词法分析

开发平台：

Visual C++

SGXMLScanner.cpp：源码内容

// the attribute should be qualified
fValidator->emitError
(
XMLValid::AttributeNotUnQualified
, attDef->getFullName()
);
((SchemaAttDef *)(attDef))->setValidity(PSVIDefs::INVALID);
}
}
}
}
}
// Find this attribute within the parent element. We pass both
// the uriID/name and the raw QName buffer, since we don't know
// how the derived validator and its elements store attributes.
bool wasAdded = false;
if (!attDef) {
attDef = elemDecl->findAttr
(
curPair->getKey()
, uriId
, suffPtr
, prefPtr
, XMLElementDecl::AddIfNotFound
, wasAdded
);
}
if(!skipThisOne && fGrammarType == Grammar::SchemaGrammarType) {
//we may have set it to invalid already, but this is the first time we are guarenteed to have the attDef
if(((SchemaAttDef *)(attDef))->getValidity() != PSVIDefs::INVALID)
((SchemaAttDef *)(attDef))->setValidity(PSVIDefs::VALID);
((SchemaAttDef *)(attDef))->setValidationAttempted(PSVIDefs::FULL);
}
if (wasAdded)
{
// This is to tell the Validator that this attribute was
// faulted-in, was not an attribute in the attdef originally
attDef->setCreateReason(XMLAttDef::JustFaultIn);
}
bool errorCondition = fValidate && !attDefForWildCard &&
attDef->getCreateReason() == XMLAttDef::JustFaultIn && !attDef->getProvided();
if (errorCondition && !skipThisOne && !laxThisOne)
{
//
// Its not valid for this element, so issue an error if we are
// validating.
//
XMLBufBid bbMsg(&fBufMgr);
XMLBuffer& bufMsg = bbMsg.getBuffer();
if (uriId != fEmptyNamespaceId) {
XMLBufBid bbURI(&fBufMgr);
XMLBuffer& bufURI = bbURI.getBuffer();
getURIText(uriId, bufURI);
bufMsg.append(chOpenCurly);
bufMsg.append(bufURI.getRawBuffer());
bufMsg.append(chCloseCurly);
}
bufMsg.append(suffPtr);
fValidator->emitError
(
XMLValid::AttNotDefinedForElement
, bufMsg.getRawBuffer()
, elemDecl->getFullName()
);
((SchemaAttDef *)attDef)->setValidity(PSVIDefs::INVALID);
}
else if(errorCondition && laxThisOne) {
((SchemaAttDef *)(attDef))->setValidationAttempted(PSVIDefs::NONE);
((SchemaAttDef *)(attDef))->setValidity(PSVIDefs::UNKNOWN);
}
// If its already provided, then there are more than one of
// this attribute in this start tag, so emit an error.
if (attDef->getProvided())
{
emitError
(
XMLErrs::AttrAlreadyUsedInSTag
, attDef->getFullName()
, elemDecl->getFullName()
);
((SchemaAttDef *)(attDef))->setValidity(PSVIDefs::INVALID);
}
else
{
attDef->setProvided(true);
}
// Now normalize the raw value since we have the attribute type. We
// don't care about the return status here. If it failed, an error
// was issued, which is all we care about.
if (attDefForWildCard) {
((SchemaAttDef*)attDef)->setAnyDatatypeValidator(((SchemaAttDef*) attDefForWildCard)->getDatatypeValidator());
normalizeAttValue
(
attDefForWildCard
, curPair->getValue()
, normBuf
);
// If we found an attdef for this one, then lets validate it.
if (fNormalizeData)
{
// normalize the attribute according to schema whitespace facet
XMLBufBid bbtemp(&fBufMgr);
XMLBuffer& tempBuf = bbtemp.getBuffer();
DatatypeValidator* tempDV = ((SchemaAttDef*) attDefForWildCard)->getDatatypeValidator();
((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, normBuf.getRawBuffer(), tempBuf);
normBuf.set(tempBuf.getRawBuffer());
}
if (fValidate && !skipThisOne) {
fValidator->validateAttrValue
(
attDefForWildCard
, normBuf.getRawBuffer()
, false
, elemDecl
);
}
// Save the type for later use
attType = attDefForWildCard->getType();
((SchemaElementDecl *)(elemDecl))->updateValidityFromAttribute((SchemaAttDef *)attDef);
DatatypeValidator* tempDV = ((SchemaAttDef*) attDefForWildCard)->getDatatypeValidator();
if(tempDV && tempDV->getType() == DatatypeValidator::Union)
((SchemaAttDef*)attDef)->setMembertypeValidator(((UnionDatatypeValidator *)tempDV)->getMemberTypeValidator());
}
else {
normalizeAttValue
(
attDef
, curPair->getValue()
, normBuf
);
// If we found an attdef for this one, then lets validate it.
if (attDef->getCreateReason() != XMLAttDef::JustFaultIn)
{
if (fNormalizeData && (fGrammarType == Grammar::SchemaGrammarType))
{
// normalize the attribute according to schema whitespace facet
XMLBufBid bbtemp(&fBufMgr);
XMLBuffer& tempBuf = bbtemp.getBuffer();
DatatypeValidator* tempDV = ((SchemaAttDef*) attDef)->getDatatypeValidator();
((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, normBuf.getRawBuffer(), tempBuf);
normBuf.set(tempBuf.getRawBuffer());
}
if (fValidate && !skipThisOne)
{
fValidator->validateAttrValue
(
attDef
, normBuf.getRawBuffer()
, false
, elemDecl
);
}
}
// Save the type for later use
attType = attDef->getType();
((SchemaElementDecl *)(elemDecl))->updateValidityFromAttribute((SchemaAttDef *)attDef);
}
}
else
{
// Just normalize as CDATA
attType = XMLAttDef::CData;
normalizeAttRawValue
(
curPair->getKey()
, curPair->getValue()
, normBuf
);
}
// Add this attribute to the attribute list that we use to pass them
// to the handler. We reuse its existing elements but expand it as
// required.
XMLAttr* curAttr;
if (retCount >= curAttListSize)
{
curAttr = new (fMemoryManager) XMLAttr
(
uriId
, suffPtr
, prefPtr
, normBuf.getRawBuffer()
, attType
, true
, fMemoryManager
);
toFill.addElement(curAttr);
}
else
{
curAttr = toFill.elementAt(retCount);
curAttr->set
(
uriId
, suffPtr
, prefPtr
, normBuf.getRawBuffer()
, attType
);
curAttr->setSpecified(true);
}
// Bump the count of attrs in the list
retCount++;
}
// Now, if there are any attributes declared by this element, let's
// go through them and make sure that any required ones are provided,
// and fault in any fixed ones and defaulted ones that are not provided
// literally.
if (hasDefs)
{
// Check after all specified attrs are scanned
// (1) report error for REQUIRED attrs that are missing (V_TAGc)
// (2) add default attrs if missing (FIXED and NOT_FIXED)
XMLAttDefList& attDefList = elemDecl->getAttDefList();
while (attDefList.hasMoreElements())
{
// Get the current att def, for convenience and its def type
XMLAttDef *curDef = &attDefList.nextElement();
const XMLAttDef::DefAttTypes defType = curDef->getDefaultType();
if (!curDef->getProvided())
{
((SchemaAttDef *)curDef)->setValidationAttempted(PSVIDefs::FULL);
((SchemaAttDef *)curDef)->setValidity(PSVIDefs::VALID);
//the attributes is not provided
if (fValidate)
{
// If we are validating and its required, then an error
if ((defType == XMLAttDef::Required) ||
(defType == XMLAttDef::Required_And_Fixed) )
{
fValidator->emitError
(
XMLValid::RequiredAttrNotProvided
, curDef->getFullName()
);
((SchemaAttDef *)(curDef))->setValidity(PSVIDefs::INVALID);
}
else if ((defType == XMLAttDef::Default) ||
(defType == XMLAttDef::Fixed) )
{
if (fStandalone && curDef->isExternal())
{
// XML 1.0 Section 2.9
// Document is standalone, so attributes must not be defaulted.
fValidator->emitError(XMLValid::NoDefAttForStandalone, curDef->getFullName(), elemDecl->getFullName());
((SchemaAttDef *)(curDef))->setValidity(PSVIDefs::INVALID);
}
}
}
// Fault in the value if needed, and bump the att count.
// We have to
if ((defType == XMLAttDef::Default)
|| (defType == XMLAttDef::Fixed))
{
// Let the validator pass judgement on the attribute value
if (fValidate)
{
fValidator->validateAttrValue
(
curDef
, curDef->getValue()
, false
, elemDecl
);
}
XMLAttr* curAtt;
if (retCount >= curAttListSize)
{
curAtt = new (fMemoryManager) XMLAttr(fMemoryManager);
fValidator->faultInAttr(*curAtt, *curDef);
fAttrList->addElement(curAtt);
}
else
{
curAtt = fAttrList->elementAt(retCount);
fValidator->faultInAttr(*curAtt, *curDef);
}
// Indicate it was not explicitly specified and bump count
curAtt->setSpecified(false);
retCount++;
}
((SchemaElementDecl *)elemDecl)->updateValidityFromAttribute((SchemaAttDef *)curDef);
}
else
{
//attribute is provided
// (schema) report error for PROHIBITED attrs that are present (V_TAGc)
if (defType == XMLAttDef::Prohibited && fValidate) {
fValidator->emitError
(
XMLValid::ProhibitedAttributePresent
, curDef->getFullName()
);
((SchemaAttDef *)curDef)->setValidity(PSVIDefs::INVALID);
}
((SchemaElementDecl *)elemDecl)->updateValidityFromAttribute((SchemaAttDef *)curDef);
}
}
}
return retCount;
}
// This method will take a raw attribute value and normalize it according to
// the rules of the attribute type. It will put the resulting value into the
// passed buffer.
//
// This code assumes that escaped characters in the original value (via char
// refs) are prefixed by a 0xFFFF character. This is because some characters
// are legal if escaped only. And some escape chars are not subject to
// normalization rules.
bool SGXMLScanner::normalizeAttValue( const XMLAttDef* const attDef
, const XMLCh* const value
, XMLBuffer& toFill)
{
// A simple state value for a whitespace processing state machine
enum States
{
InWhitespace
, InContent
};
// Get the type and name
const XMLAttDef::AttTypes type = attDef->getType();
const XMLCh* const attrName = attDef->getFullName();
// Assume its going to go fine, and empty the target buffer in preperation
bool retVal = true;
toFill.reset();
// Get attribute def - to check to see if it's declared externally or not
bool isAttExternal = attDef->isExternal();
// Loop through the chars of the source value and normalize it according
// to the type.
States curState = InContent;
bool escaped;
bool firstNonWS = false;
XMLCh nextCh;
const XMLCh* srcPtr = value;
while (*srcPtr)
{
// Get the next character from the source. We have to watch for
// escaped characters (which are indicated by a 0xFFFF value followed
// by the char that was escaped.)
nextCh = *srcPtr;
escaped = (nextCh == 0xFFFF);
if (escaped)
nextCh = *++srcPtr;
// If its not escaped, then make sure its not a < character, which is
// not allowed in attribute values.
if (!escaped && (*srcPtr == chOpenAngle))
{
emitError(XMLErrs::BracketInAttrValue, attrName);
retVal = false;
}
if (type == XMLAttDef::CData || type > XMLAttDef::Notation)
{
if (!escaped)
{
if ((nextCh == 0x09) || (nextCh == 0x0A) || (nextCh == 0x0D))
{
// Check Validity Constraint for Standalone document declaration
// XML 1.0, Section 2.9
if (fStandalone && fValidate && isAttExternal)
{
// Can't have a standalone document declaration of "yes" if attribute
// values are subject to normalisation
fValidator->emitError(XMLValid::NoAttNormForStandalone, attrName);
((SchemaAttDef *)attDef)->setValidity(PSVIDefs::INVALID);
}
nextCh = chSpace;
}
}
}
else
{
if (curState == InWhitespace)
{
if (!fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
{
if (firstNonWS)
toFill.append(chSpace);
curState = InContent;
firstNonWS = true;
}
else
{
srcPtr++;
continue;
}
}
else if (curState == InContent)
{
if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
{
curState = InWhitespace;
srcPtr++;
// Check Validity Constraint for Standalone document declaration
// XML 1.0, Section 2.9
if (fStandalone && fValidate && isAttExternal)
{
if (!firstNonWS || (nextCh != chSpace) || (!*srcPtr) || fReaderMgr.getCurrentReader()->isWhitespace(*srcPtr))
{
// Can't have a standalone document declaration of "yes" if attribute
// values are subject to normalisation
fValidator->emitError(XMLValid::NoAttNormForStandalone, attrName);
((SchemaAttDef *)attDef)->setValidity(PSVIDefs::INVALID);
}
}
continue;
}
firstNonWS = true;
}
}
// Add this char to the target buffer
toFill.append(nextCh);
// And move up to the next character in the source
srcPtr++;
}
((SchemaElementDecl *)fElemStack.topElement()->fThisElement)->updateValidityFromAttribute((SchemaAttDef *)attDef);
return retVal;
}
// This method will just normalize the input value as CDATA without
// any standalone checking.
bool SGXMLScanner::normalizeAttRawValue( const XMLCh* const attrName
, const XMLCh* const value
, XMLBuffer& toFill)
{
// Assume its going to go fine, and empty the target buffer in preperation
bool retVal = true;
toFill.reset();
// Loop through the chars of the source value and normalize it according
// to the type.
bool escaped;
XMLCh nextCh;
const XMLCh* srcPtr = value;
while (*srcPtr)
{
// Get the next character from the source. We have to watch for
// escaped characters (which are indicated by a 0xFFFF value followed
// by the char that was escaped.)
nextCh = *srcPtr;
escaped = (nextCh == 0xFFFF);
if (escaped)
nextCh = *++srcPtr;
// If its not escaped, then make sure its not a < character, which is
// not allowed in attribute values.
if (!escaped && (*srcPtr == chOpenAngle))
{
emitError(XMLErrs::BracketInAttrValue, attrName);
retVal = false;
}
if (!escaped)
{
// NOTE: Yes this is a little redundant in that a 0x20 is
// replaced with an 0x20. But its faster to do this (I think)
// than checking for 9, A, and D separately.
if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
nextCh = chSpace;
}
// Add this char to the target buffer
toFill.append(nextCh);
// And move up to the next character in the source
srcPtr++;
}
return retVal;
}
unsigned int
SGXMLScanner::resolvePrefix( const XMLCh* const prefix
, const ElemStack::MapModes mode)
{
// Watch for the special namespace prefixes. We always map these to
// special URIs. 'xml' gets mapped to the official URI that its defined
// to map to by the NS spec. xmlns gets mapped to a special place holder
// URI that we define (so that it maps to something checkable.)
if (XMLString::equals(prefix, XMLUni::fgXMLNSString))
return fXMLNSNamespaceId;
else if (XMLString::equals(prefix, XMLUni::fgXMLString))
return fXMLNamespaceId;
// Ask the element stack to search up itself for a mapping for the
// passed prefix.
bool unknown;
unsigned int uriId = fElemStack.mapPrefixToURI(prefix, mode, unknown);
// If it was unknown, then the URI was faked in but we have to issue an error
if (unknown)
emitError(XMLErrs::UnknownPrefix, prefix);
return uriId;
}
unsigned int
SGXMLScanner::resolvePrefix( const XMLCh* const prefix
, XMLBuffer& bufToFill
, const ElemStack::MapModes mode)
{
// Watch for the special namespace prefixes. We always map these to
// special URIs. 'xml' gets mapped to the official URI that its defined
// to map to by the NS spec. xmlns gets mapped to a special place holder
// URI that we define (so that it maps to something checkable.)
if (XMLString::equals(prefix, XMLUni::fgXMLNSString))
return fXMLNSNamespaceId;
else if (XMLString::equals(prefix, XMLUni::fgXMLString))
return fXMLNamespaceId;
// Ask the element stack to search up itself for a mapping for the
// passed prefix.
bool unknown;
unsigned int uriId = fElemStack.mapPrefixToURI(prefix, mode, unknown);
// If it was unknown, then the URI was faked in but we have to issue an error
if (unknown)
emitError(XMLErrs::UnknownPrefix, prefix);
getURIText(uriId,bufToFill);
return uriId;
}
// This method will reset the scanner data structures, and related plugged
// in stuff, for a new scan session. We get the input source for the primary
// XML entity, create the reader for it, and push it on the stack so that
// upon successful return from here we are ready to go.
void SGXMLScanner::scanReset(const InputSource& src)
{
// This call implicitly tells us that we are going to reuse the scanner
// if it was previously used. So tell the validator to reset itself.
//
// But, if the fUseCacheGrammar flag is set, then don't reset it.
//
// NOTE: The ReaderMgr is flushed on the way out, because that is
// required to insure that files are closed.
fGrammarResolver->cacheGrammarFromParse(fToCacheGrammar);
fGrammarResolver->useCachedGrammarInParse(fUseCachedGrammar);
fGrammar = fSchemaGrammar;
fGrammarType = Grammar::DTDGrammarType;
fRootGrammar = 0;
fValidator->setGrammar(fGrammar);
if (fValidatorFromUser) {
((SchemaValidator*) fValidator)->setErrorReporter(fErrorReporter);
((SchemaValidator*) fValidator)->setGrammarResolver(fGrammarResolver);
((SchemaValidator*) fValidator)->setExitOnFirstFatal(fExitOnFirstFatal);
}
// Reset validation
fValidate = (fValScheme == Val_Always) ? true : false;
// And for all installed handlers, send reset events. This gives them
// a chance to flush any cached data.
if (fDocHandler)
fDocHandler->resetDocument();
if (fEntityHandler)
fEntityHandler->resetEntities();
if (fErrorReporter)
fErrorReporter->resetErrors();
// Clear out the id reference list
fIDRefList->removeAll();
// Reset IdentityConstraints
fValueStoreCache->startDocument();
fMatcherStack->clear();
// Reset the element stack, and give it the latest ids for the special
// URIs it has to know about.
fElemStack.reset
(
fEmptyNamespaceId
, fUnknownNamespaceId
, fXMLNamespaceId
, fXMLNSNamespaceId
);
if (!fSchemaNamespaceId)
fSchemaNamespaceId = fURIStringPool->addOrFind(SchemaSymbols::fgURI_XSI);
// Reset some status flags
fInException = false;
fStandalone = false;
fErrorCount = 0;
fHasNoDTD = true;
fSeeXsi = false;
fDoNamespaces = true;
fDoSchema = true;
// Reset the validators
fSchemaValidator->reset();
fSchemaValidator->setErrorReporter(fErrorReporter);
fSchemaValidator->setExitOnFirstFatal(fExitOnFirstFatal);
fSchemaValidator->setGrammarResolver(fGrammarResolver);
if (fValidatorFromUser)
fValidator->reset();
// Handle the creation of the XML reader object for this input source.
// This will provide us with transcoding and basic lexing services.
XMLReader* newReader = fReaderMgr.createReader
(
src
, true
, XMLReader::RefFrom_NonLiteral
, XMLReader::Type_General
, XMLReader::Source_External
, fCalculateSrcOfs
);
if (!newReader) {
if (src.getIssueFatalErrorIfNotFound())
ThrowXML1(RuntimeException, XMLExcepts::Scan_CouldNotOpenSource, src.getSystemId());
else
ThrowXML1(RuntimeException, XMLExcepts::Scan_CouldNotOpenSource_Warning, src.getSystemId());
}
// Push this read onto the reader manager
fReaderMgr.pushReader(newReader, 0);
// and reset security-related things if necessary:
if(fSecurityManager != 0)
{
fEntityExpansionLimit = fSecurityManager->getEntityExpansionLimit();
fEntityExpansionCount = 0;
}
}
// This method is called between markup in content. It scans for character
// data that is sent to the document handler. It watches for any markup
// characters that would indicate that the character data has ended. It also
// handles expansion of general and character entities.
//
// sendData() is a local static helper for this method which handles some
// code that must be done in three different places here.
void SGXMLScanner::sendCharData(XMLBuffer& toSend)
{
// If no data in the buffer, then nothing to do
if (toSend.isEmpty())
return;
// We do different things according to whether we are validating or
// not. If not, its always just characters; else, it depends on the
// current element's content model.
if (fValidate)
{
// Get the raw data we need for the callback
const XMLCh* const rawBuf = toSend.getRawBuffer();
const unsigned int len = toSend.getLen();
// And see if the current element is a 'Children' style content model
const ElemStack::StackElem* topElem = fElemStack.topElement();
// Get the character data opts for the current element
XMLElementDecl::CharDataOpts charOpts = topElem->fThisElement->getCharDataOpts();
if (charOpts == XMLElementDecl::NoCharData)
{
// They definitely cannot handle any type of char data
fValidator->emitError(XMLValid::NoCharDataInCM);
((SchemaElementDecl *)topElem->fThisElement)->setValidity(PSVIDefs::INVALID);
}
else if (fReaderMgr.getCurrentReader()->isAllSpaces(rawBuf, len))
{
// Its all spaces. So, if they can take spaces, then send it
// as ignorable whitespace. If they can handle any char data
// send it as characters.
if (charOpts == XMLElementDecl::SpacesOk) {
if (fDocHandler)
fDocHandler->ignorableWhitespace(rawBuf, len, false);
}
else if (charOpts == XMLElementDecl::AllCharData)
{
// The normalized data can only be as large as the
// original size, so this will avoid allocating way
// too much or too little memory.
XMLBuffer toFill(len+1, fMemoryManager);
toFill.set(rawBuf);
if (fNormalizeData) {
// normalize the character according to schema whitespace facet
XMLBufBid bbtemp(&fBufMgr);
XMLBuffer& tempBuf = bbtemp.getBuffer();
DatatypeValidator* tempDV = ((SchemaElementDecl*) topElem->fThisElement)->getDatatypeValidator();
((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, toFill.getRawBuffer(), tempBuf);
toFill.set(tempBuf.getRawBuffer());
}
// tell the schema validation about the character data for checkContent later
((SchemaValidator*) fValidator)->setDatatypeBuffer(toFill.getRawBuffer());
// call all active identity constraints
if (fMatcherStack->getMatcherCount())
fContent.append(toFill.getRawBuffer(), toFill.getLen());
if (fDocHandler)
fDocHandler->docCharacters(toFill.getRawBuffer(), toFill.getLen(), false);
}
}
else
{
// If they can take any char data, then send it. Otherwise, they
// can only handle whitespace and can't handle this stuff so
// issue an error.
if (charOpts == XMLElementDecl::AllCharData)
{
// The normalized data can only be as large as the
// original size, so this will avoid allocating way
// too much or too little memory.
XMLBuffer toFill(len+1, fMemoryManager);
toFill.set(rawBuf);
if (fNormalizeData) {
// normalize the character according to schema whitespace facet
XMLBufBid bbtemp(&fBufMgr);
XMLBuffer& tempBuf = bbtemp.getBuffer();
DatatypeValidator* tempDV = ((SchemaElementDecl*) topElem->fThisElement)->getDatatypeValidator();
((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, toFill.getRawBuffer(), tempBuf);
toFill.set(tempBuf.getRawBuffer());
}
// tell the schema validation about the character data for checkContent later
((SchemaValidator*) fValidator)->setDatatypeBuffer(toFill.getRawBuffer());
// call all active identity constraints
if (fMatcherStack->getMatcherCount())
fContent.append(toFill.getRawBuffer(), toFill.getLen());
if (fDocHandler)
fDocHandler->docCharacters(toFill.getRawBuffer(), toFill.getLen(), false);
}
else
{
fValidator->emitError(XMLValid::NoCharDataInCM);
((SchemaElementDecl *)topElem->fThisElement)->setValidity(PSVIDefs::INVALID);
}
}
}
else
{
// call all active identity constraints
if (fMatcherStack->getMatcherCount())
fContent.append(toSend.getRawBuffer(), toSend.getLen());
// Always assume its just char data if not validating
if (fDocHandler)
fDocHandler->docCharacters(toSend.getRawBuffer(), toSend.getLen(), false);
}
// Reset buffer
toSend.reset();
}
// This method is called with a key/value string pair that represents an
// xmlns="yyy" or xmlns:xxx="yyy" attribute. This method will update the
// current top of the element stack based on this data. We know that when
// we get here, that it is one of these forms, so we don't bother confirming
// it.
//
// But we have to ensure
// 1. xxx is not xmlns
// 2. if xxx is xml, then yyy must match XMLUni::fgXMLURIName, and vice versa
// 3. yyy is not XMLUni::fgXMLNSURIName
// 4. if xxx is not null, then yyy cannot be an empty string.
void SGXMLScanner::updateNSMap(const XMLCh* const attrName
, const XMLCh* const attrValue)
{
// We need a buffer to normalize the attribute value into
XMLBufBid bbNormal(&fBufMgr);
XMLBuffer& normalBuf = bbNormal.getBuffer();
// Normalize the value into the passed buffer. In this case, we don't
// care about the return value. An error was issued for the error, which
// is all we care about here.
normalizeAttRawValue(attrName, attrValue, normalBuf);
XMLCh* namespaceURI = normalBuf.getRawBuffer();
// We either have the default prefix (""), or we point it into the attr
// name parameter. Note that the xmlns is not the prefix we care about
// here. To us, the 'prefix' is really the local part of the attrName
// parameter.
//
// Check 1. xxx is not xmlns
// 2. if xxx is xml, then yyy must match XMLUni::fgXMLURIName, and vice versa
// 3. yyy is not XMLUni::fgXMLNSURIName
// 4. if xxx is not null, then yyy cannot be an empty string.
const XMLCh* prefPtr = XMLUni::fgZeroLenString;
const int colonOfs = XMLString::indexOf(attrName, chColon);
if (colonOfs != -1) {
prefPtr = &attrName[colonOfs + 1];
if (XMLString::equals(prefPtr, XMLUni::fgXMLNSString))
emitError(XMLErrs::NoUseOfxmlnsAsPrefix);
else if (XMLString::equals(prefPtr, XMLUni::fgXMLString)) {
if (!XMLString::equals(namespaceURI, XMLUni::fgXMLURIName))
emitError(XMLErrs::PrefixXMLNotMatchXMLURI);
}
if (!namespaceURI)
emitError(XMLErrs::NoEmptyStrNamespace, attrName);
else if(!*namespaceURI && fXMLVersion == XMLReader::XMLV1_0)
emitError(XMLErrs::NoEmptyStrNamespace, attrName);
}
if (XMLString::equals(namespaceURI, XMLUni::fgXMLNSURIName))
emitError(XMLErrs::NoUseOfxmlnsURI);
else if (XMLString::equals(namespaceURI, XMLUni::fgXMLURIName)) {
if (!XMLString::equals(prefPtr, XMLUni::fgXMLString))
emitError(XMLErrs::XMLURINotMatchXMLPrefix);
}
// Ok, we have to get the unique id for the attribute value, which is the
// URI that this value should be mapped to. The validator has the
// namespace string pool, so we ask him to find or add this new one. Then
// we ask the element stack to add this prefix to URI Id mapping.
fElemStack.addPrefix
(
prefPtr
, fURIStringPool->addOrFind(namespaceURI)
);
}
void SGXMLScanner::scanRawAttrListforNameSpaces(const RefVectorOf<KVStringPair>* theRawAttrList, int attCount)
{
// Make an initial pass through the list and find any xmlns attributes or
// schema attributes.
// When we find one, send it off to be used to update the element stack's
// namespace mappings.
int index = 0;
for (index = 0; index < attCount; index++)
{
// each attribute has the prefix:suffix="value"
const KVStringPair* curPair = fRawAttrList->elementAt(index);
const XMLCh* rawPtr = curPair->getKey();
// If either the key begins with "xmlns:" or its just plain
// "xmlns", then use it to update the map.
if (!XMLString::compareNString(rawPtr, XMLUni::fgXMLNSColonString, 6)
|| XMLString::equals(rawPtr, XMLUni::fgXMLNSString))
{
const XMLCh* valuePtr = curPair->getValue();
updateNSMap(rawPtr, valuePtr);
// if the schema URI is seen in the the valuePtr, set the boolean seeXsi
if (XMLString::equals(valuePtr, SchemaSymbols::fgURI_XSI)) {
fSeeXsi = true;
}
}
}
// walk through the list again to deal with "xsi:...."
if (fSeeXsi)
{
// Schema Xsi Type yyyy (e.g. xsi:type="yyyyy")
XMLBufBid bbXsi(&fBufMgr);
XMLBuffer& fXsiType = bbXsi.getBuffer();
QName attName(fMemoryManager);
for (index = 0; index < attCount; index++)
{
// each attribute has the prefix:suffix="value"
const KVStringPair* curPair = fRawAttrList->elementAt(index);
const XMLCh* rawPtr = curPair->getKey();
attName.setName(rawPtr, fEmptyNamespaceId);
const XMLCh* prefPtr = attName.getPrefix();
// if schema URI has been seen, scan for the schema location and uri
// and resolve the schema grammar; or scan for schema type
if (resolvePrefix(prefPtr, ElemStack::Mode_Attribute) == fSchemaNamespaceId) {
const XMLCh* valuePtr = curPair->getValue();
const XMLCh* suffPtr = attName.getLocalPart();
if (XMLString::equals(suffPtr, SchemaSymbols::fgXSI_SCHEMALOCACTION))
parseSchemaLocation(valuePtr);
else if (XMLString::equals(suffPtr, SchemaSymbols::fgXSI_NONAMESPACESCHEMALOCACTION))
resolveSchemaGrammar(valuePtr, XMLUni::fgZeroLenString);
if (XMLString::equals(suffPtr, SchemaSymbols::fgXSI_TYPE)) {
fXsiType.set(valuePtr);
}
else if (XMLString::equals(suffPtr, SchemaSymbols::fgATT_NILL)
&& fValidator && fValidator->handlesSchema()
&& XMLString::equals(valuePtr, SchemaSymbols::fgATTVAL_TRUE)) {
((SchemaValidator*)fValidator)->setNillable(true);
}
}
}
if (fValidator && fValidator->handlesSchema()) {
if (!fXsiType.isEmpty()) {
int colonPos = -1;
unsigned int uriId = resolveQName (
fXsiType.getRawBuffer()
, fPrefixBuf
, ElemStack::Mode_Element
, colonPos
);
((SchemaValidator*)fValidator)->setXsiType(fPrefixBuf.getRawBuffer(), fXsiType.getRawBuffer() + colonPos + 1, uriId);
}
}
}
}
void SGXMLScanner::parseSchemaLocation(const XMLCh* const schemaLocationStr)
{
BaseRefVectorOf<XMLCh>* schemaLocation = XMLString::tokenizeString(schemaLocationStr);
unsigned int size = schemaLocation->size();
if (size % 2 != 0 ) {
emitError(XMLErrs::BadSchemaLocation);
} else {
for(unsigned int i=0; i<size; i=i+2) {
resolveSchemaGrammar(schemaLocation->elementAt(i+1), schemaLocation->elementAt(i));
}
}
delete schemaLocation;
}
void SGXMLScanner::resolveSchemaGrammar(const XMLCh* const loc, const XMLCh* const uri) {
Grammar* grammar = fGrammarResolver->getGrammar(uri);
if (!grammar || grammar->getGrammarType() == Grammar::DTDGrammarType) {
XSDDOMParser parser(0, fMemoryManager);
parser.setValidationScheme(XercesDOMParser::Val_Never);
parser.setDoNamespaces(true);
parser.setUserEntityHandler(fEntityHandler);
parser.setUserErrorReporter(fErrorReporter);
// Create a buffer for expanding the system id
XMLBufBid bbSys(&fBufMgr);
XMLBuffer& expSysId = bbSys.getBuffer();
XMLBuffer& normalizedSysId = bbSys.getBuffer();
normalizeURI(loc, normalizedSysId);
// Allow the entity handler to expand the system id if they choose
// to do so.
InputSource* srcToFill = 0;
const XMLCh* normalizedURI = normalizedSysId.getRawBuffer();
if (fEntityHandler)
{
if (!fEntityHandler->expandSystemId(normalizedURI, expSysId))
expSysId.set(normalizedURI);
srcToFill = fEntityHandler->resolveEntity( XMLUni::fgZeroLenString
, expSysId.getRawBuffer());
}
else
{
expSysId.set(normalizedURI);
}
// If they didn't create a source via the entity handler, then we
// have to create one on our own.
if (!srcToFill)
{
ReaderMgr::LastExtEntityInfo lastInfo;
fReaderMgr.getLastExtEntityInfo(lastInfo);
try
{
XMLURL urlTmp(lastInfo.systemId, expSysId.getRawBuffer());
if (urlTmp.isRelative())
{
ThrowXML
(
MalformedURLException
, XMLExcepts::URL_NoProtocolPresent
);
}
else {
if (fStandardUriConformant && urlTmp.hasInvalidChar())
ThrowXML(MalformedURLException, XMLExcepts::URL_MalformedURL);
srcToFill = new (fMemoryManager) URLInputSource(urlTmp, fMemoryManager);
}
}
catch(const MalformedURLException& e)
{
// Its not a URL, so lets assume its a local file name if non-standard uri is allowed
if (!fStandardUriConformant)
srcToFill = new (fMemoryManager) LocalFileInputSource
(
lastInfo.systemId
, expSysId.getRawBuffer()
, fMemoryManager
);
else
throw e;
}
}
// Put a janitor on the input source
Janitor<InputSource> janSrc(srcToFill);
// Should just issue warning if the schema is not found
const bool flag = srcToFill->getIssueFatalErrorIfNotFound();
srcToFill->setIssueFatalErrorIfNotFound(false);
parser.parse(*srcToFill);
// Reset the InputSource
srcToFill->setIssueFatalErrorIfNotFound(flag);
if (parser.getSawFatal() && fExitOnFirstFatal)
emitError(XMLErrs::SchemaScanFatalError);
DOMDocument* document = parser.getDocument(); //Our Grammar
if (document != 0) {
DOMElement* root = document->getDocumentElement();// This is what we pass to TraverserSchema
if (root != 0)
{
const XMLCh* newUri = root->getAttribute(SchemaSymbols::fgATT_TARGETNAMESPACE);
if (!XMLString::equals(newUri, uri)) {
if (fValidate || fValScheme == Val_Auto) {
fValidator->emitError(XMLValid::WrongTargetNamespace, loc, uri);
}
grammar = fGrammarResolver->getGrammar(newUri);
}
if (!grammar || grammar->getGrammarType() == Grammar::DTDGrammarType) {
// Since we have seen a grammar, set our validation flag
// at this point if the validation scheme is auto
if (fValScheme == Val_Auto && !fValidate) {
fValidate = true;
fElemStack.setValidationFlag(fValidate);
}
grammar = new (fMemoryManager) SchemaGrammar(fMemoryManager);
TraverseSchema traverseSchema
(
root
, fURIStringPool
, (SchemaGrammar*) grammar
, fGrammarResolver
, this
, srcToFill->getSystemId()
, fEntityHandler
, fErrorReporter
, fMemoryManager
);
if (fGrammarType == Grammar::DTDGrammarType) {
fGrammar = grammar;
fGrammarType = Grammar::SchemaGrammarType;
fValidator->setGrammar(fGrammar);
}
if (fValidate) {
// validate the Schema scan so far
fValidator->preContentValidation(false);
}
}
}
}
}
else {
// Since we have seen a grammar, set our validation flag
// at this point if the validation scheme is auto
if (fValScheme == Val_Auto && !fValidate) {
fValidate = true;
fElemStack.setValidationFlag(fValidate);
}
// we have seen a schema, so set up the fValidator as fSchemaValidator
if (fGrammarType == Grammar::DTDGrammarType) {
fGrammar = grammar;
fGrammarType = Grammar::SchemaGrammarType;
fValidator->setGrammar(fGrammar);
}
}
}
InputSource* SGXMLScanner::resolveSystemId(const XMLCh* const sysId)
{
// Create a buffer for expanding the system id
XMLBufBid bbSys(&fBufMgr);
XMLBuffer& expSysId = bbSys.getBuffer();
// Allow the entity handler to expand the system id if they choose
// to do so.
InputSource* srcToFill = 0;
if (fEntityHandler)
{
if (!fEntityHandler->expandSystemId(sysId, expSysId))
expSysId.set(sysId);
srcToFill = fEntityHandler->resolveEntity( XMLUni::fgZeroLenString
, expSysId.getRawBuffer());
}
else
{
expSysId.set(sysId);
}
// If they didn't create a source via the entity handler, then we
// have to create one on our own.
if (!srcToFill)
{
ReaderMgr::LastExtEntityInfo lastInfo;
fReaderMgr.getLastExtEntityInfo(lastInfo);
try
{
XMLURL urlTmp(lastInfo.systemId, expSysId.getRawBuffer());
if (urlTmp.isRelative())
{
ThrowXML
(
MalformedURLException
, XMLExcepts::URL_NoProtocolPresent
);
}
else {
if (fStandardUriConformant && urlTmp.hasInvalidChar())
ThrowXML(MalformedURLException, XMLExcepts::URL_MalformedURL);
srcToFill = new (fMemoryManager) URLInputSource(urlTmp, fMemoryManager);
}
}
catch(const MalformedURLException& e)
{
// Its not a URL, so lets assume its a local file name if non-standard uri is allowed
if (!fStandardUriConformant)
srcToFill = new (fMemoryManager) LocalFileInputSource
(
lastInfo.systemId
, expSysId.getRawBuffer()
, fMemoryManager
);
else
throw e;
}
}
return srcToFill;
}
// ---------------------------------------------------------------------------
// SGXMLScanner: Private grammar preparsing methods
// ---------------------------------------------------------------------------
Grammar* SGXMLScanner::loadXMLSchemaGrammar(const InputSource& src,
const bool toCache)
{
// Reset the validators
fSchemaValidator->reset();
fSchemaValidator->setErrorReporter(fErrorReporter);
fSchemaValidator->setExitOnFirstFatal(fExitOnFirstFatal);
fSchemaValidator->setGrammarResolver(fGrammarResolver);
if (fValidatorFromUser)
fValidator->reset();
XSDDOMParser parser(0, fMemoryManager);
parser.setValidationScheme(XercesDOMParser::Val_Never);
parser.setDoNamespaces(true);
parser.setUserEntityHandler(fEntityHandler);
parser.setUserErrorReporter(fErrorReporter);
// Should just issue warning if the schema is not found
const bool flag = src.getIssueFatalErrorIfNotFound();
((InputSource&) src).setIssueFatalErrorIfNotFound(false);
parser.parse(src);
// Reset the InputSource
((InputSource&) src).setIssueFatalErrorIfNotFound(flag);
if (parser.getSawFatal() && fExitOnFirstFatal)
emitError(XMLErrs::SchemaScanFatalError);
DOMDocument* document = parser.getDocument(); //Our Grammar
if (document != 0) {
DOMElement* root = document->getDocumentElement();// This is what we pass to TraverserSchema
if (root != 0)
{
SchemaGrammar* grammar = new (fMemoryManager) SchemaGrammar(fMemoryManager);
TraverseSchema traverseSchema
(
root
, fURIStringPool
, (SchemaGrammar*) grammar
, fGrammarResolver
, this
, src.getSystemId()
, fEntityHandler
, fErrorReporter
, fMemoryManager
);
if (fValidate) {
// validate the Schema scan so far
fValidator->setGrammar(grammar);
fValidator->preContentValidation(false, true);
}
if (toCache) {
fGrammarResolver->cacheGrammars();
}
return grammar;
}
}
return 0;
}
// ---------------------------------------------------------------------------
// SGXMLScanner: Private parsing methods
// ---------------------------------------------------------------------------
// This method is called to do a raw scan of an attribute value. It does not
// do normalization (since we don't know their types yet.) It just scans the
// value and does entity expansion.
//
// End of entity's must be dealt with here. During DTD scan, they can come
// from external entities. During content, they can come from any entity.
// We just eat the end of entity and continue with our scan until we come
// to the closing quote. If an unterminated value causes us to go through
// subsequent entities, that will cause errors back in the calling code,
// but there's little we can do about it here.
bool SGXMLScanner::basicAttrValueScan(const XMLCh* const attrName, XMLBuffer& toFill)
{
// Reset the target buffer
toFill.reset();
// Get the next char which must be a single or double quote
XMLCh quoteCh;
if (!fReaderMgr.skipIfQuote(quoteCh))
return false;
// We have to get the current reader because we have to ignore closing
// quotes until we hit the same reader again.
const unsigned int curReader = fReaderMgr.getCurrentReaderNum();
// Loop until we get the attribute value. Note that we use a double
// loop here to avoid the setup/teardown overhead of the exception
// handler on every round.
XMLCh nextCh;
XMLCh secondCh = 0;
bool gotLeadingSurrogate = false;
bool escaped;
while (true)
{
try
{
while(true)
{
nextCh = fReaderMgr.getNextChar();
if (!nextCh)
ThrowXML(UnexpectedEOFException, XMLExcepts::Gen_UnexpectedEOF);
// Check for our ending quote. It has to be in the same entity
// as where we started. Quotes in nested entities are ignored.
if (nextCh == quoteCh)
{
if (curReader == fReaderMgr.getCurrentReaderNum())
return true;
// Watch for spillover into a previous entity
if (curReader > fReaderMgr.getCurrentReaderNum())
{
emitError(XMLErrs::PartialMarkupInEntity);
return false;
}
}
// Check for an entity ref . We ignore the empty flag in
// this one.
escaped = false;
if (nextCh == chAmpersand)
{
// If it was not returned directly, then jump back up
if (scanEntityRef(true, nextCh, secondCh, escaped) != EntityExp_Returned)
{
gotLeadingSurrogate = false;
continue;
}
}
else if ((nextCh >= 0xD800) && (nextCh <= 0xDBFF))
{
// Deal with surrogate pairs
// Its a leading surrogate. If we already got one, then
// issue an error, else set leading flag to make sure that
// we look for a trailing next time.
if (gotLeadingSurrogate)
{
emitError(XMLErrs::Expected2ndSurrogateChar);
}
else
gotLeadingSurrogate = true;
}
else
{
// If its a trailing surrogate, make sure that we are
// prepared for that. Else, its just a regular char so make
// sure that we were not expected a trailing surrogate.
if ((nextCh >= 0xDC00) && (nextCh <= 0xDFFF))
{
// Its trailing, so make sure we were expecting it
if (!gotLeadingSurrogate)
emitError(XMLErrs::Unexpected2ndSurrogateChar);
}
else
{
// Its just a char, so make sure we were not expecting a
// trailing surrogate.
if (gotLeadingSurrogate) {
emitError(XMLErrs::Expected2ndSurrogateChar);
}
// Its got to at least be a valid XML character
else if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
{
XMLCh tmpBuf[9];
XMLString::binToText
(
nextCh
, tmpBuf
, 8
, 16
);
emitError(XMLErrs::InvalidCharacterInAttrValue, attrName, tmpBuf);
}
}
gotLeadingSurrogate = false;
}
// If it was escaped, then put in a 0xFFFF value. This will
// be used later during validation and normalization of the
// value to know that the following character was via an
// escape char.
if (escaped)
toFill.append(0xFFFF);
// Else add it to the buffer
toFill.append(nextCh);
if (secondCh)
toFill.append(secondCh);
}
}
catch(const EndOfEntityException&)
{
// Just eat it and continue.
gotLeadingSurrogate = false;
escaped = false;
}
}
return true;
}
// This method scans a CDATA section. It collects the character into one
// of the temp buffers and calls the document handler, if any, with the
// characters. It assumes that the <![CDATA string has been scanned before
// this call.
void SGXMLScanner::scanCDSection()
{
// This is the CDATA section opening sequence, minus the '<' character.
// We use this to watch for nested CDATA sections, which are illegal.
static const XMLCh CDataPrefix[] =
{
chBang, chOpenSquare, chLatin_C, chLatin_D, chLatin_A
, chLatin_T, chLatin_A, chOpenSquare, chNull
};
static const XMLCh CDataClose[] =
{
chCloseSquare, chCloseAngle, chNull
};
// The next character should be the opening square bracket. If not
// issue an error, but then try to recover by skipping any whitespace
// and checking again.
if (!fReaderMgr.skippedChar(chOpenSquare))
{
emitError(XMLErrs::ExpectedOpenSquareBracket);
fReaderMgr.skipPastSpaces();
// If we still don't find it, then give up, else keep going
if (!fReaderMgr.skippedChar(chOpenSquare))
return;
}
// Get a buffer for this
XMLBufBid bbCData(&fBufMgr);
// We just scan forward until we hit the end of CDATA section sequence.
// CDATA is effectively a big escape mechanism so we don't treat markup
// characters specially here.
bool emittedError = false;
bool gotLeadingSurrogate = false;
// Get the character data opts for the current element
const ElemStack::StackElem* topElem = fElemStack.topElement();
XMLElementDecl::CharDataOpts charOpts = topElem->fThisElement->getCharDataOpts();
while (true)
{
const XMLCh nextCh = fReaderMgr.getNextChar();
// Watch for unexpected end of file
if (!nextCh)
{
emitError(XMLErrs::UnterminatedCDATASection);
ThrowXML(UnexpectedEOFException, XMLExcepts::Gen_UnexpectedEOF);
}
if (fValidate && fStandalone && (fReaderMgr.getCurrentReader()->isWhitespace(nextCh)))
{
// This document is standalone; this ignorable CDATA whitespace is forbidden.
// XML 1.0, Section 2.9
// And see if the current element is a 'Children' style content model
if (topElem->fThisElement->isExternal()) {
if (charOpts == XMLElementDecl::SpacesOk) // Element Content
{
// Error - standalone should have a value of "no" as whitespace detected in an
// element type with element content whose element declaration was external
fValidator->emitError(XMLValid::NoWSForStandalone);
((SchemaElementDecl *)(topElem->fThisElement))->setValidity(PSVIDefs::INVALID);
}
}
}
// If this is a close square bracket it could be our closing
// sequence.
if (nextCh == chCloseSquare && fReaderMgr.skippedString(CDataClose))
{
// make sure we were not expecting a trailing surrogate.
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
if (fValidate) {
if (fNormalizeData) {
// normalize the character according to schema whitespace facet
XMLBufBid bbtemp(&fBufMgr);
XMLBuffer& tempBuf = bbtemp.getBuffer();
DatatypeValidator* tempDV = ((SchemaElementDecl*) topElem->fThisElement)->getDatatypeValidator();
((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, bbCData.getRawBuffer(), tempBuf);
bbCData.set(tempBuf.getRawBuffer());
}
// tell the schema validation about the character data for checkContent later
((SchemaValidator*) fValidator)->setDatatypeBuffer(bbCData.getRawBuffer());
if (charOpts != XMLElementDecl::AllCharData)
{
// They definitely cannot handle any type of char data
fValidator->emitError(XMLValid::NoCharDataInCM);
((SchemaElementDecl *)topElem->fThisElement)->setValidity(PSVIDefs::INVALID);
}
}
// call all active identity constraints
if (fMatcherStack->getMatcherCount())
fContent.append(bbCData.getRawBuffer(), bbCData.getLen());
// If we have a doc handler, call it
if (fDocHandler)
{
fDocHandler->docCharacters
(
bbCData.getRawBuffer()
, bbCData.getLen()
, true
);
}
// And we are done
break;
}
// Make sure its a valid character. But if we've emitted an error
// already, don't bother with the overhead since we've already told
// them about it.
if (!emittedError)
{
// Deal with surrogate pairs
if ((nextCh >= 0xD800) && (nextCh <= 0xDBFF))
{
// Its a leading surrogate. If we already got one, then
// issue an error, else set leading flag to make sure that
// we look for a trailing next time.
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
else
gotLeadingSurrogate = true;
}
else
{
// If its a trailing surrogate, make sure that we are
// prepared for that. Else, its just a regular char so make
// sure that we were not expected a trailing surrogate.
if ((nextCh >= 0xDC00) && (nextCh <= 0xDFFF))
{
// Its trailing, so make sure we were expecting it
if (!gotLeadingSurrogate)
emitError(XMLErrs::Unexpected2ndSurrogateChar);
}
else
{
// Its just a char, so make sure we were not expecting a
// trailing surrogate.
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
// Its got to at least be a valid XML character
else if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
{
XMLCh tmpBuf[9];
XMLString::binToText
(
nextCh
, tmpBuf
, 8
, 16
);
emitError(XMLErrs::InvalidCharacter, tmpBuf);
emittedError = true;
}
}
gotLeadingSurrogate = false;
}
}
// Add it to the buffer
bbCData.append(nextCh);
}
}
void SGXMLScanner::scanCharData(XMLBuffer& toUse)
{
// We have to watch for the stupid ]]> sequence, which is illegal in
// character data. So this is a little state machine that handles that.
enum States
{
State_Waiting
, State_GotOne
, State_GotTwo
};
// Reset the buffer before we start
toUse.reset();
// Turn on the 'throw at end' flag of the reader manager
ThrowEOEJanitor jan(&fReaderMgr, true);
// In order to be more efficient we have to use kind of a deeply nested
// set of blocks here. The outer block puts on a try and catches end of
// entity exceptions. The inner loop is the per-character loop. If we
// put the try inside the inner loop, it would work but would require
// the exception handling code setup/teardown code to be invoked for
// each character.
XMLCh nextCh;
XMLCh secondCh = 0;
States curState = State_Waiting;
bool escaped = false;
bool gotLeadingSurrogate = false;
bool notDone = true;
while (notDone)
{
try
{
while (true)
{
// Eat through as many plain content characters as possible without
// needing special handling. Moving most content characters here,
// in this one call, rather than running the overall loop once
// per content character, is a speed optimization.
if (curState == State_Waiting && !gotLeadingSurrogate)
{
fReaderMgr.movePlainContentChars(toUse);
}
// Try to get another char from the source
// The code from here on down covers all contengencies,
if (!fReaderMgr.getNextCharIfNot(chOpenAngle, nextCh))
{
// If we were waiting for a trailing surrogate, its an error
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
notDone = false;
break;
}
// Watch for a reference. Note that the escapement mechanism
// is ignored in this content.
escaped = false;
if (nextCh == chAmpersand)
{
sendCharData(toUse);
// Turn off the throwing at the end of entity during this
ThrowEOEJanitor jan(&fReaderMgr, false);
if (scanEntityRef(false, nextCh, secondCh, escaped) != EntityExp_Returned)
{
gotLeadingSurrogate = false;
continue;
}
}
else if ((nextCh >= 0xD800) && (nextCh <= 0xDBFF))
{
// Deal with surrogate pairs
// Its a leading surrogate. If we already got one, then
// issue an error, else set leading flag to make sure that
// we look for a trailing next time.
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
else
gotLeadingSurrogate = true;
}
else
{
// If its a trailing surrogate, make sure that we are
// prepared for that. Else, its just a regular char so make
// sure that we were not expected a trailing surrogate.
if ((nextCh >= 0xDC00) && (nextCh <= 0xDFFF))
{
// Its trailing, so make sure we were expecting it
if (!gotLeadingSurrogate)
emitError(XMLErrs::Unexpected2ndSurrogateChar);
}
else
{
// Its just a char, so make sure we were not expecting a
// trailing surrogate.
if (gotLeadingSurrogate)
emitError(XMLErrs::Expected2ndSurrogateChar);
// Make sure the returned char is a valid XML char
if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
{
XMLCh tmpBuf[9];
XMLString::binToText
(
nextCh
, tmpBuf
, 8
, 16
);
emitError(XMLErrs::InvalidCharacter, tmpBuf);
}
}
gotLeadingSurrogate = false;
}
// Keep the state machine up to date
if (!escaped)
{
if (nextCh == chCloseSquare)
{
if (curState == State_Waiting)
curState = State_GotOne;
else if (curState == State_GotOne)
curState = State_GotTwo;
}
else if (nextCh == chCloseAngle)
{
if (curState == State_GotTwo)
emitError(XMLErrs::BadSequenceInCharData);
curState = State_Waiting;
}
else
{
curState = State_Waiting;
}
}
else
{
curState = State_Waiting;
}
// Add this char to the buffer
toUse.append(nextCh);
if (secondCh)
toUse.append(secondCh);
}
}
catch(const EndOfEntityException& toCatch)
{
// Some entity ended, so we have to send any accumulated
// chars and send an end of entity event.
sendCharData(toUse);
gotLeadingSurrogate = false;
if (fDocHandler)
fDocHandler->endEntityReference(toCatch.getEntity());
}
}
// Check the validity constraints as per XML 1.0 Section 2.9
if (fValidate && fStandalone)
{
// See if the text contains whitespace
// Get the raw data we need for the callback
const XMLCh* rawBuf = toUse.getRawBuffer();
const unsigned int len = toUse.getLen();
const bool isSpaces = fReaderMgr.getCurrentReader()->containsWhiteSpace(rawBuf, len);
if (isSpaces)
{
// And see if the current element is a 'Children' style content model
const ElemStack::StackElem* topElem = fElemStack.topElement();
if (topElem->fThisElement->isExternal()) {
// Get the character data opts for the current element
XMLElementDecl::CharDataOpts charOpts = topElem->fThisElement->getCharDataOpts();
if (charOpts == XMLElementDecl::SpacesOk) // => Element Content
{
// Error - standalone should have a value of "no" as whitespace detected in an
// element type with element content whose element declaration was external
//
fValidator->emitError(XMLValid::NoWSForStandalone);
((SchemaElementDecl *)fElemStack.topElement()->fThisElement)->setValidity(PSVIDefs::INVALID);
}
}
}
}
// Send any char data that we accumulated into the buffer
sendCharData(toUse);
}
// This method will scan a general/character entity ref. It will either
// expand a char ref and return it directly, or push a reader for a general
// entity.
//
// The return value indicates whether the char parameters hold the value
// or whether the value was pushed as a reader, or that it failed.
//
// The escaped flag tells the caller whether the returned parameter resulted
// from a character reference, which escapes the character in some cases. It
// only makes any difference if the return value indicates the value was
// returned directly.
SGXMLScanner::EntityExpRes
SGXMLScanner::scanEntityRef( const bool inAttVal
, XMLCh& firstCh
, XMLCh& secondCh
, bool& escaped)
{
// Assume no escape
secondCh = 0;
escaped = false;
// We have to insure that its all in one entity
const unsigned int curReader = fReaderMgr.getCurrentReaderNum();
// If the next char is a pound, then its a character reference and we
// need to expand it always.
if (fReaderMgr.skippedChar(chPound))
{
// Its a character reference, so scan it and get back the numeric
// value it represents.
if (!scanCharRef(firstCh, secondCh))
return EntityExp_Failed;
escaped = true;
if (curReader != fReaderMgr.getCurrentReaderNum())
emitError(XMLErrs::PartialMarkupInEntity);
return EntityExp_Returned;
}
// Expand it since its a normal entity ref
XMLBufBid bbName(&fBufMgr);
if (!fReaderMgr.getName(bbName.getBuffer()))
{
emitError(XMLErrs::ExpectedEntityRefName);
return EntityExp_Failed;
}
// Next char must be a semi-colon. But if its not, just emit
// an error and try to continue.
if (!fReaderMgr.skippedChar(chSemiColon))
emitError(XMLErrs::UnterminatedEntityRef, bbName.getRawBuffer());
// Make sure we ended up on the same entity reader as the & char
if (curReader != fReaderMgr.getCurrentReaderNum())
emitError(XMLErrs::PartialMarkupInEntity);
// Look up the name in the general entity pool
// If it does not exist, then obviously an error
if (!fEntityTable->containsKey(bbName.getRawBuffer()))
{
// XML 1.0 Section 4.1
// Well-formedness Constraint for entity not found:
// In a document without any DTD, a document with only an internal DTD subset which contains no parameter entity references,
// or a document with "standalone='yes'", for an entity reference that does not occur within the external subset
// or a parameter entity
if (fStandalone || fHasNoDTD)
emitError(XMLErrs::EntityNotFound, bbName.getRawBuffer());
return EntityExp_Failed;
}
// here's where we need to check if there's a SecurityManager,
// how many entity references we've had
if(fSecurityManager != 0 && ++fEntityExpansionCount > fEntityExpansionLimit) {
XMLCh expLimStr[16];
XMLString::binToText(fEntityExpansionLimit, expLimStr, 15, 10);
emitError
(
XMLErrs::EntityExpansionLimitExceeded
, expLimStr
);
// there seems nothing better to be done than to reset the entity expansion limit
fEntityExpansionCount = 0;
}
firstCh = fEntityTable->get(bbName.getRawBuffer());
escaped = true;
return EntityExp_Returned;
}
bool SGXMLScanner::switchGrammar(const XMLCh* const newGrammarNameSpace)
{
Grammar* tempGrammar = fGrammarResolver->getGrammar(newGrammarNameSpace);
if (!tempGrammar) {
tempGrammar = fSchemaGrammar;
}
if (!tempGrammar)
return false;
else {
fGrammar = tempGrammar;
fGrammarType = fGrammar->getGrammarType();
if (fGrammarType == Grammar::DTDGrammarType) {
ThrowXML(RuntimeException, XMLExcepts::Gen_NoDTDValidator);
}
fValidator->setGrammar(fGrammar);
return true;
}
}
// check if we should skip or lax the validation of the element
// if skip - no validation
// if lax - validate only if the element if found
bool SGXMLScanner::laxElementValidation(QName* element, ContentLeafNameTypeVector* cv,
const XMLContentModel* const cm,
const unsigned int parentElemDepth)
{
bool skipThisOne = false;
bool laxThisOne = false;
unsigned int elementURI = element->getURI();
unsigned int currState = fElemState[parentElemDepth];
if (currState == XMLContentModel::gInvalidTrans) {
return laxThisOne;
}
SubstitutionGroupComparator comparator(fGrammarResolver, fURIStringPool);
if (cv) {
unsigned int i = 0;
unsigned int leafCount = cv->getLeafCount();
for (; i < leafCount; i++) {
QName* fElemMap = cv->getLeafNameAt(i);
unsigned int uri = fElemMap->getURI();
unsigned int nextState;
bool anyEncountered = false;
ContentSpecNode::NodeTypes type = cv->getLeafTypeAt(i);
if (type == ContentSpecNode::Leaf) {
if (((uri == elementURI)
&& XMLString::equals(fElemMap->getLocalPart(), element->getLocalPart()))
|| comparator.isEquivalentTo(element, fElemMap)) {
nextState = cm->getNextState(currState, i);
if (nextState != XMLContentModel::gInvalidTrans) {
fElemState[parentElemDepth] = nextState;
break;
}
}
} else if ((type & 0x0f) == ContentSpecNode::Any) {
anyEncountered = true;
}
else if ((type & 0x0f) == ContentSpecNode::Any_Other) {
if (uri != elementURI) {
anyEncountered = true;
}
}
else if ((type & 0x0f) == ContentSpecNode::Any_NS) {
if (uri == elementURI) {
anyEncountered = true;
}
}
if (anyEncountered) {
nextState = cm->getNextState(currState, i);
if (nextState != XMLContentModel::gInvalidTrans) {
fElemState[parentElemDepth] = nextState;
if (type == ContentSpecNode::Any_Skip ||
type == ContentSpecNode::Any_NS_Skip ||
type == ContentSpecNode::Any_Other_Skip) {
skipThisOne = true;
}
else if (type == ContentSpecNode::Any_Lax ||
type == ContentSpecNode::Any_NS_Lax ||
type == ContentSpecNode::Any_Other_Lax) {
laxThisOne = true;
}
break;
}
}
} // for
if (i == leafCount) { // no match
fElemState[parentElemDepth] = XMLContentModel::gInvalidTrans;
return laxThisOne;
}
} // if
if (skipThisOne) {
fValidate = false;
fElemStack.setValidationFlag(fValidate);
}
return laxThisOne;
}
// check if there is an AnyAttribute, and if so, see if we should lax or skip
// if skip - no validation
// if lax - validate only if the attribute if found
bool SGXMLScanner::anyAttributeValidation(SchemaAttDef* attWildCard, unsigned int uriId, bool& skipThisOne, bool& laxThisOne)
{
XMLAttDef::AttTypes wildCardType = attWildCard->getType();
bool anyEncountered = false;
skipThisOne = false;
laxThisOne = false;
if (wildCardType == XMLAttDef::Any_Any)
anyEncountered = true;
else if (wildCardType == XMLAttDef::Any_Other) {
if (attWildCard->getAttName()->getURI() != uriId
&& uriId != fEmptyNamespaceId)
anyEncountered = true;
}
else if (wildCardType == XMLAttDef::Any_List) {
ValueVectorOf<unsigned int>* nameURIList = attWildCard->getNamespaceList();
unsigned int listSize = (nameURIList) ? nameURIList->size() : 0;
if (listSize) {
for (unsigned int i=0; i < listSize; i++) {
if (nameURIList->elementAt(i) == uriId)
anyEncountered = true;
}
}
}
if (anyEncountered) {
XMLAttDef::DefAttTypes defType = attWildCard->getDefaultType();
if (defType == XMLAttDef::ProcessContents_Skip) {
// attribute should just be bypassed,
skipThisOne = true;
attWildCard->setValidationAttempted(PSVIDefs::NONE);
}
else if (defType == XMLAttDef::ProcessContents_Lax) {
laxThisOne = true;
}
}
return anyEncountered;
}
void SGXMLScanner::normalizeURI(const XMLCh* const systemURI,
XMLBuffer& normalizedURI)
{
const XMLCh* pszSrc = systemURI;
normalizedURI.reset();
while (*pszSrc) {
if ((*(pszSrc) == chPercent)
&& (*(pszSrc+1) == chDigit_2)
&& (*(pszSrc+2) == chDigit_0))
{
pszSrc += 3;
normalizedURI.append(chSpace);
}
else if (*pszSrc == 0xFFFF) { //escaped character
pszSrc++;
}
else {
normalizedURI.append(*pszSrc);
pszSrc++;
}
}
}
XERCES_CPP_NAMESPACE_END