[XERCESC-1385] UTF8 parse failure when there's a bom in the utf8 header

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.6.0
• Fix Version/s: 2.7.0
• Component/s: SAX/SAX2
• Environment: OSX, CodeWarrior 9.4
• Assignee: Unassigned
• Reporter: Miklos Fazekas
• Created: 17/Mar/05 19:57
• Updated: 12/Mar/18 04:05
• Resolved: 01/Mar/06 01:01

Description

This issue probably related to 1284. (Or a duplicate of it)

The attached sample code failes with Xerces2.6.

The problem seems to be that there's double checing for the utf8 bom. Bellow is a patch to XMLParser.cpp that resolves this issue. [ The bug is that we've already detected utf8 bom and modified fRawBufIndex, but the seconds check doesn't takes it into accout. ]
src/xercesc/internal/XMLReader.cpp

@@ -544,7 +544,7 @@
}
// If there's a utf-8 BOM (0xEF 0xBB 0xBF), skip past it.
else {

• const char* asChars = (const char*)fRawByteBuf;
  + const char* asChars = (const char*)(fRawByteBuf + fRawBufIndex);
  if ((fRawBytesAvail > XMLRecognizer::fgUTF8BOMLen )&&
  (XMLString::compareNString( asChars
  , XMLRecognizer::fgUTF8BOM

It's also possible that we should check if we detected an utf8 bom already as the following code would probably allow a double utf8 bom.