diff --git a/doc/program-sax2.xml b/doc/program-sax2.xml
index 7311281f3d786de32b525f128491fe50db70f923..5f9359ac8cf4fd03cfb926354d62da34c8e3f980 100644
--- a/doc/program-sax2.xml
+++ b/doc/program-sax2.xml
@@ -207,6 +207,15 @@ MySAX2Handler::fatalError(const SAXParseException& exception)
<p/>
+ <table>
+ <tr><td colspan="2"><em>http://apache.org/xml/features/nonvalidating/load-external-dtd</em></td></tr>
+ <tr><td><em>true:</em></td><td> Load the External DTD (default). </td></tr>
+ <tr><td><em>false:</em></td><td> Ignore the external DTD completely. This feature is
+ ignored and DTD is always loaded when validation is on.</td></tr>
+ </table>
+
+ <p/>
+
<table>
<tr><td colspan="2"><em>http://apache.org/xml/features/validation/reuse-grammar</em></td></tr>
<tr><td><em>true:</em></td><td> The parser will reuse grammar information from previous parses in subsequent parses. </td></tr>
diff --git a/src/xercesc/internal/XMLScanner.cpp b/src/xercesc/internal/XMLScanner.cpp
index 3b0ebbd3baf62f7903fa44b7185143658144aa04..d0354c55b6727e6a0a82833d2e4e73d938d71074 100644
--- a/src/xercesc/internal/XMLScanner.cpp
+++ b/src/xercesc/internal/XMLScanner.cpp
@@ -221,6 +221,7 @@ XMLScanner::XMLScanner(XMLValidator* const valToAdopt) :
, fRootElemName(0)
, fExternalSchemaLocation(0)
, fExternalNoNamespaceSchemaLocation(0)
+ , fLoadExternalDTD(true)
{
commonInit();
@@ -284,6 +285,7 @@ XMLScanner::XMLScanner( XMLDocumentHandler* const docHandler
, fRootElemName(0)
, fExternalSchemaLocation(0)
, fExternalNoNamespaceSchemaLocation(0)
+ , fLoadExternalDTD(true)
{
commonInit();
@@ -2215,28 +2217,7 @@ void XMLScanner::scanProlog()
}
else if (fReaderMgr.skippedString(XMLUni::fgDocTypeString))
{
- if (!fReuseGrammar && fValidatorFromUser && !fValidator->handlesDTD())
- {
- ThrowXML(RuntimeException, XMLExcepts::Gen_NoDTDValidator);
- }
-
- //
- // We have a doc type. So, create a DTDScanner and
- // switch the Grammar to the emptyNamespace one.
- //
-
- if (!switchGrammar(XMLUni::fgZeroLenString) && fValidate)
- {
- fValidator->emitError
- (
- XMLValid::GrammarNotFound
- , XMLUni::fgZeroLenString
- );
- }
-
- DTDScanner fDTDScanner((DTDGrammar*)fGrammar, fEntityDeclPool, fDocTypeHandler);
- fDTDScanner.setScannerInfo(this, &fReaderMgr, &fBufMgr);
- fDTDScanner.scanDocTypeDecl(fReuseGrammar);
+ scanDocTypeDecl();
// if reusing grammar, this has been validated already in first scan
// skip for performance
@@ -2297,6 +2278,302 @@ void XMLScanner::scanProlog()
}
}
+//
+// This method handles the high level logic of scanning the DOCType
+// declaration. This calls the DTDScanner and kicks off both the scanning of
+// the internal subset and the scanning of the external subset, if any.
+//
+// When we get here the '<!DOCTYPE' part has already been scanned, which is
+// what told us that we had a doc type decl to parse.
+//
+
+void XMLScanner::scanDocTypeDecl()
+{
+ if (!fReuseGrammar && fValidatorFromUser && !fValidator->handlesDTD())
+ {
+ ThrowXML(RuntimeException, XMLExcepts::Gen_NoDTDValidator);
+ }
+
+ //
+ // We have a doc type. So, create a DTDScanner and
+ // switch the Grammar to the emptyNamespace one.
+ //
+
+ if (!switchGrammar(XMLUni::fgZeroLenString) && fValidate)
+ {
+ fValidator->emitError
+ (
+ XMLValid::GrammarNotFound
+ , XMLUni::fgZeroLenString
+ );
+ }
+
+ DTDScanner dtdScanner((DTDGrammar*)fGrammar, fEntityDeclPool, fDocTypeHandler);
+ dtdScanner.setScannerInfo(this, &fReaderMgr, &fBufMgr);
+
+ if (fDocTypeHandler)
+ fDocTypeHandler->resetDocType();
+
+ // There must be some space after DOCTYPE
+ if (!fReaderMgr.skipPastSpaces())
+ {
+ emitError(XMLErrs::ExpectedWhitespace);
+
+ // Just skip the Doctype declaration and return
+ fReaderMgr.skipPastChar(chCloseAngle);
+ return;
+ }
+
+ // Get a buffer for the root element
+ XMLBufBid bbRootName(&fBufMgr);
+
+ //
+ // Get a name from the input, which should be the name of the root
+ // element of the upcoming content.
+ //
+ fReaderMgr.getName(bbRootName.getBuffer());
+ if (bbRootName.isEmpty())
+ {
+ emitError(XMLErrs::NoRootElemInDOCTYPE);
+ fReaderMgr.skipPastChar(chCloseAngle);
+ return;
+ }
+
+ //
+ // Store the root element name for later check
+ //
+ setRootElemName(bbRootName.getRawBuffer());
+
+ //
+ // This element obviously is not going to exist in the element decl
+ // pool yet, but we need to call docTypeDecl. So force it into
+ // the element decl pool, marked as being there because it was in
+ // the DOCTYPE. Later, when its declared, the status will be updated.
+ //
+ // Only do this if we are not reusing the validator! If we are reusing,
+ // then look it up instead. It has to exist!
+ //
+ DTDElementDecl* rootDecl;
+ Janitor<DTDElementDecl> janSrc(0);
+
+ if (fReuseGrammar)
+ {
+ if (fGrammar->getGrammarType() == Grammar::DTDGrammarType) {
+ rootDecl = (DTDElementDecl*) fGrammar->getElemDecl(fEmptyNamespaceId, 0, bbRootName.getRawBuffer(), Grammar::TOP_LEVEL_SCOPE);
+ if (rootDecl)
+ ((DTDGrammar*)fGrammar)->setRootElemId(rootDecl->getId());
+ else {
+ rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
+ rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
+ rootDecl->setExternalElemDeclaration(true);
+ ((DTDGrammar*)fGrammar)->setRootElemId(fGrammar->putElemDecl(rootDecl));
+ }
+ }
+ else {
+ rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
+ rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
+ rootDecl->setExternalElemDeclaration(true);
+ janSrc.reset(rootDecl);
+ }
+ }
+ else
+ {
+ rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
+ rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
+ rootDecl->setExternalElemDeclaration(true);
+ ((DTDGrammar*)fGrammar)->setRootElemId(fGrammar->putElemDecl(rootDecl));
+ }
+
+ // Skip any spaces after the name
+ fReaderMgr.skipPastSpaces();
+
+ //
+ // And now if we are looking at a >, then we are done. It is not
+ // required to have an internal or external subset, though why you
+ // would not escapes me.
+ //
+ if (fReaderMgr.skippedChar(chCloseAngle)) {
+ //
+ // If we have a doc type handler and advanced callbacks are enabled,
+ // call the doctype event.
+ //
+ if (fDocTypeHandler)
+ fDocTypeHandler->doctypeDecl(*rootDecl, 0, 0, false);
+ return;
+ }
+
+ // either internal/external subset
+ if(!fReuseGrammar) {
+ if (fValScheme == Val_Auto && !fValidate)
+ fValidate = true;
+ }
+
+
+ bool hasIntSubset = false;
+ bool hasExtSubset = false;
+ XMLCh* sysId = 0;
+ XMLCh* pubId = 0;
+
+ //
+ // If the next character is '[' then we have no external subset cause
+ // there is no system id, just the opening character of the internal
+ // subset. Else, has to be an id.
+ //
+ // Just look at the next char, don't eat it.
+ if (fReaderMgr.peekNextChar() == chOpenSquare)
+ {
+ hasIntSubset = true;
+ }
+ else
+ {
+ // Indicate we have an external subset
+ hasExtSubset = true;
+ fHasNoDTD = false;
+
+ // Get buffers for the ids
+ XMLBufBid bbPubId(&fBufMgr);
+ XMLBufBid bbSysId(&fBufMgr);
+
+ // Get the external subset id
+ if (!dtdScanner.scanId(bbPubId.getBuffer(), bbSysId.getBuffer(), DTDScanner::IDType_External))
+ {
+ fReaderMgr.skipPastChar(chCloseAngle);
+ return;
+ }
+
+ // Get copies of the ids we got
+ pubId = XMLString::replicate(bbPubId.getRawBuffer());
+ sysId = XMLString::replicate(bbSysId.getRawBuffer());
+
+ // Skip spaces and check again for the opening of an internal subset
+ fReaderMgr.skipPastSpaces();
+
+ // Just look at the next char, don't eat it.
+ if (fReaderMgr.peekNextChar() == chOpenSquare) {
+ hasIntSubset = true;
+ }
+ }
+
+ // Insure that the ids get cleaned up, if they got allocated
+ ArrayJanitor<XMLCh> janSysId(sysId);
+ ArrayJanitor<XMLCh> janPubId(pubId);
+
+ //
+ // If we have a doc type handler and advanced callbacks are enabled,
+ // call the doctype event.
+ //
+ if (fDocTypeHandler)
+ fDocTypeHandler->doctypeDecl(*rootDecl, pubId, sysId, hasIntSubset);
+
+ //
+ // Ok, if we had an internal subset, we are just past the [ character
+ // and need to parse that first.
+ //
+ if (hasIntSubset)
+ {
+ // Eat the opening square bracket
+ fReaderMgr.getNextChar();
+
+ // We can't have any internal subset if we are reusing the validator
+ if (fReuseGrammar)
+ ThrowXML(RuntimeException, XMLExcepts::Val_CantHaveIntSS);
+
+ //
+ // And try to scan the internal subset. If we fail, try to recover
+ // by skipping forward tot he close angle and returning.
+ //
+ if (!dtdScanner.scanInternalSubset())
+ {
+ fReaderMgr.skipPastChar(chCloseAngle);
+ return;
+ }
+
+ //
+ // Do a sanity check that some expanded PE did not propogate out of
+ // the doctype. This could happen if it was terminated early by bad
+ // syntax.
+ //
+ if (fReaderMgr.getReaderDepth() > 1)
+ {
+ emitError(XMLErrs::PEPropogated);
+
+ // Ask the reader manager to pop back down to the main level
+ fReaderMgr.cleanStackBackTo(1);
+ }
+
+ fReaderMgr.skipPastSpaces();
+ }
+
+ // And that should leave us at the closing > of the DOCTYPE line
+ if (!fReaderMgr.skippedChar(chCloseAngle))
+ {
+ //
+ // Do a special check for the common scenario of an extra ] char at
+ // the end. This is easy to recover from.
+ //
+ if (fReaderMgr.skippedChar(chCloseSquare)
+ && fReaderMgr.skippedChar(chCloseAngle))
+ {
+ emitError(XMLErrs::ExtraCloseSquare);
+ }
+ else
+ {
+ emitError(XMLErrs::UnterminatedDOCTYPE);
+ fReaderMgr.skipPastChar(chCloseAngle);
+ }
+ }
+
+ //
+ // If we had an external subset, then we need to deal with that one
+ // next. If we are reusing the validator, then don't scan it.
+ //
+ if (hasExtSubset && !fReuseGrammar && (fLoadExternalDTD || fValidate))
+ {
+ // And now create a reader to read this entity
+ InputSource* srcUsed;
+ XMLReader* reader = fReaderMgr.createReader
+ (
+ sysId
+ , pubId
+ , false
+ , XMLReader::RefFrom_NonLiteral
+ , XMLReader::Type_General
+ , XMLReader::Source_External
+ , srcUsed
+ );
+
+ // Put a janitor on the input source
+ Janitor<InputSource> janSrc(srcUsed);
+
+ //
+ // If it failed then throw an exception
+ //
+ if (!reader)
+ ThrowXML1(RuntimeException, XMLExcepts::Gen_CouldNotOpenDTD, srcUsed->getSystemId());
+
+ //
+ // In order to make the processing work consistently, we have to
+ // make this look like an external entity. So create an entity
+ // decl and fill it in and push it with the reader, as happens
+ // with an external entity. Put a janitor on it to insure it gets
+ // cleaned up. The reader manager does not adopt them.
+ //
+ const XMLCh gDTDStr[] = { chLatin_D, chLatin_T, chLatin_D , chNull };
+ DTDEntityDecl* declDTD = new DTDEntityDecl(gDTDStr);
+ declDTD->setSystemId(sysId);
+ Janitor<DTDEntityDecl> janDecl(declDTD);
+
+ // Mark this one as a throw at end
+ reader->setThrowAtEnd(true);
+
+ // And push it onto the stack, with its pseudo name
+ fReaderMgr.pushReader(reader, declDTD);
+
+ // Tell it its not in an include section
+ dtdScanner.scanExtSubsetDecl(false);
+ }
+}
+
bool XMLScanner::scanStartTag(bool& gotData)
{
//
diff --git a/src/xercesc/internal/XMLScanner.hpp b/src/xercesc/internal/XMLScanner.hpp
index 2350156d2116c2ee9019821e1bbc2435f0fdc573..a3374fc4f04b7ec87e54e74eea91d4e844811682 100644
--- a/src/xercesc/internal/XMLScanner.hpp
+++ b/src/xercesc/internal/XMLScanner.hpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.5 2002/05/30 16:20:57 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.4 2002/05/27 18:42:14 tng
* To get ready for 64 bit large file, use XMLSSize_t to represent line and column number.
*
@@ -396,6 +399,7 @@ public :
bool getHasNoDTD() const;
XMLCh* getExternalSchemaLocation() const;
XMLCh* getExternalNoNamespaceSchemaLocation() const;
+ bool getLoadExternalDTD() const;
// -----------------------------------------------------------------------
// Getter methods
@@ -492,6 +496,7 @@ public :
void setExternalNoNamespaceSchemaLocation(const XMLCh* const noNamespaceSchemaLocation);
void setExternalSchemaLocation(const char* const schemaLocation);
void setExternalNoNamespaceSchemaLocation(const char* const noNamespaceSchemaLocation);
+ void setLoadExternalDTD(const bool loadDTD);
// -----------------------------------------------------------------------
// Mutator methods
@@ -503,7 +508,7 @@ public :
// setValidationScheme() instead.
// -----------------------------------------------------------------------
bool getDoValidation() const;
- void setDoValidation(const bool validate, const bool setValScheme = true);
+ void setDoValidation(const bool validate);
// -----------------------------------------------------------------------
@@ -683,6 +688,7 @@ private :
XMLBuffer& toFill
, const XMLCh chEndChar
);
+ void scanDocTypeDecl();
// -----------------------------------------------------------------------
// Private helper methods
@@ -908,6 +914,9 @@ private :
// The no target namespace XML Schema Location that was specified
// externally using setExternalNoNamespaceSchemaLocation.
//
+ // fLoadExternalDTD
+ // This flag indicates whether the external DTD be loaded or not
+ //
// -----------------------------------------------------------------------
bool fDoNamespaces;
bool fExitOnFirstFatal;
@@ -964,6 +973,7 @@ private :
XMLCh* fRootElemName;
XMLCh* fExternalSchemaLocation;
XMLCh* fExternalNoNamespaceSchemaLocation;
+ bool fLoadExternalDTD;
};
@@ -1162,6 +1172,11 @@ inline XMLCh* XMLScanner::getExternalNoNamespaceSchemaLocation() const
return fExternalNoNamespaceSchemaLocation;
}
+inline bool XMLScanner::getLoadExternalDTD() const
+{
+ return fLoadExternalDTD;
+}
+
// ---------------------------------------------------------------------------
// XMLScanner: Setter methods
// ---------------------------------------------------------------------------
@@ -1281,6 +1296,11 @@ inline void XMLScanner::setExternalNoNamespaceSchemaLocation(const char* const n
fExternalNoNamespaceSchemaLocation = XMLString::transcode(noNamespaceSchemaLocation);
}
+inline void XMLScanner::setLoadExternalDTD(const bool loadDTD)
+{
+ fLoadExternalDTD = loadDTD;
+}
+
// ---------------------------------------------------------------------------
// XMLScanner: Mutator methods
@@ -1299,15 +1319,13 @@ inline bool XMLScanner::getDoValidation() const
return fValidate;
}
-inline void XMLScanner::setDoValidation(const bool validate, const bool setValScheme)
+inline void XMLScanner::setDoValidation(const bool validate)
{
fValidate = validate;
- if (setValScheme) {
- if (fValidate)
- fValScheme = Val_Always;
- else
- fValScheme = Val_Never;
- }
+ if (fValidate)
+ fValScheme = Val_Always;
+ else
+ fValScheme = Val_Never;
}
#endif
diff --git a/src/xercesc/parsers/AbstractDOMParser.cpp b/src/xercesc/parsers/AbstractDOMParser.cpp
index 5d2c65b84d023621eaae60eed8976a333370303c..ab7335fb2623da97be838debf0a76004fb75f7c6 100644
--- a/src/xercesc/parsers/AbstractDOMParser.cpp
+++ b/src/xercesc/parsers/AbstractDOMParser.cpp
@@ -227,6 +227,12 @@ XMLCh* AbstractDOMParser::getExternalNoNamespaceSchemaLocation() const
return fScanner->getExternalNoNamespaceSchemaLocation();
}
+bool AbstractDOMParser::getLoadExternalDTD() const
+{
+ return fScanner->getLoadExternalDTD();
+}
+
+
// ---------------------------------------------------------------------------
// AbstractDOMParser: Setter methods
// ---------------------------------------------------------------------------
@@ -283,6 +289,11 @@ void AbstractDOMParser::setExternalNoNamespaceSchemaLocation(const char* const n
fScanner->setExternalNoNamespaceSchemaLocation(noNamespaceSchemaLocation);
}
+void AbstractDOMParser::setLoadExternalDTD(const bool newState)
+{
+ fScanner->setLoadExternalDTD(newState);
+}
+
// ---------------------------------------------------------------------------
// AbstractDOMParser: Parsing methods
diff --git a/src/xercesc/parsers/AbstractDOMParser.hpp b/src/xercesc/parsers/AbstractDOMParser.hpp
index 60792b7cdc24d0076be6ed5db9cc6f4db727a315..00a0f05cacb3d3bb6709ce5e2b8b5a5080bf66dd 100644
--- a/src/xercesc/parsers/AbstractDOMParser.hpp
+++ b/src/xercesc/parsers/AbstractDOMParser.hpp
@@ -341,6 +341,19 @@ public :
*/
XMLCh* getExternalNoNamespaceSchemaLocation() const;
+ /** Get the 'Loading External DTD' flag
+ *
+ * This method returns the state of the parser's loading external DTD
+ * flag.
+ *
+ * @return false, if the parser is currently configured to
+ * ignore external DTD completely, true otherwise.
+ *
+ * @see #setLoadExternalDTD
+ * @see #getValidationScheme
+ */
+ bool getLoadExternalDTD() const;
+
//@}
@@ -576,6 +589,24 @@ public :
*/
void setExternalNoNamespaceSchemaLocation(const char* const noNamespaceSchemaLocation);
+ /** Set the 'Loading External DTD' flag
+ *
+ * This method allows users to enable or disable the loading of external DTD.
+ * When set to false, the parser will ignore any external DTD completely
+ * if the validationScheme is set to Val_Never.
+ *
+ * The parser's default state is: true.
+ *
+ * This flag is ignored if the validationScheme is set to Val_Always or Val_Auto.
+ *
+ * @param newState The value specifying whether external DTD should
+ * be loaded or not.
+ *
+ * @see #getLoadExternalDTD
+ * @see #setValidationScheme
+ */
+ void setLoadExternalDTD(const bool newState);
+
//@}
diff --git a/src/xercesc/parsers/SAX2XMLReaderImpl.cpp b/src/xercesc/parsers/SAX2XMLReaderImpl.cpp
index 848889370358b96eaee42b4a7b21ae4681d437af..8dfcea67323dc52b874395db86e3406417de41c6 100644
--- a/src/xercesc/parsers/SAX2XMLReaderImpl.cpp
+++ b/src/xercesc/parsers/SAX2XMLReaderImpl.cpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.7 2002/05/30 16:20:09 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.6 2002/05/29 21:37:47 knoaman
* Add baseURI to resolveEntity to support DOMInputSource.
*
@@ -1368,6 +1371,11 @@ void SAX2XMLReaderImpl::setFeature(const XMLCh* const name, const bool value)
fScanner->setValidationSchemaFullChecking(value);
}
+ else if (XMLString::compareIString(name, XMLUni::fgSAX2XercesLoadExternalDTD) == 0)
+ {
+ fScanner->setLoadExternalDTD(value);
+ }
+
else
throw SAXNotRecognizedException("Unknown Feature");
}
@@ -1390,6 +1398,8 @@ bool SAX2XMLReaderImpl::getFeature(const XMLCh* const name) const
return getDoSchema();
else if (XMLString::compareIString(name, XMLUni::fgSAX2XercesSchemaFullChecking) == 0)
return fScanner->getValidationSchemaFullChecking();
+ else if (XMLString::compareIString(name, XMLUni::fgSAX2XercesLoadExternalDTD) == 0)
+ return fScanner->getLoadExternalDTD();
else
throw SAXNotRecognizedException("Unknown Feature");
return false;
diff --git a/src/xercesc/parsers/SAX2XMLReaderImpl.hpp b/src/xercesc/parsers/SAX2XMLReaderImpl.hpp
index 0bc69782eeff51f1c01bdc4242a53b01bd327e4e..3d94ccba4850f7e5c55024ee39a1b17d43ac0d75 100644
--- a/src/xercesc/parsers/SAX2XMLReaderImpl.hpp
+++ b/src/xercesc/parsers/SAX2XMLReaderImpl.hpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.6 2002/05/30 16:20:09 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.5 2002/05/29 21:37:47 knoaman
* Add baseURI to resolveEntity to support DOMInputSource.
*
@@ -697,6 +700,7 @@ public :
* <br>http://apache.org/xml/features/validation/reuse-grammar (default: false)
* <br>http://apache.org/xml/features/validation/schema (default: true)
* <br>http://apache.org/xml/features/validation/schema-full-checking (default: false)
+ * <br>http://apache.org/xml/features/nonvalidating/load-external-dtd (default: true)
* <br>http://apache.org/xml/features/validation/reuse-validator (Deprecated) (default: false)
*
* @param name The unique identifier (URI) of the feature.
diff --git a/src/xercesc/parsers/SAXParser.cpp b/src/xercesc/parsers/SAXParser.cpp
index 809875de8fdaa13a029af4497034009a03f196c1..0cb242465f345e3c11d0ce17510e3490263be2fd 100644
--- a/src/xercesc/parsers/SAXParser.cpp
+++ b/src/xercesc/parsers/SAXParser.cpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.6 2002/05/30 16:20:09 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.5 2002/05/29 21:37:47 knoaman
* Add baseURI to resolveEntity to support DOMInputSource.
*
@@ -367,6 +370,11 @@ XMLCh* SAXParser::getExternalNoNamespaceSchemaLocation() const
return fScanner->getExternalNoNamespaceSchemaLocation();
}
+bool SAXParser::getLoadExternalDTD() const
+{
+ return fScanner->getLoadExternalDTD();
+}
+
// ---------------------------------------------------------------------------
// SAXParser: Setter methods
@@ -427,6 +435,11 @@ void SAXParser::setExternalNoNamespaceSchemaLocation(const char* const noNamespa
fScanner->setExternalNoNamespaceSchemaLocation(noNamespaceSchemaLocation);
}
+void SAXParser::setLoadExternalDTD(const bool newState)
+{
+ fScanner->setLoadExternalDTD(newState);
+}
+
// ---------------------------------------------------------------------------
// SAXParser: Overrides of the SAX Parser interface
diff --git a/src/xercesc/parsers/SAXParser.hpp b/src/xercesc/parsers/SAXParser.hpp
index e20da55dda3141be766e90c09acbfcacc11d7b2b..473931ad60e56850bcb454dd7b2e79e21de96cc5 100644
--- a/src/xercesc/parsers/SAXParser.hpp
+++ b/src/xercesc/parsers/SAXParser.hpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.6 2002/05/30 16:20:09 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.5 2002/05/29 21:37:47 knoaman
* Add baseURI to resolveEntity to support DOMInputSource.
*
@@ -407,6 +410,20 @@ public :
*/
XMLCh* getExternalNoNamespaceSchemaLocation() const;
+ /** Get the 'Loading External DTD' flag
+ *
+ * This method returns the state of the parser's loading external DTD
+ * flag.
+ *
+ * @return false, if the parser is currently configured to
+ * ignore external DTD completely, true otherwise.
+ *
+ * @see #setLoadExternalDTD
+ * @see #getValidationScheme
+ */
+ bool getLoadExternalDTD() const;
+
+
//@}
@@ -578,6 +595,25 @@ public :
*/
void setExternalNoNamespaceSchemaLocation(const char* const noNamespaceSchemaLocation);
+ /** Set the 'Loading External DTD' flag
+ *
+ * This method allows users to enable or disable the loading of external DTD.
+ * When set to false, the parser will ignore any external DTD completely
+ * if the validationScheme is set to Val_Never.
+ *
+ * The parser's default state is: true.
+ *
+ * This flag is ignored if the validationScheme is set to Val_Always or Val_Auto.
+ *
+ * @param newState The value specifying whether external DTD should
+ * be loaded or not.
+ *
+ * @see #getLoadExternalDTD
+ * @see #setValidationScheme
+ */
+ void setLoadExternalDTD(const bool newState);
+
+
//@}
diff --git a/src/xercesc/sax2/SAX2XMLReader.hpp b/src/xercesc/sax2/SAX2XMLReader.hpp
index 295a569313452c8c3f0e62bc07041730f16a74e0..c519713ec63f5f815487a0e452609e8519794e21 100644
--- a/src/xercesc/sax2/SAX2XMLReader.hpp
+++ b/src/xercesc/sax2/SAX2XMLReader.hpp
@@ -56,8 +56,11 @@
/*
* $Log$
- * Revision 1.1 2002/02/01 22:22:09 peiyongz
- * Initial revision
+ * Revision 1.2 2002/05/30 16:22:53 tng
+ * Add feature to optionally ignore external DTD.
+ *
+ * Revision 1.1.1.1 2002/02/01 22:22:09 peiyongz
+ * sane_include
*
* Revision 1.18 2002/01/28 18:45:40 knoaman
* Update documentation for SAX2 feature 'namespace-prefixes'.
@@ -519,6 +522,7 @@ public:
* <br>http://apache.org/xml/features/validation/reuse-grammar (default: false)
* <br>http://apache.org/xml/features/validation/schema (default: true)
* <br>http://apache.org/xml/features/validation/schema-full-checking (default: false)
+ * <br>http://apache.org/xml/features/nonvalidating/load-external-dtd (default: true)
* <br>http://apache.org/xml/features/validation/reuse-validator (Deprecated) (default: false)
*
* @param name The unique identifier (URI) of the feature.
diff --git a/src/xercesc/util/XMLUni.cpp b/src/xercesc/util/XMLUni.cpp
index 799a667c53f8c387a152cea8646e7d36ffeafa94..74032aaa7e80684956e60083c86e15213e11d2c8 100644
--- a/src/xercesc/util/XMLUni.cpp
+++ b/src/xercesc/util/XMLUni.cpp
@@ -914,6 +914,22 @@ const XMLCh XMLUni::fgSAX2XercesSchemaFullChecking[] =
, chLatin_e, chLatin_c, chLatin_k, chLatin_i, chLatin_n, chLatin_g, chNull
};
+//Xerces: http://apache.org/xml/features/nonvalidating/load-external-dtd
+const XMLCh XMLUni::fgSAX2XercesLoadExternalDTD[] =
+{
+ chLatin_h, chLatin_t, chLatin_t, chLatin_p, chColon, chForwardSlash
+ , chForwardSlash, chLatin_a, chLatin_p, chLatin_a, chLatin_c, chLatin_h
+ , chLatin_e, chPeriod, chLatin_o, chLatin_r, chLatin_g, chForwardSlash
+ , chLatin_x, chLatin_m, chLatin_l, chForwardSlash, chLatin_f, chLatin_e
+ , chLatin_a, chLatin_t, chLatin_u, chLatin_r, chLatin_e, chLatin_s
+ , chForwardSlash, chLatin_n, chLatin_o, chLatin_n
+ , chLatin_v, chLatin_a, chLatin_l, chLatin_i, chLatin_d
+ , chLatin_a, chLatin_t, chLatin_i, chLatin_n, chLatin_g, chForwardSlash
+ , chLatin_l, chLatin_o, chLatin_a, chLatin_d, chDash
+ , chLatin_e, chLatin_x, chLatin_t, chLatin_e, chLatin_r, chLatin_n, chLatin_a, chLatin_l, chDash
+ , chLatin_d, chLatin_t, chLatin_d, chNull
+};
+
//deprecated
//Xerces: http://apache.org/xml/features/validation/reuse-validator
const XMLCh XMLUni::fgSAX2XercesReuseValidator[] =
@@ -977,7 +993,7 @@ const XMLCh XMLUni::fgDOMCDATASections[] =
, chLatin_e, chLatin_c, chLatin_t, chLatin_i, chLatin_o, chLatin_n, chLatin_s, chNull
};
-const XMLCh XMLUni::fgDOMComments[] =
+const XMLCh XMLUni::fgDOMComments[] =
{
chLatin_c, chLatin_o, chLatin_m, chLatin_m, chLatin_e, chLatin_n, chLatin_t
, chLatin_s, chNull
diff --git a/src/xercesc/util/XMLUni.hpp b/src/xercesc/util/XMLUni.hpp
index 04fa7e224f7a39e7b786aab0b40c437d3110aaf3..d21cb9b15bb970dc941f58c108d0c5bf9ef742c2 100644
--- a/src/xercesc/util/XMLUni.hpp
+++ b/src/xercesc/util/XMLUni.hpp
@@ -227,6 +227,7 @@ public :
static const XMLCh fgSAX2XercesSchemaFullChecking[];
static const XMLCh fgSAX2XercesSchemaExternalSchemaLocation[];
static const XMLCh fgSAX2XercesSchemaExternalNoNameSpaceSchemaLocation[];
+ static const XMLCh fgSAX2XercesLoadExternalDTD[];
// DOMBuilder features
static const XMLCh fgDOMCanonicalForm[];
diff --git a/src/xercesc/validators/DTD/DTDScanner.cpp b/src/xercesc/validators/DTD/DTDScanner.cpp
index 48af45383f23412b896ca2357730dc757ebd7622..e49c899ef1f8b936a7cda359fdc518f2bacf1793 100644
--- a/src/xercesc/validators/DTD/DTDScanner.cpp
+++ b/src/xercesc/validators/DTD/DTDScanner.cpp
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.5 2002/05/30 16:17:19 tng
+ * Add feature to optionally ignore external DTD.
+ *
* Revision 1.4 2002/05/03 14:51:16 peiyongz
* Bug#8769: UMR detected by memory tool - patch from Kenneth Palsson
*
@@ -1621,288 +1624,6 @@ void DTDScanner::scanDefaultDecl(DTDAttDef& toFill)
}
-//
-// This method handles the high level logic of scanning the DOCType
-// declaration. This kicks off both the scanning of the internal subset and
-// the scanning of the external subset, if any.
-//
-// When we get here the '<!DOCTYPE' part has already been scanned, which is
-// what told us that we had a doc type decl to parse.
-//
-
-void DTDScanner::scanDocTypeDecl(const bool reuseGrammar)
-{
- if (fDocTypeHandler)
- fDocTypeHandler->resetDocType();
-
- // There must be some space after DOCTYPE
- if (!fReaderMgr->skipPastSpaces())
- {
- fScanner->emitError(XMLErrs::ExpectedWhitespace);
-
- // Just skip the Doctype declaration and return
- fReaderMgr->skipPastChar(chCloseAngle);
- return;
- }
-
- // Get a buffer for the root element
- XMLBufBid bbRootName(fBufMgr);
-
- //
- // Get a name from the input, which should be the name of the root
- // element of the upcoming content.
- //
- fReaderMgr->getName(bbRootName.getBuffer());
- if (bbRootName.isEmpty())
- {
- fScanner->emitError(XMLErrs::NoRootElemInDOCTYPE);
- fReaderMgr->skipPastChar(chCloseAngle);
- return;
- }
-
- //
- // Store the root element name for later check
- //
- fScanner->setRootElemName(bbRootName.getRawBuffer());
-
- //
- // This element obviously is not going to exist in the element decl
- // pool yet, but we need to call docTypeDecl. So force it into
- // the element decl pool, marked as being there because it was in
- // the DOCTYPE. Later, when its declared, the status will be updated.
- //
- // Only do this if we are not reusing the validator! If we are reusing,
- // then look it up instead. It has to exist!
- //
- DTDElementDecl* rootDecl;
- Janitor<DTDElementDecl> janSrc(0);
-
- if (reuseGrammar)
- {
- Grammar* fGrammar = fDTDGrammar;
- if (fGrammar->getGrammarType() == Grammar::DTDGrammarType) {
- rootDecl = (DTDElementDecl*) fDTDGrammar->getElemDecl(fEmptyNamespaceId, 0, bbRootName.getRawBuffer(), Grammar::TOP_LEVEL_SCOPE);
- if (rootDecl)
- fDTDGrammar->setRootElemId(rootDecl->getId());
- else {
- rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
- rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
- rootDecl->setExternalElemDeclaration(isReadingExternalEntity());
- fDTDGrammar->setRootElemId(fDTDGrammar->putElemDecl(rootDecl));
- }
- }
- else {
- rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
- rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
- rootDecl->setExternalElemDeclaration(isReadingExternalEntity());
- janSrc.reset(rootDecl);
- }
- }
- else
- {
- rootDecl = new DTDElementDecl(bbRootName.getRawBuffer(), fEmptyNamespaceId);
- rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
- rootDecl->setExternalElemDeclaration(isReadingExternalEntity());
- fDTDGrammar->setRootElemId(fDTDGrammar->putElemDecl(rootDecl));
- }
-
- // Skip any spaces after the name
- fReaderMgr->skipPastSpaces();
-
- //
- // And now if we are looking at a >, then we are done. It is not
- // required to have an internal or external subset, though why you
- // would not escapes me.
- //
- if (fReaderMgr->skippedChar(chCloseAngle)) {
- //
- // If we have a doc type handler and advanced callbacks are enabled,
- // call the doctype event.
- //
- if (fDocTypeHandler)
- fDocTypeHandler->doctypeDecl(*rootDecl, 0, 0, false);
- return;
- }
-
- // either internal/external subset
- if(!reuseGrammar) {
- if (fScanner->getValidationScheme() == XMLScanner::Val_Auto)
- fScanner->setDoValidation(true, false);
- }
-
-
- bool hasIntSubset = false;
- bool hasExtSubset = false;
- XMLCh* sysId = 0;
- XMLCh* pubId = 0;
-
- //
- // If the next character is '[' then we have no external subset cause
- // there is no system id, just the opening character of the internal
- // subset. Else, has to be an id.
- //
- // Just look at the next char, don't eat it.
- if (fReaderMgr->peekNextChar() == chOpenSquare)
- {
- hasIntSubset = true;
- }
- else
- {
- // Indicate we have an external subset
- hasExtSubset = true;
- fScanner->setHasNoDTD(false);
-
- // Get buffers for the ids
- XMLBufBid bbPubId(fBufMgr);
- XMLBufBid bbSysId(fBufMgr);
-
- // Get the external subset id
- if (!scanId(bbPubId.getBuffer(), bbSysId.getBuffer(), IDType_External))
- {
- fReaderMgr->skipPastChar(chCloseAngle);
- return;
- }
-
- // Get copies of the ids we got
- pubId = XMLString::replicate(bbPubId.getRawBuffer());
- sysId = XMLString::replicate(bbSysId.getRawBuffer());
-
- // Skip spaces and check again for the opening of an internal subset
- fReaderMgr->skipPastSpaces();
-
- // Just look at the next char, don't eat it.
- if (fReaderMgr->peekNextChar() == chOpenSquare) {
- hasIntSubset = true;
- }
- }
-
- // Insure that the ids get cleaned up, if they got allocated
- ArrayJanitor<XMLCh> janSysId(sysId);
- ArrayJanitor<XMLCh> janPubId(pubId);
-
- //
- // If we have a doc type handler and advanced callbacks are enabled,
- // call the doctype event.
- //
- if (fDocTypeHandler)
- fDocTypeHandler->doctypeDecl(*rootDecl, pubId, sysId, hasIntSubset);
-
- //
- // Ok, if we had an internal subset, we are just past the [ character
- // and need to parse that first.
- //
- if (hasIntSubset)
- {
- // Eat the opening square bracket
- fReaderMgr->getNextChar();
-
- // We can't have any internal subset if we are reusing the validator
- if (reuseGrammar)
- ThrowXML(RuntimeException, XMLExcepts::Val_CantHaveIntSS);
-
- // Indicate we are in the internal subset now
- FlagJanitor<bool> janContentFlag(&fInternalSubset, true);
-
- //
- // And try to scan the internal subset. If we fail, try to recover
- // by skipping forward tot he close angle and returning.
- //
- if (!scanInternalSubset())
- {
- fReaderMgr->skipPastChar(chCloseAngle);
- return;
- }
-
- //
- // Do a sanity check that some expanded PE did not propogate out of
- // the doctype. This could happen if it was terminated early by bad
- // syntax.
- //
- if (fReaderMgr->getReaderDepth() > 1)
- {
- fScanner->emitError(XMLErrs::PEPropogated);
-
- // Ask the reader manager to pop back down to the main level
- fReaderMgr->cleanStackBackTo(1);
- }
-
- fReaderMgr->skipPastSpaces();
- }
-
- // And that should leave us at the closing > of the DOCTYPE line
- if (!fReaderMgr->skippedChar(chCloseAngle))
- {
- //
- // Do a special check for the common scenario of an extra ] char at
- // the end. This is easy to recover from.
- //
- if (fReaderMgr->skippedChar(chCloseSquare)
- && fReaderMgr->skippedChar(chCloseAngle))
- {
- fScanner->emitError(XMLErrs::ExtraCloseSquare);
- }
- else
- {
- fScanner->emitError(XMLErrs::UnterminatedDOCTYPE);
- fReaderMgr->skipPastChar(chCloseAngle);
- }
- }
-
- //
- // If we had an external subset, then we need to deal with that one
- // next. If we are reusing the validator, then don't scan it.
- //
- if (hasExtSubset && !reuseGrammar)
- {
- // Indicate we are in the external subset now
- FlagJanitor<bool> janContentFlag(&fInternalSubset, false);
-
- // And now create a reader to read this entity
- InputSource* srcUsed;
- XMLReader* reader = fReaderMgr->createReader
- (
- sysId
- , pubId
- , false
- , XMLReader::RefFrom_NonLiteral
- , XMLReader::Type_General
- , XMLReader::Source_External
- , srcUsed
- );
-
- // Put a janitor on the input source
- Janitor<InputSource> janSrc(srcUsed);
-
- //
- // If it failed then throw an exception
- //
- if (!reader)
- ThrowXML1(RuntimeException, XMLExcepts::Gen_CouldNotOpenDTD, srcUsed->getSystemId());
-
- //
- // In order to make the processing work consistently, we have to
- // make this look like an external entity. So create an entity
- // decl and fill it in and push it with the reader, as happens
- // with an external entity. Put a janitor on it to insure it gets
- // cleaned up. The reader manager does not adopt them.
- //
- const XMLCh gDTDStr[] = { chLatin_D, chLatin_T, chLatin_D , chNull };
- DTDEntityDecl* declDTD = new DTDEntityDecl(gDTDStr);
- declDTD->setSystemId(sysId);
- Janitor<DTDEntityDecl> janDecl(declDTD);
-
- // Mark this one as a throw at end
- reader->setThrowAtEnd(true);
-
- // And push it onto the stack, with its pseudo name
- fReaderMgr->pushReader(reader, declDTD);
-
- // Tell it its not in an include section
- scanExtSubsetDecl(false);
- }
-}
-
-
//
// This is called after seeing '<!ELEMENT' which indicates that an element
// markup is starting. This guy scans the rest of it and adds it to the
@@ -2680,6 +2401,10 @@ bool DTDScanner::scanEq()
//
void DTDScanner::scanExtSubsetDecl(const bool inIncludeSect)
{
+ // Indicate we are in the external subset now
+ FlagJanitor<bool> janContentFlag(&fInternalSubset, false);
+
+
bool bAcceptDecl = !inIncludeSect;
// Get a buffer for whitespace
@@ -3095,6 +2820,9 @@ void DTDScanner::scanIgnoredSection()
//
bool DTDScanner::scanInternalSubset()
{
+ // Indicate we are in the internal subset now
+ FlagJanitor<bool> janContentFlag(&fInternalSubset, true);
+
// If we have a doc type handler, tell it the internal subset starts
if (fDocTypeHandler)
fDocTypeHandler->startIntSubset();
diff --git a/src/xercesc/validators/DTD/DTDScanner.hpp b/src/xercesc/validators/DTD/DTDScanner.hpp
index aa4e91da74b6b72de69179ecf20a18c049803ccb..77261e2f92191890c35a8fb2066232b98f7db90e 100644
--- a/src/xercesc/validators/DTD/DTDScanner.hpp
+++ b/src/xercesc/validators/DTD/DTDScanner.hpp
@@ -56,8 +56,11 @@
/*
* $Log$
- * Revision 1.1 2002/02/01 22:22:44 peiyongz
- * Initial revision
+ * Revision 1.2 2002/05/30 16:17:19 tng
+ * Add feature to optionally ignore external DTD.
+ *
+ * Revision 1.1.1.1 2002/02/01 22:22:44 peiyongz
+ * sane_include
*
* Revision 1.4 2001/06/21 14:25:56 knoaman
* Fix for bug 1946
@@ -92,12 +95,12 @@ public:
// -----------------------------------------------------------------------
// Class specific types
//
- // NOTE: This should really be private, but some of the compilers we
- // have to support cannot understand that.
- //
// EntityExpRes
// Returned from scanEntityRef() to indicate how the expanded text
// was treated.
+ //
+ // IDTypes
+ // Type of the ID
// -----------------------------------------------------------------------
enum EntityExpRes
{
@@ -106,6 +109,14 @@ public:
, EntityExp_Returned
};
+ enum IDTypes
+ {
+ IDType_Public
+ , IDType_External
+ , IDType_Either
+ };
+
+
// -----------------------------------------------------------------------
// Constructors and Destructor
@@ -137,20 +148,16 @@ public:
DocTypeHandler* const handlerToSet
);
- void scanDocTypeDecl(const bool reuseGrammar);
+ void scanExtSubsetDecl(const bool inIncludeSect);
+ bool scanInternalSubset();
+ bool scanId
+ (
+ XMLBuffer& pubIdToFill
+ , XMLBuffer& sysIdToFill
+ , const IDTypes whatKind
+ );
private:
- // -----------------------------------------------------------------------
- // Private class types
- // -----------------------------------------------------------------------
- enum IDTypes
- {
- IDType_Public
- , IDType_External
- , IDType_Either
- };
-
-
// -----------------------------------------------------------------------
// Private DTD scanning methods. These are all in XMLValidator2.cpp
// -----------------------------------------------------------------------
@@ -199,15 +206,7 @@ private:
, const bool notation
);
bool scanEq();
- void scanExtSubsetDecl(const bool inIncludeSect);
- bool scanId
- (
- XMLBuffer& pubIdToFill
- , XMLBuffer& sysIdToFill
- , const IDTypes whatKind
- );
void scanIgnoredSection();
- bool scanInternalSubset();
void scanMarkupDecl(const bool parseTextDecl);
bool scanMixed(DTDElementDecl& toFill);
void scanNotationDecl();