mozilla
/
pjs
зеркало из https://github.com/mozilla/pjs.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

updated comments

This commit is contained in:
rickg 1998-05-12 00:59:32 +00:00
Родитель 45f5c658fc
Коммит ba2be950ad
10 изменённых файлов: 1590 добавлений и 278 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

140
htmlparser/src/CNavDelegate.h
Просмотреть файл

@ -46,33 +46,171 @@
class CNavDelegate : public ITokenizerDelegate {
public:
/**
* Default constructor
* @update gess5/11/98
*/
CNavDelegate();
/**
* Copy constructor
* @update gess 5/11/98
*/
CNavDelegate(CNavDelegate& aDelegate);
/**
* Consume next token from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 GetToken(CScanner& aScanner,CToken*& aToken);
/**
* Ask if its ok to add this token
* @update gess 5/11/98
* @param aToken is the token to be added
* @return True if ok to add the given token
*/
virtual PRBool WillAddToken(CToken& aToken);
/**
* Called as a preprocess -- tells delegate that tokenization will begin
* @update gess 5/11/98
* @param aIncremental tells us if tokenization is incremental
* @return TRUE if ok to continue -- FALSE if process should stop
*/
virtual PRBool WillTokenize(PRBool aIncremental);
/**
* Postprocess -- called to say that tokenization has concluded
* @update gess 5/11/98
* @param aIncremental tells us if tokenization was incremental
* @return TRUE if all went well--FALSE if you encountered an error
*/
virtual PRBool DidTokenize(PRBool aIncremental);
/**
* DEPRECATED. Tells us what mode the delegate is operating in.
* @update gess 5/11/98
* @return parse mode
*/
virtual eParseMode GetParseMode(void) const;
/**
* Retrieve the DTD required by this delegate
* (The parser will call this prior to tokenization)
* @update gess 5/11/98
* @return ptr to DTD -- should NOT be null.
*/
virtual nsIDTD* GetDTD(void) const;
/**
* Conduct self test.
* @update gess 5/11/98
*/
static void SelfTest();
protected:
/**
* Called to cause delegate to create a token of given type.
* @update gess 5/11/98
* @param aType represents the kind of token you want to create.
* @return new token or NULL
*/
virtual CToken* CreateTokenOfType(eHTMLTokenTypes aType);
/**
* Retrieve the next TAG from the given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve next START tag from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeStartTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve collection of HTML/XML attributes from given scanner
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeAttributes(PRUnichar aChar,CScanner& aScanner);
/**
* Retrieve a sequence of text from given scanner.
* @update gess 5/11/98
* @param aString will contain retrieved text.
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeText(const nsString& aString,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve an entity from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeEntity(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a whitespace sequence from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeWhitespace(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a comment from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeComment(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve newlines from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeNewline(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
//the only special case method...
/**
* Causes content to be skipped up to sequence contained in aString.
* @update gess 5/11/98
* @param aString ????
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 ConsumeContentToEndTag(const nsString& aString,PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
nsDeque mTokenDeque;

140
htmlparser/src/COtherDelegate.h
Просмотреть файл

@ -46,33 +46,171 @@
class COtherDelegate : public ITokenizerDelegate {
public:
/**
* Default constructor
* @update gess5/11/98
*/
COtherDelegate();
/**
* Copy constructor
* @update gess 5/11/98
*/
COtherDelegate(COtherDelegate& aDelegate);
/**
* Consume next token from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 GetToken(CScanner& aScanner,CToken*& aToken);
/**
* Ask if its ok to add this token
* @update gess 5/11/98
* @param aToken is the token to be added
* @return True if ok to add the given token
*/
virtual PRBool WillAddToken(CToken& aToken);
/**
* Called as a preprocess -- tells delegate that tokenization will begin
* @update gess 5/11/98
* @param aIncremental tells us if tokenization is incremental
* @return TRUE if ok to continue -- FALSE if process should stop
*/
virtual PRBool WillTokenize(PRBool aIncremental);
/**
* Postprocess -- called to say that tokenization has concluded
* @update gess 5/11/98
* @param aIncremental tells us if tokenization was incremental
* @return TRUE if all went well--FALSE if you encountered an error
*/
virtual PRBool DidTokenize(PRBool aIncremental);
/**
* DEPRECATED. Tells us what mode the delegate is operating in.
* @update gess 5/11/98
* @return parse mode
*/
virtual eParseMode GetParseMode(void) const;
/**
* Retrieve the DTD required by this delegate
* (The parser will call this prior to tokenization)
* @update gess 5/11/98
* @return ptr to DTD -- should NOT be null.
*/
virtual nsIDTD* GetDTD(void) const;
/**
* Conduct self test.
* @update gess 5/11/98
*/
static void SelfTest();
protected:
/**
* Called to cause delegate to create a token of given type.
* @update gess 5/11/98
* @param aType represents the kind of token you want to create.
* @return new token or NULL
*/
virtual CToken* CreateTokenOfType(eHTMLTokenTypes aType);
/**
* Retrieve the next TAG from the given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve next START tag from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeStartTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve collection of HTML/XML attributes from given scanner
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeAttributes(PRUnichar aChar,CScanner& aScanner);
/**
* Retrieve a sequence of text from given scanner.
* @update gess 5/11/98
* @param aString will contain retrieved text.
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeText(const nsString& aString,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve an entity from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeEntity(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a whitespace sequence from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeWhitespace(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a comment from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeComment(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve newlines from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeNewline(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
//the only special case method...
/**
* Causes content to be skipped up to sequence contained in aString.
* @update gess 5/11/98
* @param aString ????
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 ConsumeContentToEndTag(const nsString& aString,PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
nsDeque mTokenDeque;

18
htmlparser/src/nsHTMLParser.cpp
Просмотреть файл

@ -879,7 +879,7 @@ PRBool nsHTMLParser::HandleSkippedContentToken(CToken* aToken) {
PRBool result=PR_TRUE;
if(IsWithinBody()) {
if(HasOpenContainer(eHTMLTag_body)) {
nsCParserNode aNode((CHTMLToken*)aToken);
result=AddLeaf(aNode);
}
@ -937,22 +937,6 @@ PRBool nsHTMLParser::HandleStyleToken(CToken* aToken){
return result;
}
/**
* This method gets called to determine whether a given
* tag is itself a container
*
* @update gess 3/25/98
* @param aTag -- tag to test for containership
* @return PR_TRUE if given tag can contain other tags
*/
PRBool nsHTMLParser::IsWithinBody(void) const {
for(int i=0;i<mContextStackPos;i++) {
if(eHTMLTag_body==mContextStack[i])
return PR_TRUE;
}
return PR_FALSE;
}
/**
* This method does two things: 1st, help construct

353
htmlparser/src/nsHTMLParser.h
Просмотреть файл

@ -81,72 +81,417 @@ class nsHTMLParser : public nsIParser {
friend class CTokenHandler;
NS_DECL_ISUPPORTS
/**
* default constructor
* @update gess5/11/98
*/
nsHTMLParser();
/**
* Destructor
* @update gess5/11/98
*/
~nsHTMLParser();
/**
* Select given content sink into parser for parser output
* @update gess5/11/98
* @param aSink is the new sink to be used by parser
* @return old sink, or NULL
*/
virtual nsIContentSink* SetContentSink(nsIContentSink* aSink);
/**
* Cause parser to parse input from given URL
* @update gess5/11/98
* @param aURL is a descriptor for source document
* @return TRUE if all went well -- FALSE otherwise
*/
virtual PRBool Parse(nsIURL* aURL);
/**
* Cause parser to parse input from given URL in given mode
* @update gess5/11/98
* @param aURL is a descriptor for source document
* @param aMode is the desired parser mode (Nav, other, etc.)
* @return TRUE if all went well -- FALSE otherwise
*/
virtual PRBool Parse(nsIURL* aURL,eParseMode aMode);
/**
* This method gets called (automatically) during incremental parsing
* @update gess5/11/98
* @return TRUE if all went well, otherwise FALSE
*/
virtual PRBool ResumeParse();
/**
* Retrieve ptr to internal context vector stack
* @update gess5/11/98
* @param stack ptr to be set
* @return count of items on stack (may be 0)
*/
virtual PRInt32 GetStack(PRInt32* aStackPtr);
/**
* Ask parser if a given container is open ANYWHERE on stack
* @update gess5/11/98
* @param id of container you want to test for
* @return TRUE if the given container type is open -- otherwise FALSE
*/
virtual PRBool HasOpenContainer(PRInt32 aContainer) const;
/**
* This method gets called when a start token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the start token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleStartToken(CToken* aToken);
/**
* This method gets called when a start token has been consumed, and
* we want to use default start token handling behavior.
* This method gets called automatically by handleStartToken.
*
* @update gess5/11/98
* @param aToken is the start token to be handled
* @param aChildTag is the tag-type of given token
* @param aNode is a node be updated with info from given token
* @return TRUE if the token was handled.
*/
PRBool HandleDefaultStartToken(CToken* aToken,eHTMLTags aChildTag,nsCParserNode& aNode);
/**
* This method gets called when an end token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the end token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleEndToken(CToken* aToken);
/**
* This method gets called when an entity token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the entity token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleEntityToken(CToken* aToken);
/**
* This method gets called when a comment token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the comment token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleCommentToken(CToken* aToken);
/**
* This method gets called when a skipped-content token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the skipped-content token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleSkippedContentToken(CToken* aToken);
/**
* This method gets called when an attribute token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the attribute token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleAttributeToken(CToken* aToken);
/**
* This method gets called when a script token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the script token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleScriptToken(CToken* aToken);
/**
* This method gets called when a style token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the style token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleStyleToken(CToken* aToken);
PRBool IsWithinBody(void) const;
protected:
/**
* This method gets called when the tokens have been consumed, and it's time
* to build the model via the content sink.
* @update gess5/11/98
* @return YES if model building went well -- NO otherwise.
*/
PRBool IterateTokens();
/**
* Retrieves the tag type for the element on the context vector stack at given pos.
* @update gess5/11/98
* @param pos of item you want to retrieve
* @return tag type (may be unknown)
*/
eHTMLTags NodeAt(PRInt32 aPos) const;
/**
* Retrieve the tag type of the topmost item on context vector stack
* @update gess5/11/98
* @return tag type (may be unknown)
*/
eHTMLTags GetTopNode() const;
/**
* Retrieve the number of items on context vector stack
* @update gess5/11/98
* @return count of items on stack -- may be 0
*/
PRInt32 GetStackPos() const;
/**
* Causes the parser to scan foward, collecting nearby (sequential)
* attribute tokens into the given node.
* @update gess5/11/98
* @param node to store attributes
* @return number of attributes added to node.
*/
PRInt32 CollectAttributes(nsCParserNode& aNode);
/**
* Causes the next skipped-content token (if any) to
* be consumed by this node.
* @update gess5/11/98
* @param node to consume skipped-content
* @return number of skipped-content tokens consumed.
*/
PRInt32 CollectSkippedContent(nsCParserNode& aNode);
/**
* Causes token handlers to be registered for this parser.
* DO NOT CALL THIS! IT'S DEPRECATED!
* @update gess5/11/98
*/
void InitializeDefaultTokenHandlers();
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* GetTokenHandler(const nsString& aString) const;
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* GetTokenHandler(eHTMLTokenTypes aType) const;
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* AddTokenHandler(CTokenHandler* aHandler);
/**
* DEPRECATED
* @update gess5/11/98
*/
nsHTMLParser& DeleteTokenHandlers(void);
protected:
//*************************************************
//these cover methods mimic the sink, and are used
//by the parser to manage its context-stack.
//*************************************************
/**
* This cover method opens the given node as a HTML item in
* content sink.
* @update gess5/11/98
* @param HTML (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenHTML(const nsIParserNode& aNode);
/**
*
* @update gess5/11/98
* @param
* @return
*/
PRBool CloseHTML(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a head item in
* content sink.
* @update gess5/11/98
* @param HEAD (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenHead(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink head to be closed
* @update gess5/11/98
* @param aNode is the node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseHead(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a body item in
* content sink.
* @update gess5/11/98
* @param BODY (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenBody(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink body to be closed
* @update gess5/11/98
* @param aNode is the body node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseBody(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a form item in
* content sink.
* @update gess5/11/98
* @param FORM (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenForm(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink form to be closed
* @update gess5/11/98
* @param aNode is the form node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseForm(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a frameset item in
* content sink.
* @update gess5/11/98
* @param FRAMESET (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenFrameset(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink frameset to be closed
* @update gess5/11/98
* @param aNode is the frameeset node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseFrameset(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a generic container in
* content sink.
* @update gess5/11/98
* @param generic container (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenContainer(const nsIParserNode& aNode);
/**
* This cover method causes a generic containre in the content-sink to be closed
* @update gess5/11/98
* @param aNode is the node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseContainer(const nsIParserNode& aNode);
/**
* This cover method causes the topmost container to be closed in sink
* @update gess5/11/98
* @return TRUE if all went well.
*/
PRBool CloseTopmostContainer();
/**
* Cause all containers down to topmost given tag to be closed
* @update gess5/11/98
* @param aTag is the tag at which auto-closure should stop (inclusive)
* @return TRUE if all went well -- otherwise FALSE
*/
PRBool CloseContainersTo(eHTMLTags aTag);
/**
* Cause all containers down to given position to be closed
* @update gess5/11/98
* @param anIndex is the stack pos at which auto-closure should stop (inclusive)
* @return TRUE if all went well -- otherwise FALSE
*/
PRBool CloseContainersTo(PRInt32 anIndex);
/**
* Causes leaf to be added to sink at current vector pos.
* @update gess5/11/98
* @param aNode is leaf node to be added.
* @return TRUE if all went well -- FALSE otherwise.
*/
PRBool AddLeaf(const nsIParserNode& aNode);
/**
* Determine if the given tag is open in the context vector stack.
* @update gess5/11/98
* @param aTag is the type of tag you want to test for openness
* @return TRUE if open, FALSE otherwise.
*/
PRBool IsOpen(eHTMLTags aTag) const;
/**
* Finds the topmost occurance of given tag within context vector stack.
* @update gess5/11/98
* @param tag to be found
* @return index of topmost tag occurance -- may be -1 (kNotFound).
*/
PRInt32 GetTopmostIndex(eHTMLTags aTag) const;
/**
* Causes auto-closures of context vector stack in order to find a
* proper home for the given child. Propagation may also occur as
* a fall out.
* @update gess5/11/98
* @param child to be added (somewhere) to context vector stack.
* @return TRUE if succeeds, otherwise FALSE
*/
PRBool ReduceContextStackFor(PRInt32 aChildTag);
/**
* Attempt forward and/or backward propagation for the given
* child within the current context vector stack.
* @update gess5/11/98
* @param type of child to be propagated.
* @return TRUE if succeeds, otherwise FALSE
*/
PRBool CreateContextStackFor(PRInt32 aChildTag);
//*********************************************
// And now, some data members...
//*********************************************
nsIHTMLContentSink* mSink;
CTokenizer* mTokenizer;
// eHTMLTags mContextStack[50];
PRInt32 mContextStack[50];
PRInt32 mContextStackPos;

59
htmlparser/src/nsIParserNode.h
Просмотреть файл

@ -52,21 +52,72 @@ class nsIParserNode {
public:
/**
* Retrieve the name of the node
* @update gess5/11/98
* @return string containing node name
*/
virtual const nsString& GetName() const =0; //to get name of tag
/**
* Retrieve the text from the given node
* @update gess5/11/98
* @return string containing node text
*/
virtual const nsString& GetText() const =0; //get plain text if available
/**
* Retrieve skipped context from node
* @update gess5/11/98
* @return string containing skipped content
*/
virtual const nsString& GetSkippedContent() const =0;
//methods for determining the type of parser node...
/**
* Retrieve the type of the parser node.
* @update gess5/11/98
* @return node type.
*/
virtual PRInt32 GetNodeType() const =0;
/**
* Retrieve token type of parser node
* @update gess5/11/98
* @return token type
*/
virtual PRInt32 GetTokenType() const =0;
//methods for accessing key/value pairs
/**
* Retrieve the number of attributes in this node.
* @update gess5/11/98
* @return count of attributes (may be 0)
*/
virtual PRInt32 GetAttributeCount(void) const =0;
/**
* Retrieve the key (of key/value pair) at given index
* @update gess5/11/98
* @param anIndex is the index of the key you want
* @return string containing key.
*/
virtual const nsString& GetKeyAt(PRInt32 anIndex) const =0;
/**
* Retrieve the value (of key/value pair) at given index
* @update gess5/11/98
* @param anIndex is the index of the value you want
* @return string containing value.
*/
virtual const nsString& GetValueAt(PRInt32 anIndex) const =0;
// When the node is an entity, this will translate the entity to
// it's unicode value.
/**
* NOTE: When the node is an entity, this will translate the entity
* to it's unicode value, and store it in aString.
* @update gess5/11/98
* @param aString will contain the resulting unicode string value
* @return int (unicode char or unicode index from table)
*/
virtual PRInt32 TranslateToUnicodeStr(nsString& aString) const = 0;
};

140
parser/htmlparser/src/CNavDelegate.h
Просмотреть файл

@ -46,33 +46,171 @@
class CNavDelegate : public ITokenizerDelegate {
public:
/**
* Default constructor
* @update gess5/11/98
*/
CNavDelegate();
/**
* Copy constructor
* @update gess 5/11/98
*/
CNavDelegate(CNavDelegate& aDelegate);
/**
* Consume next token from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 GetToken(CScanner& aScanner,CToken*& aToken);
/**
* Ask if its ok to add this token
* @update gess 5/11/98
* @param aToken is the token to be added
* @return True if ok to add the given token
*/
virtual PRBool WillAddToken(CToken& aToken);
/**
* Called as a preprocess -- tells delegate that tokenization will begin
* @update gess 5/11/98
* @param aIncremental tells us if tokenization is incremental
* @return TRUE if ok to continue -- FALSE if process should stop
*/
virtual PRBool WillTokenize(PRBool aIncremental);
/**
* Postprocess -- called to say that tokenization has concluded
* @update gess 5/11/98
* @param aIncremental tells us if tokenization was incremental
* @return TRUE if all went well--FALSE if you encountered an error
*/
virtual PRBool DidTokenize(PRBool aIncremental);
/**
* DEPRECATED. Tells us what mode the delegate is operating in.
* @update gess 5/11/98
* @return parse mode
*/
virtual eParseMode GetParseMode(void) const;
/**
* Retrieve the DTD required by this delegate
* (The parser will call this prior to tokenization)
* @update gess 5/11/98
* @return ptr to DTD -- should NOT be null.
*/
virtual nsIDTD* GetDTD(void) const;
/**
* Conduct self test.
* @update gess 5/11/98
*/
static void SelfTest();
protected:
/**
* Called to cause delegate to create a token of given type.
* @update gess 5/11/98
* @param aType represents the kind of token you want to create.
* @return new token or NULL
*/
virtual CToken* CreateTokenOfType(eHTMLTokenTypes aType);
/**
* Retrieve the next TAG from the given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve next START tag from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeStartTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve collection of HTML/XML attributes from given scanner
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeAttributes(PRUnichar aChar,CScanner& aScanner);
/**
* Retrieve a sequence of text from given scanner.
* @update gess 5/11/98
* @param aString will contain retrieved text.
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeText(const nsString& aString,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve an entity from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeEntity(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a whitespace sequence from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeWhitespace(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a comment from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeComment(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve newlines from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeNewline(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
//the only special case method...
/**
* Causes content to be skipped up to sequence contained in aString.
* @update gess 5/11/98
* @param aString ????
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 ConsumeContentToEndTag(const nsString& aString,PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
nsDeque mTokenDeque;

140
parser/htmlparser/src/COtherDelegate.h
Просмотреть файл

@ -46,33 +46,171 @@
class COtherDelegate : public ITokenizerDelegate {
public:
/**
* Default constructor
* @update gess5/11/98
*/
COtherDelegate();
/**
* Copy constructor
* @update gess 5/11/98
*/
COtherDelegate(COtherDelegate& aDelegate);
/**
* Consume next token from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 GetToken(CScanner& aScanner,CToken*& aToken);
/**
* Ask if its ok to add this token
* @update gess 5/11/98
* @param aToken is the token to be added
* @return True if ok to add the given token
*/
virtual PRBool WillAddToken(CToken& aToken);
/**
* Called as a preprocess -- tells delegate that tokenization will begin
* @update gess 5/11/98
* @param aIncremental tells us if tokenization is incremental
* @return TRUE if ok to continue -- FALSE if process should stop
*/
virtual PRBool WillTokenize(PRBool aIncremental);
/**
* Postprocess -- called to say that tokenization has concluded
* @update gess 5/11/98
* @param aIncremental tells us if tokenization was incremental
* @return TRUE if all went well--FALSE if you encountered an error
*/
virtual PRBool DidTokenize(PRBool aIncremental);
/**
* DEPRECATED. Tells us what mode the delegate is operating in.
* @update gess 5/11/98
* @return parse mode
*/
virtual eParseMode GetParseMode(void) const;
/**
* Retrieve the DTD required by this delegate
* (The parser will call this prior to tokenization)
* @update gess 5/11/98
* @return ptr to DTD -- should NOT be null.
*/
virtual nsIDTD* GetDTD(void) const;
/**
* Conduct self test.
* @update gess 5/11/98
*/
static void SelfTest();
protected:
/**
* Called to cause delegate to create a token of given type.
* @update gess 5/11/98
* @param aType represents the kind of token you want to create.
* @return new token or NULL
*/
virtual CToken* CreateTokenOfType(eHTMLTokenTypes aType);
/**
* Retrieve the next TAG from the given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve next START tag from given scanner.
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeStartTag(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve collection of HTML/XML attributes from given scanner
* @update gess 5/11/98
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeAttributes(PRUnichar aChar,CScanner& aScanner);
/**
* Retrieve a sequence of text from given scanner.
* @update gess 5/11/98
* @param aString will contain retrieved text.
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeText(const nsString& aString,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve an entity from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeEntity(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a whitespace sequence from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeWhitespace(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve a comment from the given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeComment(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
/**
* Retrieve newlines from given scanner
* @update gess 5/11/98
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
PRInt32 ConsumeNewline(PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
//the only special case method...
/**
* Causes content to be skipped up to sequence contained in aString.
* @update gess 5/11/98
* @param aString ????
* @param aChar last char read from scanner
* @param aScanner is the input source
* @param aToken is the next token (or null)
* @return error code
*/
virtual PRInt32 ConsumeContentToEndTag(const nsString& aString,PRUnichar aChar,CScanner& aScanner,CToken*& aToken);
nsDeque mTokenDeque;

18
parser/htmlparser/src/nsHTMLParser.cpp
Просмотреть файл

@ -879,7 +879,7 @@ PRBool nsHTMLParser::HandleSkippedContentToken(CToken* aToken) {
PRBool result=PR_TRUE;
if(IsWithinBody()) {
if(HasOpenContainer(eHTMLTag_body)) {
nsCParserNode aNode((CHTMLToken*)aToken);
result=AddLeaf(aNode);
}
@ -937,22 +937,6 @@ PRBool nsHTMLParser::HandleStyleToken(CToken* aToken){
return result;
}
/**
* This method gets called to determine whether a given
* tag is itself a container
*
* @update gess 3/25/98
* @param aTag -- tag to test for containership
* @return PR_TRUE if given tag can contain other tags
*/
PRBool nsHTMLParser::IsWithinBody(void) const {
for(int i=0;i<mContextStackPos;i++) {
if(eHTMLTag_body==mContextStack[i])
return PR_TRUE;
}
return PR_FALSE;
}
/**
* This method does two things: 1st, help construct

353
parser/htmlparser/src/nsHTMLParser.h
Просмотреть файл

@ -81,72 +81,417 @@ class nsHTMLParser : public nsIParser {
friend class CTokenHandler;
NS_DECL_ISUPPORTS
/**
* default constructor
* @update gess5/11/98
*/
nsHTMLParser();
/**
* Destructor
* @update gess5/11/98
*/
~nsHTMLParser();
/**
* Select given content sink into parser for parser output
* @update gess5/11/98
* @param aSink is the new sink to be used by parser
* @return old sink, or NULL
*/
virtual nsIContentSink* SetContentSink(nsIContentSink* aSink);
/**
* Cause parser to parse input from given URL
* @update gess5/11/98
* @param aURL is a descriptor for source document
* @return TRUE if all went well -- FALSE otherwise
*/
virtual PRBool Parse(nsIURL* aURL);
/**
* Cause parser to parse input from given URL in given mode
* @update gess5/11/98
* @param aURL is a descriptor for source document
* @param aMode is the desired parser mode (Nav, other, etc.)
* @return TRUE if all went well -- FALSE otherwise
*/
virtual PRBool Parse(nsIURL* aURL,eParseMode aMode);
/**
* This method gets called (automatically) during incremental parsing
* @update gess5/11/98
* @return TRUE if all went well, otherwise FALSE
*/
virtual PRBool ResumeParse();
/**
* Retrieve ptr to internal context vector stack
* @update gess5/11/98
* @param stack ptr to be set
* @return count of items on stack (may be 0)
*/
virtual PRInt32 GetStack(PRInt32* aStackPtr);
/**
* Ask parser if a given container is open ANYWHERE on stack
* @update gess5/11/98
* @param id of container you want to test for
* @return TRUE if the given container type is open -- otherwise FALSE
*/
virtual PRBool HasOpenContainer(PRInt32 aContainer) const;
/**
* This method gets called when a start token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the start token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleStartToken(CToken* aToken);
/**
* This method gets called when a start token has been consumed, and
* we want to use default start token handling behavior.
* This method gets called automatically by handleStartToken.
*
* @update gess5/11/98
* @param aToken is the start token to be handled
* @param aChildTag is the tag-type of given token
* @param aNode is a node be updated with info from given token
* @return TRUE if the token was handled.
*/
PRBool HandleDefaultStartToken(CToken* aToken,eHTMLTags aChildTag,nsCParserNode& aNode);
/**
* This method gets called when an end token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the end token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleEndToken(CToken* aToken);
/**
* This method gets called when an entity token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the entity token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleEntityToken(CToken* aToken);
/**
* This method gets called when a comment token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the comment token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleCommentToken(CToken* aToken);
/**
* This method gets called when a skipped-content token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the skipped-content token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleSkippedContentToken(CToken* aToken);
/**
* This method gets called when an attribute token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the attribute token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleAttributeToken(CToken* aToken);
/**
* This method gets called when a script token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the script token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleScriptToken(CToken* aToken);
/**
* This method gets called when a style token has been consumed and needs
* to be handled (possibly added to content model via sink).
* @update gess5/11/98
* @param aToken is the style token to be handled
* @return TRUE if the token was handled.
*/
PRBool HandleStyleToken(CToken* aToken);
PRBool IsWithinBody(void) const;
protected:
/**
* This method gets called when the tokens have been consumed, and it's time
* to build the model via the content sink.
* @update gess5/11/98
* @return YES if model building went well -- NO otherwise.
*/
PRBool IterateTokens();
/**
* Retrieves the tag type for the element on the context vector stack at given pos.
* @update gess5/11/98
* @param pos of item you want to retrieve
* @return tag type (may be unknown)
*/
eHTMLTags NodeAt(PRInt32 aPos) const;
/**
* Retrieve the tag type of the topmost item on context vector stack
* @update gess5/11/98
* @return tag type (may be unknown)
*/
eHTMLTags GetTopNode() const;
/**
* Retrieve the number of items on context vector stack
* @update gess5/11/98
* @return count of items on stack -- may be 0
*/
PRInt32 GetStackPos() const;
/**
* Causes the parser to scan foward, collecting nearby (sequential)
* attribute tokens into the given node.
* @update gess5/11/98
* @param node to store attributes
* @return number of attributes added to node.
*/
PRInt32 CollectAttributes(nsCParserNode& aNode);
/**
* Causes the next skipped-content token (if any) to
* be consumed by this node.
* @update gess5/11/98
* @param node to consume skipped-content
* @return number of skipped-content tokens consumed.
*/
PRInt32 CollectSkippedContent(nsCParserNode& aNode);
/**
* Causes token handlers to be registered for this parser.
* DO NOT CALL THIS! IT'S DEPRECATED!
* @update gess5/11/98
*/
void InitializeDefaultTokenHandlers();
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* GetTokenHandler(const nsString& aString) const;
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* GetTokenHandler(eHTMLTokenTypes aType) const;
/**
* DEPRECATED
* @update gess5/11/98
*/
CTokenHandler* AddTokenHandler(CTokenHandler* aHandler);
/**
* DEPRECATED
* @update gess5/11/98
*/
nsHTMLParser& DeleteTokenHandlers(void);
protected:
//*************************************************
//these cover methods mimic the sink, and are used
//by the parser to manage its context-stack.
//*************************************************
/**
* This cover method opens the given node as a HTML item in
* content sink.
* @update gess5/11/98
* @param HTML (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenHTML(const nsIParserNode& aNode);
/**
*
* @update gess5/11/98
* @param
* @return
*/
PRBool CloseHTML(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a head item in
* content sink.
* @update gess5/11/98
* @param HEAD (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenHead(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink head to be closed
* @update gess5/11/98
* @param aNode is the node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseHead(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a body item in
* content sink.
* @update gess5/11/98
* @param BODY (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenBody(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink body to be closed
* @update gess5/11/98
* @param aNode is the body node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseBody(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a form item in
* content sink.
* @update gess5/11/98
* @param FORM (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenForm(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink form to be closed
* @update gess5/11/98
* @param aNode is the form node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseForm(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a frameset item in
* content sink.
* @update gess5/11/98
* @param FRAMESET (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenFrameset(const nsIParserNode& aNode);
/**
* This cover method causes the content-sink frameset to be closed
* @update gess5/11/98
* @param aNode is the frameeset node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseFrameset(const nsIParserNode& aNode);
/**
* This cover method opens the given node as a generic container in
* content sink.
* @update gess5/11/98
* @param generic container (node) to be opened in content sink.
* @return TRUE if all went well.
*/
PRBool OpenContainer(const nsIParserNode& aNode);
/**
* This cover method causes a generic containre in the content-sink to be closed
* @update gess5/11/98
* @param aNode is the node to be closed in sink (usually ignored)
* @return TRUE if all went well.
*/
PRBool CloseContainer(const nsIParserNode& aNode);
/**
* This cover method causes the topmost container to be closed in sink
* @update gess5/11/98
* @return TRUE if all went well.
*/
PRBool CloseTopmostContainer();
/**
* Cause all containers down to topmost given tag to be closed
* @update gess5/11/98
* @param aTag is the tag at which auto-closure should stop (inclusive)
* @return TRUE if all went well -- otherwise FALSE
*/
PRBool CloseContainersTo(eHTMLTags aTag);
/**
* Cause all containers down to given position to be closed
* @update gess5/11/98
* @param anIndex is the stack pos at which auto-closure should stop (inclusive)
* @return TRUE if all went well -- otherwise FALSE
*/
PRBool CloseContainersTo(PRInt32 anIndex);
/**
* Causes leaf to be added to sink at current vector pos.
* @update gess5/11/98
* @param aNode is leaf node to be added.
* @return TRUE if all went well -- FALSE otherwise.
*/
PRBool AddLeaf(const nsIParserNode& aNode);
/**
* Determine if the given tag is open in the context vector stack.
* @update gess5/11/98
* @param aTag is the type of tag you want to test for openness
* @return TRUE if open, FALSE otherwise.
*/
PRBool IsOpen(eHTMLTags aTag) const;
/**
* Finds the topmost occurance of given tag within context vector stack.
* @update gess5/11/98
* @param tag to be found
* @return index of topmost tag occurance -- may be -1 (kNotFound).
*/
PRInt32 GetTopmostIndex(eHTMLTags aTag) const;
/**
* Causes auto-closures of context vector stack in order to find a
* proper home for the given child. Propagation may also occur as
* a fall out.
* @update gess5/11/98
* @param child to be added (somewhere) to context vector stack.
* @return TRUE if succeeds, otherwise FALSE
*/
PRBool ReduceContextStackFor(PRInt32 aChildTag);
/**
* Attempt forward and/or backward propagation for the given
* child within the current context vector stack.
* @update gess5/11/98
* @param type of child to be propagated.
* @return TRUE if succeeds, otherwise FALSE
*/
PRBool CreateContextStackFor(PRInt32 aChildTag);
//*********************************************
// And now, some data members...
//*********************************************
nsIHTMLContentSink* mSink;
CTokenizer* mTokenizer;
// eHTMLTags mContextStack[50];
PRInt32 mContextStack[50];
PRInt32 mContextStackPos;

59
parser/htmlparser/src/nsIParserNode.h
Просмотреть файл

@ -52,21 +52,72 @@ class nsIParserNode {
public:
/**
* Retrieve the name of the node
* @update gess5/11/98
* @return string containing node name
*/
virtual const nsString& GetName() const =0; //to get name of tag
/**
* Retrieve the text from the given node
* @update gess5/11/98
* @return string containing node text
*/
virtual const nsString& GetText() const =0; //get plain text if available
/**
* Retrieve skipped context from node
* @update gess5/11/98
* @return string containing skipped content
*/
virtual const nsString& GetSkippedContent() const =0;
//methods for determining the type of parser node...
/**
* Retrieve the type of the parser node.
* @update gess5/11/98
* @return node type.
*/
virtual PRInt32 GetNodeType() const =0;
/**
* Retrieve token type of parser node
* @update gess5/11/98
* @return token type
*/
virtual PRInt32 GetTokenType() const =0;
//methods for accessing key/value pairs
/**
* Retrieve the number of attributes in this node.
* @update gess5/11/98
* @return count of attributes (may be 0)
*/
virtual PRInt32 GetAttributeCount(void) const =0;
/**
* Retrieve the key (of key/value pair) at given index
* @update gess5/11/98
* @param anIndex is the index of the key you want
* @return string containing key.
*/
virtual const nsString& GetKeyAt(PRInt32 anIndex) const =0;
/**
* Retrieve the value (of key/value pair) at given index
* @update gess5/11/98
* @param anIndex is the index of the value you want
* @return string containing value.
*/
virtual const nsString& GetValueAt(PRInt32 anIndex) const =0;
// When the node is an entity, this will translate the entity to
// it's unicode value.
/**
* NOTE: When the node is an entity, this will translate the entity
* to it's unicode value, and store it in aString.
* @update gess5/11/98
* @param aString will contain the resulting unicode string value
* @return int (unicode char or unicode index from table)
*/
virtual PRInt32 TranslateToUnicodeStr(nsString& aString) const = 0;
};