defer(function() {
var err;
try {
parser.parse(xml);
resolveReferences();
} catch (e) {
err = e;
}
var element = rootHandler.element;
// handle the situation that we could not extract
// the desired root element from the document
if (!err && !element) {
err = error('failed to parse document as <' + rootHandler.type.$descriptor.name + '>');
}
done(err, err ? undefined : element, context);
});
Reader.prototype.fromXML = function(xml, options, done) {
var rootHandler = options.rootHandler;
if (options instanceof ElementHandler) {
// root handler passed via (xml, { rootHandler: ElementHandler }, ...)
rootHandler = options;
options = {};
} else {
if (typeof options === 'string') {
// rootHandler passed via (xml, 'someString', ...)
rootHandler = this.handler(options);
options = {};
} else if (typeof rootHandler === 'string') {
// rootHandler passed via (xml, { rootHandler: 'someString' }, ...)
rootHandler = this.handler(rootHandler);
}
}
var model = this.model,
lax = this.lax;
var context = new Context(assign({}, options, { rootHandler: rootHandler })),
parser = new SaxParser({ proxy: true }),
stack = createStack();
rootHandler.context = context;
// push root handler
stack.push(rootHandler);
/**
* Handle error.
*
* @param {Error} err
* @param {Function} getContext
* @param {boolean} lax
*
* @return {boolean} true if handled
*/
function handleError(err, getContext, lax) {
var ctx = getContext();
var line = ctx.line,
column = ctx.column,
data = ctx.data;
// we receive the full context data here,
// for elements trim down the information
// to the tag name, only
if (data.charAt(0) === '<' && data.indexOf(' ') !== -1) {
data = data.slice(0, data.indexOf(' ')) + '>';
}
var message =
'unparsable content ' + (data ? data + ' ' : '') + 'detected\n\t' +
'line: ' + line + '\n\t' +
'column: ' + column + '\n\t' +
'nested error: ' + err.message;
if (lax) {
context.addWarning({
message: message,
error: err
});
return true;
} else {
throw error(message);
}
}
function handleWarning(err, getContext) {
// just like handling errors in <lax=true> mode
return handleError(err, getContext, true);
}
/**
* Resolve collected references on parse end.
*/
function resolveReferences() {
var elementsById = context.elementsById;
var references = context.references;
var i, r;
for (i = 0; (r = references[i]); i++) {
var element = r.element;
var reference = elementsById[r.id];
var property = getModdleDescriptor(element).propertiesByName[r.property];
if (!reference) {
context.addWarning({
message: 'unresolved reference <' + r.id + '>',
element: r.element,
property: r.property,
value: r.id
});
}
if (property.isMany) {
var collection = element.get(property.name),
idx = collection.indexOf(r);
// we replace an existing place holder (idx != -1) or
// append to the collection instead
if (idx === -1) {
idx = collection.length;
}
if (!reference) {
// remove unresolvable reference
collection.splice(idx, 1);
} else {
// add or update reference in collection
collection[idx] = reference;
}
} else {
element.set(property.name, reference);
}
}
}
function handleClose() {
stack.pop().handleEnd();
}
var PREAMBLE_START_PATTERN = /^<\?xml /i;
var ENCODING_PATTERN = / encoding="([^"]+)"/i;
var UTF_8_PATTERN = /^utf-8$/i;
function handleQuestion(question) {
if (!PREAMBLE_START_PATTERN.test(question)) {
return;
}
var match = ENCODING_PATTERN.exec(question);
var encoding = match && match[1];
if (!encoding || UTF_8_PATTERN.test(encoding)) {
return;
}
context.addWarning({
message:
'unsupported document encoding <' + encoding + '>, ' +
'falling back to UTF-8'
});
}
function handleOpen(node, getContext) {
var handler = stack.peek();
try {
stack.push(handler.handleNode(node));
} catch (err) {
if (handleError(err, getContext, lax)) {
stack.push(new NoopHandler());
}
}
}
function handleCData(text, getContext) {
try {
stack.peek().handleText(text);
} catch (err) {
handleWarning(err, getContext);
}
}
function handleText(text, getContext) {
// strip whitespace only nodes, i.e. before
// <!CDATA[ ... ]> sections and in between tags
text = text.trim();
if (!text) {
return;
}
handleCData(text, getContext);
}
var uriMap = model.getPackages().reduce(function(uriMap, p) {
uriMap[p.uri] = p.prefix;
return uriMap;
}, {});
parser
.ns(uriMap)
.on('openTag', function(obj, decodeStr, selfClosing, getContext) {
// gracefully handle unparsable attributes (attrs=false)
var attrs = obj.attrs || {};
var decodedAttrs = Object.keys(attrs).reduce(function(d, key) {
var value = decodeStr(attrs[key]);
d[key] = value;
return d;
}, {});
var node = {
name: obj.name,
originalName: obj.originalName,
attributes: decodedAttrs,
ns: obj.ns
};
handleOpen(node, getContext);
})
.on('question', handleQuestion)
.on('closeTag', handleClose)
.on('cdata', handleCData)
.on('text', function(text, decodeEntities, getContext) {
handleText(decodeEntities(text), getContext);
})
.on('error', handleError)
.on('warn', handleWarning);
// deferred parse XML to make loading really ascnchronous
// this ensures the execution environment (node or browser)
// is kept responsive and that certain optimization strategies
// can kick in
defer(function() {
var err;
try {
parser.parse(xml);
resolveReferences();
} catch (e) {
err = e;
}
var element = rootHandler.element;
// handle the situation that we could not extract
// the desired root element from the document
if (!err && !element) {
err = error('failed to parse document as <' + rootHandler.type.$descriptor.name + '>');
}
done(err, err ? undefined : element, context);
});
};