some bugs and correction

Ambiera Forum

Discussions, Help and Support.

	Ambiera Forum > irrXML
	some bugs and correction

Menuki Guest
Quote
2013-10-22 12:32:36

Hello,

Being a long time user of irrXML, I post some correction I made.
Here the source code of each method I corrected.
All are in CXMLReaderImpl.h

Line 74 : I made parseCurrentNode() returning a boolean to correctly handle the end of file. So now read() look like:

	virtual bool read()
	{
		// if not end reached, parse the node
		if (P && (unsigned int)(P - TextBegin) < TextSize - 1 && *P != 0)
		{
			return parseCurrentNode();
		}

		_IRR_IMPLEMENT_MANAGED_MARSHALLING_BUGFIX;
		return false;
	}

Line 93 : It is not a big problem but this line cause a unsigned to signed conversion warning with VS2010 so:

	virtual int getAttributeCount() const
	{
		return (int)Attributes.size();
	}

Line 213 : parseCurrentNode() now returns a boolean. It returns false when it reach the end of file.

	bool parseCurrentNode()
	{
		char_type* start = P;

		// more forward until '<' found
		while(*P != L'<' && *P)
			++P;

		if (!*P)
			return false;

		if (P - start > 0)
		{
			// we found some text, store it
			if (setText(start, P))
				return true;
		}

		++P;

		// based on current token, parse and report next element
		switch(*P)
		{
		case L'/':
			parseClosingXMLElement(); 
			break;
		case L'?':
			ignoreDefinition();	
			break;
		case L'!':
			if (!parseCDATA())
				parseComment();	
			break;
		default:
			parseOpeningXMLElement();
			break;
		}
		return true;
	}

Line 424: In bad formated file, if closing bracket is missing, parseClosingXMLElement() crashed because there is no "end of file" check in this line. So the while condition :

while(*P != L'>')

become

while(P && (unsigned int)(P - TextBegin) < TextSize - 1 && *P != 0 && *P != L'>')

Line 542: Replacing special characters do not always copy all trailing characters because of the '<' comperand which must be a '<=' comperand. The comparaison becomes :

if (oldPos <= origstr.size()-1)

Line 614: UTF8 file are not detected with VS2010. data8[] element are signed extented to int (0xEF becomes 0xFFFFFFEF) while UTF8[] element are non signed extended to int (0xEF becomes 0xEF). So the condition is never true. We have to tell the compiler to do a non signed extension of data8[] elements.

if (size >= 3 && data8[0] == UTF8[0] && data8[1] == UTF8[1] && data8[2] == UTF8[2])

become

if (size >= 3 &&
	reinterpret_cast<unsigned char *>(data8)[0] == UTF8[0] &&
	reinterpret_cast<unsigned char *>(data8)[1] == UTF8[1] &&
	reinterpret_cast<unsigned char *>(data8)[2] == UTF8[2])

Menuki Guest
Quote
2013-10-22 12:38:22

Finally I implemented a true UTF8 to UNICODE conversion in convertTextData().

Line 659, the code

			TextData = new char_type[sizeWithoutHeader];

			for (int i=0; i<sizeWithoutHeader; ++i)
				TextData[i] = (char_type)source[i];

			TextBegin = TextData;
			TextSize = sizeWithoutHeader;

			// delete original data because no longer needed
			delete [] pointerToStore;

become

define         MASKBITS                0x3F
define         MASKBYTE                0x80
define         MASK2BYTES              0xC0
define         MASK3BYTES              0xE0
define         MASK4BYTES              0xF0
define         MASK5BYTES              0xF8
define         MASK6BYTES              0xFC

			TextData = new char_type[sizeWithoutHeader];
			int i,n;
			for (i=n=0; i<sizeWithoutHeader;)
			{
				char_type ch;

				// 1110xxxx 10xxxxxx 10xxxxxx
				if ((source[i] & MASK3BYTES) == MASK3BYTES)
				{
					ch = ((source[i] & 0x0F) << 12) | (
						  (source[i+1] & MASKBITS) << 6)
						| (source[i+2] & MASKBITS);
					i += 3;
				}
				// 110xxxxx 10xxxxxx
				else if ((source[i] & MASK2BYTES) == MASK2BYTES)
				{
					ch = ((source[i] & 0x1F) << 6) | (source[i+1] & MASKBITS);
					i += 2;
				}
				// 0xxxxxxx
				else if (source[i] < MASKBYTE)
				{
					ch = (char_type)source[i];
					i += 1;
				}
				else
				{
					++i;
					continue;
				}
				TextData[n++] = ch;
			}

			TextBegin = TextData;
			TextSize = n;

			// delete original data because no longer needed
			delete [] pointerToStore;

I hope this can help irrXML users to get the best of this good tool.

niko Moderator
Quote
2013-10-23 05:58:23

Thanks :)

kyiv Registered User
Quote
2014-05-24 19:16:39

At first glance the Menuki's fix looks correct and promising.
It was done more than a half year ago.
But I didn't see these changes in the available currently irrXML archive with sources.
Does anybody have explanation why?
Has irrXML available source repository and bug tracker?
I can't find them in SF, but only IrrlichtEngine :(

niko Moderator
Quote
2014-05-25 20:56:12

Yes, irrXML isn't actively developed anymore, sorry.

Create reply:

Possible Codes

Feature	Code
Link	[url] www.example.com [/url]
Bold	[b]bold text[/b]
Image	[img]http://www.example.com/image.jpg[/img]
Quote	[quote]quoted text[/quote]
Code	[code]source code[/code]

Emoticons