зеркало из https://github.com/mozilla/gecko-dev.git
Bug 110544, improving comment parsing by avioding string copies and other trickery. r=harishd, sr=jst, a=asa.
This commit is contained in:
heikki%netscape.com
Родитель
bfb4027790
Коммит
3116e5c047
|
|
@ -75,10 +75,6 @@ public:
|
|||
#endif
|
||||
|
||||
// nsITextContent
|
||||
NS_IMETHOD SetText(const PRUnichar* aBuffer, PRInt32 aLength,
|
||||
PRBool aNotify);
|
||||
NS_IMETHOD SetText(const nsAReadableString& aStr, PRBool aNotify);
|
||||
NS_IMETHOD SetText(const char* aBuffer, PRInt32 aLength, PRBool aNotify);
|
||||
NS_IMETHOD CloneContent(PRBool aCloneText, nsITextContent** aClone);
|
||||
};
|
||||
|
||||
|
|
@ -185,75 +181,3 @@ nsCommentNode::List(FILE* out, PRInt32 aIndent) const
|
|||
return NS_OK;
|
||||
}
|
||||
#endif
|
||||
|
||||
// This would ideally be done by the parser, but for the sake
|
||||
// of "genericity" it's being done in the comment content code
|
||||
static void
|
||||
StripCommentDelimiters(nsString& aCommentString)
|
||||
{
|
||||
PRInt32 offset;
|
||||
static const char kCommentStart[] = "<!";
|
||||
static const char kCommentEnd[] = "->";
|
||||
static const char kCommentAlternateEnd[] = "--!>";
|
||||
static const char kMinus = '-';
|
||||
|
||||
offset = aCommentString.Find(kCommentStart);
|
||||
if (-1 != offset) {
|
||||
// Take up to 2 '-' characters
|
||||
offset += strlen(kCommentStart);
|
||||
if (kMinus == aCommentString.CharAt(offset)) {
|
||||
offset++;
|
||||
if (kMinus == aCommentString.CharAt(offset)) {
|
||||
offset++;
|
||||
}
|
||||
}
|
||||
aCommentString.Cut(0, offset);
|
||||
}
|
||||
|
||||
offset = aCommentString.RFind(kCommentEnd);
|
||||
if (offset > 0) {
|
||||
// Take up to 1 more '-'
|
||||
if (kMinus == aCommentString.CharAt(offset-1)) {
|
||||
offset--;
|
||||
}
|
||||
aCommentString.Cut(offset, aCommentString.Length()-offset);
|
||||
}
|
||||
else {
|
||||
offset = aCommentString.RFind(kCommentAlternateEnd);
|
||||
if (-1 != offset) {
|
||||
aCommentString.Cut(offset, aCommentString.Length()-offset);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
NS_IMETHODIMP
|
||||
nsCommentNode::SetText(const PRUnichar* aBuffer,
|
||||
PRInt32 aLength,
|
||||
PRBool aNotify)
|
||||
{
|
||||
nsAutoString str(aBuffer);
|
||||
|
||||
StripCommentDelimiters(str);
|
||||
return nsGenericDOMDataNode::SetText(str, aNotify);
|
||||
}
|
||||
|
||||
NS_IMETHODIMP
|
||||
nsCommentNode::SetText(const nsAReadableString& aStr,
|
||||
PRBool aNotify)
|
||||
{
|
||||
nsAutoString str(aStr);
|
||||
|
||||
StripCommentDelimiters(str);
|
||||
return nsGenericDOMDataNode::SetText(str, aNotify);
|
||||
}
|
||||
|
||||
NS_IMETHODIMP
|
||||
nsCommentNode::SetText(const char* aBuffer,
|
||||
PRInt32 aLength,
|
||||
PRBool aNotify)
|
||||
{
|
||||
nsAutoString str; str.AssignWithConversion(aBuffer);
|
||||
|
||||
StripCommentDelimiters(str);
|
||||
return nsGenericDOMDataNode::SetText(str, aNotify);
|
||||
}
|
||||
|
|
|
|||
|
|
@ -1005,172 +1005,223 @@ CCommentToken::CCommentToken(const nsAReadableString& aName) : CHTMLToken(eHTMLT
|
|||
mTextValue.Assign(aName);
|
||||
}
|
||||
|
||||
/*
|
||||
* This method consumes a comment using the (CORRECT) comment parsing
|
||||
* algorithm according to ISO 8879 (SGML).
|
||||
*
|
||||
* @update gess 01/04/99
|
||||
* @param
|
||||
* @param
|
||||
* @return
|
||||
*/
|
||||
// We must test if we have whitespace and >
|
||||
static PRBool IsGoodCommentEnd(
|
||||
const nsReadingIterator<PRUnichar>& aCurrent,
|
||||
const nsReadingIterator<PRUnichar>& aEnd,
|
||||
nsReadingIterator<PRUnichar>& aGt)
|
||||
{
|
||||
nsReadingIterator<PRUnichar> current = aCurrent;
|
||||
|
||||
while (current != aEnd) {
|
||||
if (*current == kGreaterThan) {
|
||||
aGt = current;
|
||||
return PR_TRUE;
|
||||
}
|
||||
if (!nsString::IsSpace(*current))
|
||||
break;
|
||||
++current;
|
||||
}
|
||||
|
||||
return PR_FALSE;
|
||||
}
|
||||
|
||||
static
|
||||
nsresult ConsumeStrictComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString) {
|
||||
nsresult result=NS_OK;
|
||||
static const PRUnichar theTerminalsChars[] =
|
||||
{ PRUnichar('-'), PRUnichar('>'), PRUnichar(0) };
|
||||
static const nsReadEndCondition endCommentCondition(theTerminalsChars);
|
||||
|
||||
nsresult ConsumeStrictComment(nsScanner& aScanner, nsString& aString) {
|
||||
aString.Truncate();
|
||||
|
||||
// <!--[... -- ... -- ...]*-->
|
||||
/*********************************************************
|
||||
NOTE: This algorithm does a fine job of handling comments
|
||||
when they're formatted per spec, but if they're not
|
||||
we don't handle them well. For example, we gack
|
||||
on the following:
|
||||
|
||||
<!-- xx -- xx -->
|
||||
we don't handle them well.
|
||||
*********************************************************/
|
||||
nsReadingIterator<PRUnichar> end, current, gt;
|
||||
aScanner.EndReading(end);
|
||||
aScanner.CurrentPosition(current);
|
||||
|
||||
aString.Assign(NS_LITERAL_STRING("<!"));
|
||||
while(NS_OK==result) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
aString+=aChar;
|
||||
if(kMinus==aChar) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
//in this case, we're reading a long-form comment <-- xxx -->
|
||||
aString+=aChar;
|
||||
if(NS_OK==result) {
|
||||
//Find the first ending sequence '--'
|
||||
PRBool found = PR_FALSE;
|
||||
nsAutoString temp;
|
||||
do {
|
||||
result = aScanner.ReadUntil(temp,kMinus,PR_TRUE);
|
||||
if (NS_FAILED(result)) return result;
|
||||
|
||||
result = aScanner.Peek(aChar);
|
||||
if (NS_FAILED(result)) return result;
|
||||
|
||||
if (aChar == kMinus) {
|
||||
// found matching '--'
|
||||
found = PR_TRUE;
|
||||
aScanner.GetChar(aChar);
|
||||
aString += temp;
|
||||
aString += aChar;
|
||||
// Read until "-" or ">"
|
||||
result = aScanner.ReadUntil(aString,endCommentCondition,PR_FALSE);
|
||||
}
|
||||
} while (!found);
|
||||
}
|
||||
} //
|
||||
else break; //go find '>'
|
||||
}
|
||||
}//if
|
||||
else if(kGreaterThan==aChar) {
|
||||
return result;
|
||||
nsReadingIterator<PRUnichar> currentEnd = end, beginData = end;
|
||||
PRBool balancedComment = PR_TRUE;
|
||||
|
||||
static NS_NAMED_LITERAL_STRING(dashes,"--");
|
||||
|
||||
while (FindInReadable(dashes, current, currentEnd)) {
|
||||
current.advance(2);
|
||||
if (beginData == end) {
|
||||
beginData = current;
|
||||
}
|
||||
balancedComment = !balancedComment; // We need to match '--' with '--'
|
||||
|
||||
if (balancedComment && IsGoodCommentEnd(current, end, gt)) {
|
||||
// done
|
||||
current.advance(-2);
|
||||
if (beginData != current) { // protects from <!---->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
else break; //go find '>'
|
||||
}//if
|
||||
}//while
|
||||
if(NS_OK==result) {
|
||||
//Read up to the closing '>', unless you already did! (such as <!>).
|
||||
if(kGreaterThan!=aChar) {
|
||||
result=aScanner.ReadUntil(aString,kGreaterThan,PR_TRUE);
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
// Continue after the last '--'
|
||||
currentEnd = end;
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
|
||||
/*
|
||||
* This method consumes a comment using common (actually non-standard)
|
||||
* algorithm that seems to work against the content on the web.
|
||||
*
|
||||
* @update gess 01/04/99
|
||||
* @param
|
||||
* @param
|
||||
* @return
|
||||
*/
|
||||
static
|
||||
nsresult ConsumeComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString) {
|
||||
|
||||
nsresult ConsumeComment(nsScanner& aScanner, nsString& aString) {
|
||||
aString.Truncate();
|
||||
|
||||
nsresult result=NS_OK;
|
||||
|
||||
// <![-[-]] ... [[-]-|--!]>
|
||||
/*********************************************************
|
||||
NOTE: This algorithm does a fine job of handling comments
|
||||
commonly used, but it doesn't really consume them
|
||||
per spec (But then, neither does IE or Nav).
|
||||
*********************************************************/
|
||||
nsReadingIterator<PRUnichar> end, current;
|
||||
aScanner.EndReading(end);
|
||||
aScanner.CurrentPosition(current);
|
||||
nsReadingIterator<PRUnichar> beginData = current, beginLastMinus = end;
|
||||
|
||||
nsReadingIterator<PRUnichar> theStartOffset, theCurrOffset, theBestAltPos, endPos;
|
||||
aScanner.EndReading(endPos);
|
||||
theBestAltPos = endPos;
|
||||
// When we get here, we have always already consumed <!
|
||||
// Skip over possible leading minuses
|
||||
if (current != end && *current == kMinus) {
|
||||
beginLastMinus = current;
|
||||
++current;
|
||||
++beginData;
|
||||
if (current != end && *current == kMinus) { // <!--
|
||||
beginLastMinus = current;
|
||||
++current;
|
||||
++beginData;
|
||||
// Long form comment
|
||||
|
||||
aScanner.CurrentPosition(theStartOffset);
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
//in this case, we're reading a long-form comment <-- xxx -->
|
||||
|
||||
theCurrOffset = theStartOffset;
|
||||
|
||||
while((NS_OK==result)) {
|
||||
if (FindCharInReadable(PRUnichar(kGreaterThan), theCurrOffset, endPos)) {
|
||||
++theCurrOffset;
|
||||
nsReadingIterator<PRUnichar> temp = theCurrOffset;
|
||||
temp.advance(-3);
|
||||
aChar=*temp;
|
||||
if(kMinus==aChar) {
|
||||
++temp;
|
||||
aChar=*temp;
|
||||
if(kMinus==aChar) {
|
||||
theStartOffset.advance(-2); // Include "<!" also..
|
||||
CopyUnicodeTo(theStartOffset, theCurrOffset, aString);
|
||||
aScanner.SetPosition(theCurrOffset);
|
||||
return result; // We have found the dflt end comment delimiter ("-->")
|
||||
}
|
||||
}
|
||||
if(theBestAltPos == endPos) {
|
||||
// If we did not find the dflt then assume that '>' is the end comment
|
||||
// until we find '-->'. Nav. Compatibility -- Ref: Bug# 24006
|
||||
theBestAltPos=theCurrOffset;
|
||||
}
|
||||
}
|
||||
else {
|
||||
result=kEOF;
|
||||
}
|
||||
} //while
|
||||
if((endPos==theCurrOffset) && (!aScanner.IsIncremental())) {
|
||||
//if you're here, then we're in a special state.
|
||||
//The problem at hand is that we've hit the end of the document without finding the normal endcomment delimiter "-->".
|
||||
//In this case, the first thing we try is to see if we found one of the alternate endcomment delimiter ">".
|
||||
//If so, rewind just pass than, and use everything up to that point as your comment.
|
||||
//If not, the document has no end comment and should be treated as one big comment.
|
||||
if(endPos != theBestAltPos) {
|
||||
theStartOffset.advance(-2);// Include "<!" also..
|
||||
CopyUnicodeTo(theStartOffset, theBestAltPos, aString);
|
||||
aScanner.SetPosition(theBestAltPos);
|
||||
result=NS_OK;
|
||||
nsReadingIterator<PRUnichar> currentEnd = end, gt = end;
|
||||
|
||||
// Find the end of the comment
|
||||
while (FindCharInReadable(kGreaterThan, current, currentEnd)) {
|
||||
gt = current;
|
||||
--current;
|
||||
PRBool goodComment = PR_FALSE;
|
||||
if (current != beginLastMinus && *current == kMinus) { // ->
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) { // -->
|
||||
goodComment = PR_TRUE;
|
||||
--current;
|
||||
}
|
||||
} else if (current != beginLastMinus && *current == '!') {
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) {
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) { // --!>
|
||||
--current;
|
||||
goodComment = PR_TRUE;
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
} //if
|
||||
}//if
|
||||
}//if
|
||||
}//if
|
||||
if(NS_OK==result) {
|
||||
//Read up to the closing '>', unless you already did! (such as <!>).
|
||||
if(kGreaterThan!=aChar) {
|
||||
aString.Append(NS_LITERAL_STRING("<!- "));
|
||||
result=aScanner.ReadUntil(aString,kGreaterThan,PR_TRUE);
|
||||
} else if (current == beginLastMinus) {
|
||||
goodComment = PR_TRUE;
|
||||
}
|
||||
|
||||
if (goodComment) {
|
||||
// done
|
||||
if (beginLastMinus != current) { // protects from <!---->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
// try again starting after the last '>'
|
||||
current = ++gt;
|
||||
currentEnd = end;
|
||||
}
|
||||
} //while
|
||||
|
||||
if (gt != end && !aScanner.IsIncremental()) {
|
||||
// If you're here, then we're in a special state.
|
||||
// The problem at hand is that we've hit the end of the document without finding the normal endcomment delimiter "-->".
|
||||
// In this case, the first thing we try is to see if we found one of the alternate endcomment delimiter ">".
|
||||
// If so, rewind just pass that, and use everything up to that point as your comment.
|
||||
// If not, the document has no end comment and should be treated as one big comment.
|
||||
if (beginData != gt) { // protects from <!-->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++gt);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++gt);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, gt, dest);
|
||||
#endif
|
||||
} else {
|
||||
++gt;
|
||||
}
|
||||
aScanner.SetPosition(gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
// This could be short form of comment
|
||||
// Find the end of the comment
|
||||
current = beginData;
|
||||
if (FindCharInReadable(kGreaterThan, current, end)) {
|
||||
nsReadingIterator<PRUnichar> gt = current;
|
||||
if (current != beginData) {
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // ->
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // -->
|
||||
--current;
|
||||
}
|
||||
} else if (current != beginData && *current == '!') { // !>
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // -!>
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // --!>
|
||||
--current;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (current != gt) {
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
}
|
||||
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
|
||||
/*
|
||||
|
|
@ -1187,10 +1238,10 @@ nsresult CCommentToken::Consume(PRUnichar aChar, nsScanner& aScanner,PRInt32 aFl
|
|||
|
||||
if (aFlag & NS_IPARSER_FLAG_STRICT_MODE) {
|
||||
//Enabling strict comment parsing for Bug 53011 and 2749 contradicts!!!!
|
||||
result=ConsumeStrictComment(aChar,aScanner,mTextValue);
|
||||
result=ConsumeStrictComment(aScanner,mTextValue);
|
||||
}
|
||||
else {
|
||||
result=ConsumeComment(aChar,aScanner,mTextValue);
|
||||
result=ConsumeComment(aScanner,mTextValue);
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -278,8 +278,9 @@ void nsScanner::RewindToMark(void){
|
|||
*/
|
||||
void nsScanner::Mark() {
|
||||
if (mSlidingBuffer) {
|
||||
mSlidingBuffer->DiscardPrefix(mCurrentPosition);
|
||||
mSlidingBuffer->BeginReading(mCurrentPosition);
|
||||
mMarkPosition = mCurrentPosition;
|
||||
mSlidingBuffer->DiscardPrefix(mMarkPosition);
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1169,8 +1169,11 @@ NS_IMETHODIMP CViewSourceHTML::HandleToken(CToken* aToken,nsIParser* aParser) {
|
|||
|
||||
case eToken_comment:
|
||||
{
|
||||
const nsAReadableString& commentValue = aToken->GetStringValue();
|
||||
result=WriteTag(mCommentTag,commentValue,0,PR_TRUE);
|
||||
nsAutoString theStr;
|
||||
theStr.Assign(NS_LITERAL_STRING("<!--") +
|
||||
aToken->GetStringValue() +
|
||||
NS_LITERAL_STRING("-->") );
|
||||
result=WriteTag(mCommentTag,theStr,0,PR_TRUE);
|
||||
}
|
||||
break;
|
||||
|
||||
|
|
|
|||
|
|
@ -1005,172 +1005,223 @@ CCommentToken::CCommentToken(const nsAReadableString& aName) : CHTMLToken(eHTMLT
|
|||
mTextValue.Assign(aName);
|
||||
}
|
||||
|
||||
/*
|
||||
* This method consumes a comment using the (CORRECT) comment parsing
|
||||
* algorithm according to ISO 8879 (SGML).
|
||||
*
|
||||
* @update gess 01/04/99
|
||||
* @param
|
||||
* @param
|
||||
* @return
|
||||
*/
|
||||
// We must test if we have whitespace and >
|
||||
static PRBool IsGoodCommentEnd(
|
||||
const nsReadingIterator<PRUnichar>& aCurrent,
|
||||
const nsReadingIterator<PRUnichar>& aEnd,
|
||||
nsReadingIterator<PRUnichar>& aGt)
|
||||
{
|
||||
nsReadingIterator<PRUnichar> current = aCurrent;
|
||||
|
||||
while (current != aEnd) {
|
||||
if (*current == kGreaterThan) {
|
||||
aGt = current;
|
||||
return PR_TRUE;
|
||||
}
|
||||
if (!nsString::IsSpace(*current))
|
||||
break;
|
||||
++current;
|
||||
}
|
||||
|
||||
return PR_FALSE;
|
||||
}
|
||||
|
||||
static
|
||||
nsresult ConsumeStrictComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString) {
|
||||
nsresult result=NS_OK;
|
||||
static const PRUnichar theTerminalsChars[] =
|
||||
{ PRUnichar('-'), PRUnichar('>'), PRUnichar(0) };
|
||||
static const nsReadEndCondition endCommentCondition(theTerminalsChars);
|
||||
|
||||
nsresult ConsumeStrictComment(nsScanner& aScanner, nsString& aString) {
|
||||
aString.Truncate();
|
||||
|
||||
// <!--[... -- ... -- ...]*-->
|
||||
/*********************************************************
|
||||
NOTE: This algorithm does a fine job of handling comments
|
||||
when they're formatted per spec, but if they're not
|
||||
we don't handle them well. For example, we gack
|
||||
on the following:
|
||||
|
||||
<!-- xx -- xx -->
|
||||
we don't handle them well.
|
||||
*********************************************************/
|
||||
nsReadingIterator<PRUnichar> end, current, gt;
|
||||
aScanner.EndReading(end);
|
||||
aScanner.CurrentPosition(current);
|
||||
|
||||
aString.Assign(NS_LITERAL_STRING("<!"));
|
||||
while(NS_OK==result) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
aString+=aChar;
|
||||
if(kMinus==aChar) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
//in this case, we're reading a long-form comment <-- xxx -->
|
||||
aString+=aChar;
|
||||
if(NS_OK==result) {
|
||||
//Find the first ending sequence '--'
|
||||
PRBool found = PR_FALSE;
|
||||
nsAutoString temp;
|
||||
do {
|
||||
result = aScanner.ReadUntil(temp,kMinus,PR_TRUE);
|
||||
if (NS_FAILED(result)) return result;
|
||||
|
||||
result = aScanner.Peek(aChar);
|
||||
if (NS_FAILED(result)) return result;
|
||||
|
||||
if (aChar == kMinus) {
|
||||
// found matching '--'
|
||||
found = PR_TRUE;
|
||||
aScanner.GetChar(aChar);
|
||||
aString += temp;
|
||||
aString += aChar;
|
||||
// Read until "-" or ">"
|
||||
result = aScanner.ReadUntil(aString,endCommentCondition,PR_FALSE);
|
||||
}
|
||||
} while (!found);
|
||||
}
|
||||
} //
|
||||
else break; //go find '>'
|
||||
}
|
||||
}//if
|
||||
else if(kGreaterThan==aChar) {
|
||||
return result;
|
||||
nsReadingIterator<PRUnichar> currentEnd = end, beginData = end;
|
||||
PRBool balancedComment = PR_TRUE;
|
||||
|
||||
static NS_NAMED_LITERAL_STRING(dashes,"--");
|
||||
|
||||
while (FindInReadable(dashes, current, currentEnd)) {
|
||||
current.advance(2);
|
||||
if (beginData == end) {
|
||||
beginData = current;
|
||||
}
|
||||
balancedComment = !balancedComment; // We need to match '--' with '--'
|
||||
|
||||
if (balancedComment && IsGoodCommentEnd(current, end, gt)) {
|
||||
// done
|
||||
current.advance(-2);
|
||||
if (beginData != current) { // protects from <!---->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
else break; //go find '>'
|
||||
}//if
|
||||
}//while
|
||||
if(NS_OK==result) {
|
||||
//Read up to the closing '>', unless you already did! (such as <!>).
|
||||
if(kGreaterThan!=aChar) {
|
||||
result=aScanner.ReadUntil(aString,kGreaterThan,PR_TRUE);
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
// Continue after the last '--'
|
||||
currentEnd = end;
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
|
||||
/*
|
||||
* This method consumes a comment using common (actually non-standard)
|
||||
* algorithm that seems to work against the content on the web.
|
||||
*
|
||||
* @update gess 01/04/99
|
||||
* @param
|
||||
* @param
|
||||
* @return
|
||||
*/
|
||||
static
|
||||
nsresult ConsumeComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString) {
|
||||
|
||||
nsresult ConsumeComment(nsScanner& aScanner, nsString& aString) {
|
||||
aString.Truncate();
|
||||
|
||||
nsresult result=NS_OK;
|
||||
|
||||
// <![-[-]] ... [[-]-|--!]>
|
||||
/*********************************************************
|
||||
NOTE: This algorithm does a fine job of handling comments
|
||||
commonly used, but it doesn't really consume them
|
||||
per spec (But then, neither does IE or Nav).
|
||||
*********************************************************/
|
||||
nsReadingIterator<PRUnichar> end, current;
|
||||
aScanner.EndReading(end);
|
||||
aScanner.CurrentPosition(current);
|
||||
nsReadingIterator<PRUnichar> beginData = current, beginLastMinus = end;
|
||||
|
||||
nsReadingIterator<PRUnichar> theStartOffset, theCurrOffset, theBestAltPos, endPos;
|
||||
aScanner.EndReading(endPos);
|
||||
theBestAltPos = endPos;
|
||||
// When we get here, we have always already consumed <!
|
||||
// Skip over possible leading minuses
|
||||
if (current != end && *current == kMinus) {
|
||||
beginLastMinus = current;
|
||||
++current;
|
||||
++beginData;
|
||||
if (current != end && *current == kMinus) { // <!--
|
||||
beginLastMinus = current;
|
||||
++current;
|
||||
++beginData;
|
||||
// Long form comment
|
||||
|
||||
aScanner.CurrentPosition(theStartOffset);
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
result=aScanner.GetChar(aChar);
|
||||
if(NS_OK==result) {
|
||||
if(kMinus==aChar) {
|
||||
//in this case, we're reading a long-form comment <-- xxx -->
|
||||
|
||||
theCurrOffset = theStartOffset;
|
||||
|
||||
while((NS_OK==result)) {
|
||||
if (FindCharInReadable(PRUnichar(kGreaterThan), theCurrOffset, endPos)) {
|
||||
++theCurrOffset;
|
||||
nsReadingIterator<PRUnichar> temp = theCurrOffset;
|
||||
temp.advance(-3);
|
||||
aChar=*temp;
|
||||
if(kMinus==aChar) {
|
||||
++temp;
|
||||
aChar=*temp;
|
||||
if(kMinus==aChar) {
|
||||
theStartOffset.advance(-2); // Include "<!" also..
|
||||
CopyUnicodeTo(theStartOffset, theCurrOffset, aString);
|
||||
aScanner.SetPosition(theCurrOffset);
|
||||
return result; // We have found the dflt end comment delimiter ("-->")
|
||||
}
|
||||
}
|
||||
if(theBestAltPos == endPos) {
|
||||
// If we did not find the dflt then assume that '>' is the end comment
|
||||
// until we find '-->'. Nav. Compatibility -- Ref: Bug# 24006
|
||||
theBestAltPos=theCurrOffset;
|
||||
}
|
||||
}
|
||||
else {
|
||||
result=kEOF;
|
||||
}
|
||||
} //while
|
||||
if((endPos==theCurrOffset) && (!aScanner.IsIncremental())) {
|
||||
//if you're here, then we're in a special state.
|
||||
//The problem at hand is that we've hit the end of the document without finding the normal endcomment delimiter "-->".
|
||||
//In this case, the first thing we try is to see if we found one of the alternate endcomment delimiter ">".
|
||||
//If so, rewind just pass than, and use everything up to that point as your comment.
|
||||
//If not, the document has no end comment and should be treated as one big comment.
|
||||
if(endPos != theBestAltPos) {
|
||||
theStartOffset.advance(-2);// Include "<!" also..
|
||||
CopyUnicodeTo(theStartOffset, theBestAltPos, aString);
|
||||
aScanner.SetPosition(theBestAltPos);
|
||||
result=NS_OK;
|
||||
nsReadingIterator<PRUnichar> currentEnd = end, gt = end;
|
||||
|
||||
// Find the end of the comment
|
||||
while (FindCharInReadable(kGreaterThan, current, currentEnd)) {
|
||||
gt = current;
|
||||
--current;
|
||||
PRBool goodComment = PR_FALSE;
|
||||
if (current != beginLastMinus && *current == kMinus) { // ->
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) { // -->
|
||||
goodComment = PR_TRUE;
|
||||
--current;
|
||||
}
|
||||
} else if (current != beginLastMinus && *current == '!') {
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) {
|
||||
--current;
|
||||
if (current != beginLastMinus && *current == kMinus) { // --!>
|
||||
--current;
|
||||
goodComment = PR_TRUE;
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
} //if
|
||||
}//if
|
||||
}//if
|
||||
}//if
|
||||
if(NS_OK==result) {
|
||||
//Read up to the closing '>', unless you already did! (such as <!>).
|
||||
if(kGreaterThan!=aChar) {
|
||||
aString.Append(NS_LITERAL_STRING("<!- "));
|
||||
result=aScanner.ReadUntil(aString,kGreaterThan,PR_TRUE);
|
||||
} else if (current == beginLastMinus) {
|
||||
goodComment = PR_TRUE;
|
||||
}
|
||||
|
||||
if (goodComment) {
|
||||
// done
|
||||
if (beginLastMinus != current) { // protects from <!---->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
// try again starting after the last '>'
|
||||
current = ++gt;
|
||||
currentEnd = end;
|
||||
}
|
||||
} //while
|
||||
|
||||
if (gt != end && !aScanner.IsIncremental()) {
|
||||
// If you're here, then we're in a special state.
|
||||
// The problem at hand is that we've hit the end of the document without finding the normal endcomment delimiter "-->".
|
||||
// In this case, the first thing we try is to see if we found one of the alternate endcomment delimiter ">".
|
||||
// If so, rewind just pass that, and use everything up to that point as your comment.
|
||||
// If not, the document has no end comment and should be treated as one big comment.
|
||||
if (beginData != gt) { // protects from <!-->
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++gt);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++gt);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, gt, dest);
|
||||
#endif
|
||||
} else {
|
||||
++gt;
|
||||
}
|
||||
aScanner.SetPosition(gt);
|
||||
return NS_OK;
|
||||
} else {
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
}
|
||||
}
|
||||
return result;
|
||||
|
||||
// This could be short form of comment
|
||||
// Find the end of the comment
|
||||
current = beginData;
|
||||
if (FindCharInReadable(kGreaterThan, current, end)) {
|
||||
nsReadingIterator<PRUnichar> gt = current;
|
||||
if (current != beginData) {
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // ->
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // -->
|
||||
--current;
|
||||
}
|
||||
} else if (current != beginData && *current == '!') { // !>
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // -!>
|
||||
--current;
|
||||
if (current != beginData && *current == kMinus) { // --!>
|
||||
--current;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (current != gt) {
|
||||
#if 0
|
||||
// XXX We should do this, but it HANGS until bug 112943 is fixed:
|
||||
aString = Substring(beginData, ++current);
|
||||
#else
|
||||
// XXX Instead we can do this EVIL HACK (from jag):
|
||||
PRUint32 len = Distance(beginData, ++current);
|
||||
aString.SetLength(len);
|
||||
PRUnichar* dest = NS_CONST_CAST(PRUnichar*, aString.get());
|
||||
copy_string(beginData, current, dest);
|
||||
#endif
|
||||
}
|
||||
aScanner.SetPosition(++gt);
|
||||
return NS_OK;
|
||||
}
|
||||
|
||||
return kEOF; // not really an nsresult, but...
|
||||
}
|
||||
|
||||
/*
|
||||
|
|
@ -1187,10 +1238,10 @@ nsresult CCommentToken::Consume(PRUnichar aChar, nsScanner& aScanner,PRInt32 aFl
|
|||
|
||||
if (aFlag & NS_IPARSER_FLAG_STRICT_MODE) {
|
||||
//Enabling strict comment parsing for Bug 53011 and 2749 contradicts!!!!
|
||||
result=ConsumeStrictComment(aChar,aScanner,mTextValue);
|
||||
result=ConsumeStrictComment(aScanner,mTextValue);
|
||||
}
|
||||
else {
|
||||
result=ConsumeComment(aChar,aScanner,mTextValue);
|
||||
result=ConsumeComment(aScanner,mTextValue);
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -278,8 +278,9 @@ void nsScanner::RewindToMark(void){
|
|||
*/
|
||||
void nsScanner::Mark() {
|
||||
if (mSlidingBuffer) {
|
||||
mSlidingBuffer->DiscardPrefix(mCurrentPosition);
|
||||
mSlidingBuffer->BeginReading(mCurrentPosition);
|
||||
mMarkPosition = mCurrentPosition;
|
||||
mSlidingBuffer->DiscardPrefix(mMarkPosition);
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1169,8 +1169,11 @@ NS_IMETHODIMP CViewSourceHTML::HandleToken(CToken* aToken,nsIParser* aParser) {
|
|||
|
||||
case eToken_comment:
|
||||
{
|
||||
const nsAReadableString& commentValue = aToken->GetStringValue();
|
||||
result=WriteTag(mCommentTag,commentValue,0,PR_TRUE);
|
||||
nsAutoString theStr;
|
||||
theStr.Assign(NS_LITERAL_STRING("<!--") +
|
||||
aToken->GetStringValue() +
|
||||
NS_LITERAL_STRING("-->") );
|
||||
result=WriteTag(mCommentTag,theStr,0,PR_TRUE);
|
||||
}
|
||||
break;
|
||||
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче