end position of current event in tokenizer biffer

start position of current event in tokenizer biffer

mapping of names prefixes to uris

should parser report namespace xmlns* attributes ?

Have we seen root element

what is current event type as returned from next()?

should parser support namespaces?

what is current token returned from tokeizer

XML tokenizer that is doing actual tokenizning of input stream.

Create instance of pull parser.

Make sure that in attributes temporary array is enough space.

Make sure that we have enough space to keep element stack if passed size.

Return string describing current position of parser in input stream.

Return true if just read CONTENT contained only white spaces.

This is key method - it reads more from input stream and returns next event type (such as START_TAG, END_TAG, CONTENT). or END_DOCUMENT if no more input.

This is simple automata (in pseudo-code):

 byte next() {
    while(state != END_DOCUMENT) {
      token = tokenizer.next();  // get next XML token
      switch(token) {

      case Tokenizer.END_DOCUMENT:
        return state = END_DOCUMENT

      case Tokenizer.CONTENT:
        // check if content allowed - only inside element
        return state = CONTENT

     case Tokenizer.ETAG_NAME:
        // popup element from stack - compare if matched start and end tag
        // if namespaces supported restore namespaces prefix mappings
        return state = END_TAG;

      case Tokenizer.STAG_NAME:
        // create new  element push it on stack
        // process attributes (including namespaces)
        // set emptyElement = true; if empty element
        // check atribute uniqueness (including nmespacese prefixes)
        return state = START_TAG;

      }
    }
 }
 

Actual parsing is more complex especilly for start tag due to dealing with attributes reported separately from tokenizer and declaring namespace prefixes and uris.

Return String that contains just read CONTENT.

Read value of just read END_TAG into passed as argument EndTag.

Return namespace prefixes for element at depth

Return namespace URIs for element at depth

Read value of just read START_TAG into passed as argument StartTag.

Reset parser state so it can be used to parse new

Allow for mixed element content. Enabled by default. When disbaled element must containt either text or other elements.

Reset parser and set new input.

Reset parser and set new input.

Make parser to report xmlns* attributes. Disabled by default. Only meaningful when namespaces are enabled (when namespaces are disabled all attributes are always reported).

Set support of namespaces. Disabled by default.

If parser has just read start tag it allows to skip whoole subtree contined in this element. Returns when encounters end tag matching the start tag.